API tools faq
paste
NickCanCode

beellama logs

NickCanCode
Jun 5th, 2026
48
0
Never
Add comment
Not a member of Pastebin yet? Sign Up, it unlocks many cool features!
text 148.66 KB | Software | 0 0
raw download clone embed print report
  1. 0.00.088.073 I dflash: setting -cd to 256 (drafter doesn't need the full main ctx; pass -cd N to override)
  2. 0.00.088.115 I log_info: verbosity = 3 (adjust with the `-lv N` CLI arg)
  3. 0.00.088.116 I device_info:
  4. 0.00.247.001 I - CUDA0 : NVIDIA GeForce RTX 5070 Ti (16302 MiB, 15037 MiB free)
  5. 0.00.394.408 I - CUDA1 : NVIDIA GeForce RTX 5070 Ti (16302 MiB, 15037 MiB free)
  6. 0.00.394.421 I - CPU : AMD Ryzen 5 5600GT with Radeon Graphics (48512 MiB, 32809 MiB free)
  7. 0.00.394.484 I system_info: n_threads = 6 (n_threads_batch = 6) / 12 | CUDA : ARCHS = 1200 | USE_GRAPHS = 1 | PEER_MAX_BATCH_SIZE = 128 | FA_ALL_QUANTS = 1 | BLACKWELL_NATIVE_FP4 = 1 | CPU : SSE3 = 1 | SSSE3 = 1 | AVX = 1 | AVX2 = 1 | F16C = 1 | FMA = 1 | LLAMAFILE = 1 | OPENMP = 1 | REPACK = 1 |
  8. 0.00.394.558 I srv init: api_keys: ****lama
  9. 0.00.394.569 I srv init: using 11 threads for HTTP server
  10. 0.00.394.743 I srv start: binding port with default address family
  11. 0.00.406.126 I srv llama_server: loading model
  12. 0.00.406.151 I srv load_model: loading model 'A:\models\Qwen3.6-27B-Q2_K_MIXED-AutoRound-MTP.gguf'
  13. 0.00.933.474 I srv load_model: [spec] estimated memory usage of draft model is 3825.44 MiB
  14. 0.00.933.573 I common_init_result: fitting params to device memory ...
  15. 0.00.933.574 I common_init_result: (for bugs during this step try to reproduce them with -fit off, or provide --verbose logs if the bug only occurs with -fit on)
  16. 0.06.622.650 W llama_context: n_ctx_seq (160000) < n_ctx_train (262144) -- the full capacity of the model will not be utilized
  17. 0.06.772.329 I common_init_from_params: warming up the model with an empty run - please wait ... (--no-warmup to disable)
  18. TCQ decode: context-adaptive V alpha enabled
  19. 0.07.021.390 I srv load_model: shrunk recurrent state to 1 cells before draft load (deferred 1 backup cells)
  20. 0.07.021.396 I srv load_model: loading draft model 'S:\LLMs\Qwen3.6-27B-DFlash-IQ4_XS.gguf'
  21. 0.07.021.468 I srv load_model: DFlash draft model will use target output device CUDA1 by default; pass --spec-draft-device to override
  22. 0.07.952.910 I srv load_model: expanded recurrent state to 2 cells before speculative GPU buffers
  23. 0.07.952.914 I srv load_model: initializing slots, n_slots = 1
  24. 0.08.138.673 W seq_rm: RS-ROLLBACK-OVERFLOW seq=0 p0=1 cell.pos=1 rollback=1 > n_rs_seq=0 -> CANNOT ROLLBACK
  25. 0.08.153.643 W srv load_model: speculative decoding will use checkpoints
  26. 0.08.153.650 I srv load_model: DFlash enabled for all 1 slots
  27. 0.08.153.738 W llama_context: n_ctx_seq (256) < n_ctx_train (262144) -- the full capacity of the model will not be utilized
  28. 0.08.230.883 I common_speculative_create_ctx_dft: draft model warmup complete
  29. 0.08.230.921 I common_speculative_init: adding implementation dflash
  30. 0.08.230.979 I dflash: contract ok: block_size=16 mask_token=248070 target_layer_ids=[1,16,31,46,61] n_target_layers=5 n_embd=5120 n_target_features=25600 target_layers=65 cross_ctx=1024
  31. 0.08.230.986 I dflash: target/drafter info: target_ctx_train=262144 target_vocab=248320 drafter_vocab=248320 vocab_match=1 capture_min=1 capture_max=61
  32. 0.08.341.488 I dflash: multi-GPU placement detected (target=2 devices, drafter=1 devices); enabling GPU cross ring and graph hidden capture
  33. 0.08.341.499 I dflash: GPU hidden capture policy: allowed=1 forced_cpu=0 requested=1 target_devices=2 drafter_devices=1
  34. dflash gpu ring: allocated 5 layers x 1024 slots x 5120 embd + staging (~200 MB)
  35. 0.08.344.082 I dflash: GPU cross ring enabled (5 layers x 1024 slots x 5120 embd)
  36. 0.08.344.111 I slot load_model: id 0 | task -1 | speculative decoding context initialized
  37. 0.08.344.113 I slot load_model: id 0 | task -1 | new slot, n_ctx = 160000
  38. 0.08.350.274 I srv load_model: prompt cache is enabled, size limit: 8192 MiB
  39. 0.08.350.279 I srv load_model: use `--cache-ram 0` to disable the prompt cache
  40. 0.08.350.280 I srv load_model: for more info see https://github.com/ggml-org/llama.cpp/pull/16391
  41. 0.08.350.281 I srv load_model: context checkpoints enabled, max = 32, min spacing = 1024
  42. 0.08.350.361 I srv init: idle slots will be saved to prompt cache and cleared upon starting a new task
  43. 0.08.383.301 I init: chat template, example_format: '<|im_start|>system
  44. You are a helpful assistant<|im_end|>
  45. <|im_start|>user
  46. Hello<|im_end|>
  47. <|im_start|>assistant
  48. Hi there<|im_end|>
  49. <|im_start|>user
  50. How are you?<|im_end|>
  51. <|im_start|>assistant
  52. <think>
  53. '
  54. 0.08.402.578 I srv init: init: chat template, thinking = 1
  55. 0.08.402.622 I srv llama_server: model loaded
  56. 0.08.402.627 I srv llama_server: server is listening on http://127.0.0.1:8080
  57. 0.08.402.651 I srv update_slots: all slots are idle
  58. 1.14.196.406 I srv params_from_: Chat format: peg-native
  59. 1.14.200.028 I slot get_availabl: id 0 | task -1 | selected slot by LRU, t_last = -1
  60. 1.14.200.043 I slot get_availabl: id 0 | task -1 | adaptive dm: reset state for LRU slot selection
  61. 1.14.414.730 I srv recurrent_sh: shrunk recurrent state to 1 cells for prompt cache (before prompt cache save/load, removed 1 backup cells)
  62. 1.14.414.735 I srv get_availabl: updating prompt cache
  63. 1.14.414.740 I srv load: - looking for better prompt, base f_keep = -1.000, sim = 0.000
  64. 1.14.414.747 I srv update: - cache state: 0 prompts, 0.000 MiB (limits: 8192.000 MiB, 160000 tokens, 8589934592 est)
  65. 1.14.596.433 I srv recurrent_ex: expanded recurrent state to 2 cells after prompt cache (after prompt cache save/load)
  66. 1.14.596.438 I srv get_availabl: prompt cache update took 181.70 ms
  67. 1.14.596.506 I reasoning-budget: activated, budget=2147483647 tokens
  68. 1.14.596.507 I reasoning-budget: deactivated (natural end)
  69. 1.14.596.534 I slot launch_slot_: id 0 | task 0 | processing task, is_child = 0
  70. 1.16.439.369 I dflash: drafter K/V projection cache enabled (1024-token window)
  71. 1.17.154.411 I slot create_check: id 0 | task 0 | created context checkpoint 1 of 32 (pos_min = 2883, pos_max = 2883, n_tokens = 2884, size = 247.771 MiB)
  72. 1.17.424.606 I slot operator (): id 0 | task 0 | adaptive dm profit: cur=0 recommended=4 score=24.8 action=apply
  73. 1.17.663.236 I slot operator (): id 0 | task 0 | adaptive dm profit: cur=4 recommended=8 score=24.8 action=apply
  74. 1.17.884.879 I slot operator (): id 0 | task 0 | adaptive dm profit: cur=8 recommended=15 score=24.8 action=apply
  75. 1.19.578.986 I slot print_timing: id 0 | task 0 | n_decoded = 101, tg = 44.44 t/s
  76. 1.22.579.202 I slot print_timing: id 0 | task 0 | n_decoded = 414, tg = 78.49 t/s
  77. 1.25.583.873 I slot print_timing: id 0 | task 0 | n_decoded = 697, tg = 84.20 t/s
  78. 1.28.594.709 I slot print_timing: id 0 | task 0 | n_decoded = 1106, tg = 97.98 t/s
  79. 1.31.621.542 I slot print_timing: id 0 | task 0 | n_decoded = 1504, tg = 105.05 t/s
  80. 1.34.622.204 I slot print_timing: id 0 | task 0 | n_decoded = 1798, tg = 103.83 t/s
  81. 1.37.629.411 I slot print_timing: id 0 | task 0 | n_decoded = 2096, tg = 103.13 t/s
  82. 1.40.647.300 I slot print_timing: id 0 | task 0 | n_decoded = 2302, tg = 98.62 t/s
  83. 1.43.679.854 I slot print_timing: id 0 | task 0 | n_decoded = 2487, tg = 94.29 t/s
  84. 1.46.700.524 I slot print_timing: id 0 | task 0 | n_decoded = 2951, tg = 100.39 t/s
  85. 1.49.704.542 I slot print_timing: id 0 | task 0 | n_decoded = 3196, tg = 98.64 t/s
  86. 1.52.739.561 I slot print_timing: id 0 | task 0 | n_decoded = 3478, tg = 98.15 t/s
  87. 1.55.747.506 I slot print_timing: id 0 | task 0 | n_decoded = 3657, tg = 95.13 t/s
  88. 1.58.795.158 I slot print_timing: id 0 | task 0 | n_decoded = 3831, tg = 92.34 t/s
  89. 2.01.809.421 I slot print_timing: id 0 | task 0 | n_decoded = 4103, tg = 92.19 t/s
  90. 2.04.815.885 I slot print_timing: id 0 | task 0 | n_decoded = 4354, tg = 91.64 t/s
  91. 2.06.604.343 I slot print_timing: id 0 | task 0 | prompt eval time = 2707.04 ms / 2903 tokens ( 0.93 ms per token, 1072.39 tokens per second)
  92. 2.06.604.350 I slot print_timing: id 0 | task 0 | eval time = 49299.91 ms / 4481 tokens ( 11.00 ms per token, 90.89 tokens per second)
  93. 2.06.604.353 I slot print_timing: id 0 | task 0 | total time = 52006.95 ms / 7384 tokens
  94. 2.06.604.367 I slot print_timing: id 0 | task 0 | graphs reused = 928
  95. 2.06.604.369 I slot print_timing: id 0 | task 0 | draft acceptance = 0.24091 ( 3503 accepted / 14541 generated)
  96. 2.06.604.370 I slot print_timing: id 0 | task 0 | adaptive dm: fringe=0.00 n_max=15
  97. 2.06.604.400 I statistics dflash: #calls(b,g,a) = 1 976 976, #gen drafts = 976, #acc drafts = 803, #gen tokens = 14541, #acc tokens = 3503, dur(b,g,a) = 0.001, 7583.688, 0.321 ms
  98. 2.06.604.900 I slot release: id 0 | task 0 | stop processing: n_tokens = 7385, truncated = 0
  99. 2.06.604.951 I srv update_slots: all slots are idle
  100. 28.38.115.599 I srv params_from_: Chat format: peg-native
  101. 28.38.168.636 I slot get_availabl: id 0 | task -1 | selected slot by LRU, t_last = 119583427
  102. 28.38.168.644 I slot get_availabl: id 0 | task -1 | adaptive dm: reset state for LRU slot selection
  103. 28.38.480.709 I srv recurrent_sh: shrunk recurrent state to 1 cells for prompt cache (before prompt cache save/load, removed 1 backup cells)
  104. 28.38.480.714 I srv get_availabl: updating prompt cache
  105. 28.38.481.533 W srv prompt_save: - saving prompt with length 7385, total state size = 394.972 MiB (draft: 0.000 MiB)
  106. 28.38.788.859 I srv load: - looking for better prompt, base f_keep = 0.000, sim = 0.000
  107. 28.38.788.868 I srv update: - cache state: 1 prompts, 642.743 MiB (limits: 8192.000 MiB, 160000 tokens, 160000 est)
  108. 28.38.788.870 I srv update: - prompt 0000017C07DA3EE0: 7385 tokens, checkpoints: 1, 642.743 MiB
  109. 28.38.949.714 I srv recurrent_ex: expanded recurrent state to 2 cells after prompt cache (after prompt cache save/load)
  110. 28.38.949.720 I srv get_availabl: prompt cache update took 469.00 ms
  111. 28.38.950.779 I reasoning-budget: activated, budget=2147483647 tokens
  112. 28.38.951.157 I slot launch_slot_: id 0 | task 985 | processing task, is_child = 0
  113. 28.38.951.182 I slot update_slots: id 0 | task 985 | Checking checkpoint with [2883, 2883] against 0...
  114. 28.38.951.196 W slot update_slots: id 0 | task 985 | forcing full prompt re-processing due to lack of cache data (likely due to SWA or hybrid/recurrent memory, see https://github.com/ggml-org/llama.cpp/pull/13194#issuecomment-2868343055)
  115. 28.38.951.202 W slot update_slots: id 0 | task 985 | erased invalidated context checkpoint (pos_min = 2883, pos_max = 2883, n_tokens = 2884, n_swa = 0, pos_next = 0, size = 247.771 MiB)
  116. 28.39.152.709 W find_slot: non-consecutive token position 511 after 3 for sequence 0 with 512 new tokens
  117. 28.39.154.508 W find_slot: non-consecutive token position 511 after 3 for sequence 0 with 512 new tokens
  118. 28.41.951.803 I slot print_timing: id 0 | task 985 | prompt processing, n_tokens = 4096, progress = 0.20, t = 3.00 s / 1365.05 tokens per second
  119. 28.43.257.008 I slot print_timing: id 0 | task 985 | prompt processing, n_tokens = 6144, progress = 0.30, t = 4.31 s / 1426.91 tokens per second
  120. 28.44.616.071 I slot print_timing: id 0 | task 985 | prompt processing, n_tokens = 8192, progress = 0.40, t = 5.66 s / 1446.10 tokens per second
  121. 28.45.996.391 I slot print_timing: id 0 | task 985 | prompt processing, n_tokens = 10240, progress = 0.50, t = 7.05 s / 1453.47 tokens per second
  122. 28.47.389.147 I slot print_timing: id 0 | task 985 | prompt processing, n_tokens = 12288, progress = 0.60, t = 8.44 s / 1456.28 tokens per second
  123. 28.48.833.007 I slot print_timing: id 0 | task 985 | prompt processing, n_tokens = 14336, progress = 0.70, t = 9.88 s / 1450.74 tokens per second
  124. 28.50.307.402 I slot print_timing: id 0 | task 985 | prompt processing, n_tokens = 16384, progress = 0.80, t = 11.36 s / 1442.73 tokens per second
  125. 28.51.822.041 I slot print_timing: id 0 | task 985 | prompt processing, n_tokens = 18432, progress = 0.90, t = 12.87 s / 1432.07 tokens per second
  126. 28.53.137.810 I slot print_timing: id 0 | task 985 | prompt processing, n_tokens = 20056, progress = 0.97, t = 14.19 s / 1413.73 tokens per second
  127. 28.53.530.043 I slot print_timing: id 0 | task 985 | prompt processing, n_tokens = 20517, progress = 1.00, t = 14.58 s / 1407.31 tokens per second
  128. 28.53.623.682 I slot create_check: id 0 | task 985 | created context checkpoint 1 of 32 (pos_min = 20516, pos_max = 20516, n_tokens = 20517, size = 244.255 MiB)
  129. 28.53.704.531 I slot print_timing: id 0 | task 985 | prompt processing, n_tokens = 20568, progress = 1.00, t = 14.75 s / 1394.12 tokens per second
  130. 28.53.879.064 I slot operator (): id 0 | task 985 | adaptive dm profit: cur=0 recommended=4 score=25.9 action=apply
  131. 28.54.083.477 I slot operator (): id 0 | task 985 | adaptive dm profit: cur=4 recommended=8 score=25.9 action=apply
  132. 28.54.277.016 I slot operator (): id 0 | task 985 | adaptive dm profit: cur=8 recommended=15 score=25.9 action=apply
  133. 28.55.671.956 I slot print_timing: id 0 | task 985 | n_decoded = 104, tg = 54.55 t/s
  134. 28.55.829.268 I reasoning-budget: deactivated (natural end)
  135. 28.56.030.819 W slot process_toke: id 0 | task 985 | raw tool marker observed while lazy grammar is enabled; keeping DFlash governed by active grammar boundary in_reasoning=0 n_decoded=117 reasoning_tokens=108 visible_tokens=9
  136. 28.58.724.464 I slot print_timing: id 0 | task 985 | n_decoded = 231, tg = 46.57 t/s
  137. 29.00.478.573 I slot print_timing: id 0 | task 985 | prompt eval time = 14812.08 ms / 20572 tokens ( 0.72 ms per token, 1388.87 tokens per second)
  138. 29.00.478.580 I slot print_timing: id 0 | task 985 | eval time = 6712.47 ms / 282 tokens ( 23.80 ms per token, 42.01 tokens per second)
  139. 29.00.478.582 I slot print_timing: id 0 | task 985 | total time = 21524.55 ms / 20854 tokens
  140. 29.00.478.583 I slot print_timing: id 0 | task 985 | graphs reused = 1016
  141. 29.00.478.585 I slot print_timing: id 0 | task 985 | draft acceptance = 0.10962 ( 171 accepted / 1560 generated)
  142. 29.00.478.586 I slot print_timing: id 0 | task 985 | adaptive dm: fringe=0.00 n_max=15
  143. 29.00.478.614 I statistics dflash: #calls(b,g,a) = 2 1085 1085, #gen drafts = 1085, #acc drafts = 881, #gen tokens = 16101, #acc tokens = 3674, dur(b,g,a) = 0.001, 8440.465, 0.351 ms
  144. 29.00.480.088 I slot release: id 0 | task 985 | stop processing: n_tokens = 20855, truncated = 0
  145. 29.00.480.120 I srv update_slots: all slots are idle
  146. 29.03.462.915 I srv params_from_: Chat format: peg-native
  147. 29.03.497.386 I slot get_availabl: id 0 | task -1 | selected slot by LCP similarity, sim_best = 0.985 (> 0.100 thold), f_keep = 1.000
  148. 29.03.497.393 I slot get_availabl: id 0 | task -1 | adaptive dm: preserving state for continuation (sim=0.985, keep=1.000, n_max=15)
  149. 29.03.498.169 I reasoning-budget: activated, budget=2147483647 tokens
  150. 29.03.498.492 I slot launch_slot_: id 0 | task 1111 | processing task, is_child = 0
  151. 29.06.118.911 I slot print_timing: id 0 | task 1111 | n_decoded = 101, tg = 45.35 t/s
  152. 29.06.434.475 I reasoning-budget: deactivated (natural end)
  153. 29.07.449.972 W slot process_toke: id 0 | task 1111 | raw tool marker observed while lazy grammar is enabled; keeping DFlash governed by active grammar boundary in_reasoning=0 n_decoded=175 reasoning_tokens=120 visible_tokens=55
  154. 29.09.141.081 I slot print_timing: id 0 | task 1111 | n_decoded = 235, tg = 44.77 t/s
  155. 29.09.292.911 I slot print_timing: id 0 | task 1111 | prompt eval time = 391.10 ms / 326 tokens ( 1.20 ms per token, 833.54 tokens per second)
  156. 29.09.292.917 I slot print_timing: id 0 | task 1111 | eval time = 5401.66 ms / 240 tokens ( 22.51 ms per token, 44.43 tokens per second)
  157. 29.09.292.918 I slot print_timing: id 0 | task 1111 | total time = 5792.76 ms / 566 tokens
  158. 29.09.292.919 I slot print_timing: id 0 | task 1111 | graphs reused = 1103
  159. 29.09.292.921 I slot print_timing: id 0 | task 1111 | draft acceptance = 0.09323 ( 139 accepted / 1491 generated)
  160. 29.09.292.923 I slot print_timing: id 0 | task 1111 | adaptive dm: fringe=0.00 n_max=15
  161. 29.09.292.944 I statistics dflash: #calls(b,g,a) = 3 1186 1186, #gen drafts = 1186, #acc drafts = 950, #gen tokens = 17592, #acc tokens = 3813, dur(b,g,a) = 0.002, 9168.691, 0.375 ms
  162. 29.09.294.096 I slot release: id 0 | task 1111 | stop processing: n_tokens = 21421, truncated = 0
  163. 29.09.294.232 I srv update_slots: all slots are idle
  164. 29.14.807.598 I srv params_from_: Chat format: peg-native
  165. 29.14.855.442 I slot get_availabl: id 0 | task -1 | selected slot by LCP similarity, sim_best = 0.997 (> 0.100 thold), f_keep = 1.000
  166. 29.14.855.448 I slot get_availabl: id 0 | task -1 | adaptive dm: preserving state for continuation (sim=0.997, keep=1.000, n_max=15)
  167. 29.14.856.402 I reasoning-budget: activated, budget=2147483647 tokens
  168. 29.14.856.723 I slot launch_slot_: id 0 | task 1215 | processing task, is_child = 0
  169. 29.15.811.124 I reasoning-budget: deactivated (natural end)
  170. 29.15.813.337 W slot process_toke: id 0 | task 1215 | raw tool marker observed while lazy grammar is enabled; keeping DFlash governed by active grammar boundary in_reasoning=0 n_decoded=32 reasoning_tokens=30 visible_tokens=2
  171. 29.17.655.407 I slot print_timing: id 0 | task 1215 | prompt eval time = 233.87 ms / 61 tokens ( 3.83 ms per token, 260.82 tokens per second)
  172. 29.17.655.457 I slot print_timing: id 0 | task 1215 | eval time = 2561.93 ms / 83 tokens ( 30.87 ms per token, 32.40 tokens per second)
  173. 29.17.655.459 I slot print_timing: id 0 | task 1215 | total time = 2795.81 ms / 144 tokens
  174. 29.17.655.463 I slot print_timing: id 0 | task 1215 | graphs reused = 1145
  175. 29.17.655.466 I slot print_timing: id 0 | task 1215 | draft acceptance = 0.05778 ( 39 accepted / 675 generated)
  176. 29.17.655.467 I slot print_timing: id 0 | task 1215 | adaptive dm: fringe=0.00 n_max=15
  177. 29.17.655.492 I statistics dflash: #calls(b,g,a) = 4 1231 1231, #gen drafts = 1231, #acc drafts = 977, #gen tokens = 18267, #acc tokens = 3852, dur(b,g,a) = 0.002, 9482.354, 0.386 ms
  178. 29.17.657.078 I slot release: id 0 | task 1215 | stop processing: n_tokens = 21566, truncated = 0
  179. 29.17.657.253 I srv update_slots: all slots are idle
  180. 29.19.080.189 I srv params_from_: Chat format: peg-native
  181. 29.19.115.585 I slot get_availabl: id 0 | task -1 | selected slot by LCP similarity, sim_best = 0.990 (> 0.100 thold), f_keep = 0.998
  182. 29.19.115.592 I slot get_availabl: id 0 | task -1 | adaptive dm: preserving state for continuation (sim=0.990, keep=0.998, n_max=15)
  183. 29.19.116.382 I reasoning-budget: activated, budget=2147483647 tokens
  184. 29.19.116.553 I slot launch_slot_: id 0 | task 1264 | processing task, is_child = 0
  185. 29.19.116.576 I slot update_slots: id 0 | task 1264 | Checking checkpoint with [20516, 20516] against 21514...
  186. 29.19.210.036 W slot update_slots: id 0 | task 1264 | restored context checkpoint (pos_min = 20516, pos_max = 20516, n_tokens = 20517, n_past = 20517, size = 244.255 MiB)
  187. 29.19.214.585 W find_slot: non-consecutive token position 21028 after 20518 for sequence 0 with 512 new tokens
  188. 29.19.215.023 W find_slot: non-consecutive token position 21028 after 20518 for sequence 0 with 512 new tokens
  189. 29.20.180.025 I slot create_check: id 0 | task 1264 | created context checkpoint 2 of 32 (pos_min = 21565, pos_max = 21565, n_tokens = 21566, size = 249.626 MiB)
  190. 29.20.799.275 I reasoning-budget: deactivated (natural end)
  191. 29.21.198.219 W slot process_toke: id 0 | task 1264 | raw tool marker observed while lazy grammar is enabled; keeping DFlash governed by active grammar boundary in_reasoning=0 n_decoded=38 reasoning_tokens=15 visible_tokens=23
  192. 29.22.828.853 I slot print_timing: id 0 | task 1264 | prompt eval time = 1276.95 ms / 1219 tokens ( 1.05 ms per token, 954.61 tokens per second)
  193. 29.22.828.858 I slot print_timing: id 0 | task 1264 | eval time = 2434.52 ms / 85 tokens ( 28.64 ms per token, 34.91 tokens per second)
  194. 29.22.828.859 I slot print_timing: id 0 | task 1264 | total time = 3711.47 ms / 1304 tokens
  195. 29.22.828.860 I slot print_timing: id 0 | task 1264 | graphs reused = 1178
  196. 29.22.828.862 I slot print_timing: id 0 | task 1264 | draft acceptance = 0.09444 ( 51 accepted / 540 generated)
  197. 29.22.828.862 I slot print_timing: id 0 | task 1264 | adaptive dm: fringe=0.00 n_max=15
  198. 29.22.828.883 I statistics dflash: #calls(b,g,a) = 5 1267 1267, #gen drafts = 1267, #acc drafts = 1001, #gen tokens = 18807, #acc tokens = 3903, dur(b,g,a) = 0.004, 9747.140, 0.391 ms
  199. 29.22.830.160 I slot release: id 0 | task 1264 | stop processing: n_tokens = 21823, truncated = 0
  200. 29.22.830.304 I srv update_slots: all slots are idle
  201. 29.25.057.511 I srv params_from_: Chat format: peg-native
  202. 29.25.092.346 I slot get_availabl: id 0 | task -1 | selected slot by LCP similarity, sim_best = 0.994 (> 0.100 thold), f_keep = 0.998
  203. 29.25.092.390 I slot get_availabl: id 0 | task -1 | adaptive dm: preserving state for continuation (sim=0.994, keep=0.998, n_max=15)
  204. 29.25.093.101 I reasoning-budget: activated, budget=2147483647 tokens
  205. 29.25.093.260 I slot launch_slot_: id 0 | task 1305 | processing task, is_child = 0
  206. 29.25.093.287 I slot update_slots: id 0 | task 1305 | Checking checkpoint with [21565, 21565] against 21774...
  207. 29.25.202.538 W slot update_slots: id 0 | task 1305 | restored context checkpoint (pos_min = 21565, pos_max = 21565, n_tokens = 21566, n_past = 21566, size = 249.626 MiB)
  208. 29.25.207.649 W find_slot: non-consecutive token position 21821 after 21568 for sequence 0 with 256 new tokens
  209. 29.25.207.846 W find_slot: non-consecutive token position 21821 after 21568 for sequence 0 with 256 new tokens
  210. 29.25.709.382 I reasoning-budget: deactivated (natural end)
  211. 29.26.726.284 I slot print_timing: id 0 | task 1305 | prompt eval time = 551.22 ms / 337 tokens ( 1.64 ms per token, 611.37 tokens per second)
  212. 29.26.726.290 I slot print_timing: id 0 | task 1305 | eval time = 1078.89 ms / 69 tokens ( 15.64 ms per token, 63.95 tokens per second)
  213. 29.26.726.292 I slot print_timing: id 0 | task 1305 | total time = 1630.11 ms / 406 tokens
  214. 29.26.726.293 I slot print_timing: id 0 | task 1305 | graphs reused = 1198
  215. 29.26.726.295 I slot print_timing: id 0 | task 1305 | draft acceptance = 0.15556 ( 49 accepted / 315 generated)
  216. 29.26.726.296 I slot print_timing: id 0 | task 1305 | adaptive dm: fringe=0.00 n_max=15
  217. 29.26.726.321 I statistics dflash: #calls(b,g,a) = 6 1288 1288, #gen drafts = 1288, #acc drafts = 1018, #gen tokens = 19122, #acc tokens = 3952, dur(b,g,a) = 0.004, 9905.156, 0.395 ms
  218. 29.26.728.460 I slot release: id 0 | task 1305 | stop processing: n_tokens = 21973, truncated = 0
  219. 29.26.728.592 I srv update_slots: all slots are idle
  220. 33.07.088.487 I srv params_from_: Chat format: peg-native
  221. 33.07.122.524 I slot get_availabl: id 0 | task -1 | selected slot by LCP similarity, sim_best = 0.851 (> 0.100 thold), f_keep = 0.936
  222. 33.07.122.531 I slot get_availabl: id 0 | task -1 | adaptive dm: reset state for prompt change (sim=0.851, keep=0.936, preserve_policy=0, n_max=15)
  223. 33.07.123.206 I reasoning-budget: activated, budget=2147483647 tokens
  224. 33.07.123.370 I slot launch_slot_: id 0 | task 1330 | processing task, is_child = 0
  225. 33.07.123.395 I slot update_slots: id 0 | task 1330 | Checking checkpoint with [21565, 21565] against 20569...
  226. 33.07.123.396 I slot update_slots: id 0 | task 1330 | Checking checkpoint with [20516, 20516] against 20569...
  227. 33.07.349.982 W slot update_slots: id 0 | task 1330 | restored context checkpoint (pos_min = 20516, pos_max = 20516, n_tokens = 20517, n_past = 20517, size = 244.255 MiB)
  228. 33.07.349.990 W slot update_slots: id 0 | task 1330 | erased invalidated context checkpoint (pos_min = 21565, pos_max = 21565, n_tokens = 21566, n_swa = 0, pos_next = 20517, size = 249.626 MiB)
  229. 33.07.383.148 W find_slot: non-consecutive token position 21028 after 20521 for sequence 0 with 512 new tokens
  230. 33.07.384.220 W find_slot: non-consecutive token position 21028 after 20521 for sequence 0 with 512 new tokens
  231. 33.08.722.912 I slot create_check: id 0 | task 1330 | created context checkpoint 2 of 32 (pos_min = 21682, pos_max = 21682, n_tokens = 21683, size = 244.255 MiB)
  232. 33.10.263.839 I slot print_timing: id 0 | task 1330 | prompt processing, n_tokens = 3126, progress = 0.98, t = 3.14 s / 995.41 tokens per second
  233. 33.10.380.768 I slot create_check: id 0 | task 1330 | created context checkpoint 3 of 32 (pos_min = 23642, pos_max = 23642, n_tokens = 23643, size = 249.626 MiB)
  234. 33.10.793.046 I slot print_timing: id 0 | task 1330 | prompt processing, n_tokens = 3638, progress = 1.00, t = 3.67 s / 991.37 tokens per second
  235. 33.11.585.677 I reasoning-budget: deactivated (natural end)
  236. 33.11.843.069 W slot process_toke: id 0 | task 1330 | raw tool marker observed while lazy grammar is enabled; keeping DFlash governed by active grammar boundary in_reasoning=0 n_decoded=64 reasoning_tokens=47 visible_tokens=17
  237. 33.13.195.205 I slot print_timing: id 0 | task 1330 | n_decoded = 100, tg = 42.66 t/s
  238. 33.15.425.571 I slot print_timing: id 0 | task 1330 | prompt eval time = 3726.45 ms / 3642 tokens ( 1.02 ms per token, 977.34 tokens per second)
  239. 33.15.425.577 I slot print_timing: id 0 | task 1330 | eval time = 4573.82 ms / 207 tokens ( 22.10 ms per token, 45.26 tokens per second)
  240. 33.15.425.579 I slot print_timing: id 0 | task 1330 | total time = 8300.27 ms / 3849 tokens
  241. 33.15.425.580 I slot print_timing: id 0 | task 1330 | graphs reused = 1265
  242. 33.15.425.581 I slot print_timing: id 0 | task 1330 | draft acceptance = 0.10265 ( 124 accepted / 1208 generated)
  243. 33.15.425.583 I slot print_timing: id 0 | task 1330 | adaptive dm: fringe=0.00 n_max=15
  244. 33.15.425.604 I statistics dflash: #calls(b,g,a) = 7 1370 1370, #gen drafts = 1370, #acc drafts = 1073, #gen tokens = 20330, #acc tokens = 4076, dur(b,g,a) = 0.004, 10519.234, 0.426 ms
  245. 33.15.427.012 I slot release: id 0 | task 1330 | stop processing: n_tokens = 24365, truncated = 0
  246. 33.15.427.217 I srv update_slots: all slots are idle
  247. 33.18.294.133 I srv params_from_: Chat format: peg-native
  248. 33.18.335.569 I slot get_availabl: id 0 | task -1 | selected slot by LCP similarity, sim_best = 0.921 (> 0.100 thold), f_keep = 1.000
  249. 33.18.335.574 I slot get_availabl: id 0 | task -1 | adaptive dm: preserving state for continuation (sim=0.921, keep=1.000, n_max=15)
  250. 33.18.336.303 I reasoning-budget: activated, budget=2147483647 tokens
  251. 33.18.336.605 I slot launch_slot_: id 0 | task 1417 | processing task, is_child = 0
  252. 33.19.922.009 I slot create_check: id 0 | task 1417 | created context checkpoint 4 of 32 (pos_min = 25940, pos_max = 25940, n_tokens = 25941, size = 199.236 MiB)
  253. 33.22.644.994 I slot print_timing: id 0 | task 1417 | n_decoded = 102, tg = 45.48 t/s
  254. 33.25.675.681 I slot print_timing: id 0 | task 1417 | n_decoded = 246, tg = 46.64 t/s
  255. 33.28.706.818 I slot print_timing: id 0 | task 1417 | n_decoded = 362, tg = 43.58 t/s
  256. 33.28.853.362 I reasoning-budget: deactivated (natural end)
  257. 33.30.592.321 W slot process_toke: id 0 | task 1417 | raw tool marker observed while lazy grammar is enabled; keeping DFlash governed by active grammar boundary in_reasoning=0 n_decoded=467 reasoning_tokens=372 visible_tokens=95
  258. 33.31.728.005 I slot print_timing: id 0 | task 1417 | n_decoded = 495, tg = 43.70 t/s
  259. 33.32.601.124 I slot print_timing: id 0 | task 1417 | prompt eval time = 2063.24 ms / 2092 tokens ( 0.99 ms per token, 1013.94 tokens per second)
  260. 33.32.601.130 I slot print_timing: id 0 | task 1417 | eval time = 12199.09 ms / 535 tokens ( 22.80 ms per token, 43.86 tokens per second)
  261. 33.32.601.207 I slot print_timing: id 0 | task 1417 | total time = 14262.33 ms / 2627 tokens
  262. 33.32.601.211 I slot print_timing: id 0 | task 1417 | graphs reused = 1441
  263. 33.32.601.214 I slot print_timing: id 0 | task 1417 | draft acceptance = 0.10105 ( 319 accepted / 3157 generated)
  264. 33.32.601.216 I slot print_timing: id 0 | task 1417 | adaptive dm: fringe=0.00 n_max=15
  265. 33.32.601.237 I statistics dflash: #calls(b,g,a) = 8 1586 1586, #gen drafts = 1586, #acc drafts = 1219, #gen tokens = 23487, #acc tokens = 4395, dur(b,g,a) = 0.005, 12151.169, 0.493 ms
  266. 33.32.602.714 I slot release: id 0 | task 1417 | stop processing: n_tokens = 26992, truncated = 0
  267. 33.32.602.900 I srv update_slots: all slots are idle
  268. 33.34.067.039 I srv params_from_: Chat format: peg-native
  269. 33.34.107.066 I slot get_availabl: id 0 | task -1 | selected slot by LCP similarity, sim_best = 0.994 (> 0.100 thold), f_keep = 0.998
  270. 33.34.107.072 I slot get_availabl: id 0 | task -1 | adaptive dm: preserving state for continuation (sim=0.994, keep=0.998, n_max=15)
  271. 33.34.107.759 I reasoning-budget: activated, budget=2147483647 tokens
  272. 33.34.108.067 I slot launch_slot_: id 0 | task 1638 | processing task, is_child = 0
  273. 33.34.108.099 I slot update_slots: id 0 | task 1638 | Checking checkpoint with [25940, 25940] against 26924...
  274. 33.34.172.840 W slot update_slots: id 0 | task 1638 | restored context checkpoint (pos_min = 25940, pos_max = 25940, n_tokens = 25941, n_past = 25941, size = 199.236 MiB)
  275. 33.34.176.426 W find_slot: non-consecutive token position 26452 after 25941 for sequence 0 with 512 new tokens
  276. 33.34.176.888 W find_slot: non-consecutive token position 26452 after 25941 for sequence 0 with 512 new tokens
  277. 33.35.252.826 I slot create_check: id 0 | task 1638 | created context checkpoint 5 of 32 (pos_min = 26992, pos_max = 26992, n_tokens = 26993, size = 249.626 MiB)
  278. 33.35.411.149 I reasoning-budget: deactivated (natural end)
  279. 33.35.479.404 W slot process_toke: id 0 | task 1638 | raw tool marker observed while lazy grammar is enabled; keeping DFlash governed by active grammar boundary in_reasoning=0 n_decoded=3 reasoning_tokens=1 visible_tokens=2
  280. 33.36.732.635 I slot print_timing: id 0 | task 1638 | prompt eval time = 1303.08 ms / 1137 tokens ( 1.15 ms per token, 872.55 tokens per second)
  281. 33.36.732.642 I slot print_timing: id 0 | task 1638 | eval time = 1320.65 ms / 50 tokens ( 26.41 ms per token, 37.86 tokens per second)
  282. 33.36.732.643 I slot print_timing: id 0 | task 1638 | total time = 2623.73 ms / 1187 tokens
  283. 33.36.732.644 I slot print_timing: id 0 | task 1638 | graphs reused = 1458
  284. 33.36.732.646 I slot print_timing: id 0 | task 1638 | draft acceptance = 0.08036 ( 27 accepted / 336 generated)
  285. 33.36.732.647 I slot print_timing: id 0 | task 1638 | adaptive dm: fringe=0.00 n_max=15
  286. 33.36.732.671 I statistics dflash: #calls(b,g,a) = 9 1609 1609, #gen drafts = 1609, #acc drafts = 1232, #gen tokens = 23823, #acc tokens = 4422, dur(b,g,a) = 0.006, 12322.516, 0.502 ms
  287. 33.36.734.058 I slot release: id 0 | task 1638 | stop processing: n_tokens = 27128, truncated = 0
  288. 33.36.734.238 I srv update_slots: all slots are idle
  289. 33.38.398.242 I srv params_from_: Chat format: peg-native
  290. 33.38.439.968 I slot get_availabl: id 0 | task -1 | selected slot by LCP similarity, sim_best = 0.685 (> 0.100 thold), f_keep = 0.998
  291. 33.38.439.976 I slot get_availabl: id 0 | task -1 | adaptive dm: reset state for prompt change (sim=0.685, keep=0.998, preserve_policy=0, n_max=15)
  292. 33.38.440.658 I reasoning-budget: activated, budget=2147483647 tokens
  293. 33.38.440.948 I slot launch_slot_: id 0 | task 1666 | processing task, is_child = 0
  294. 33.38.440.979 I slot update_slots: id 0 | task 1666 | Checking checkpoint with [26992, 26992] against 27076...
  295. 33.38.524.399 W slot update_slots: id 0 | task 1666 | restored context checkpoint (pos_min = 26992, pos_max = 26992, n_tokens = 26993, n_past = 26993, size = 249.626 MiB)
  296. 33.38.529.319 W find_slot: non-consecutive token position 27128 after 26994 for sequence 0 with 136 new tokens
  297. 33.38.529.466 W find_slot: non-consecutive token position 27128 after 26994 for sequence 0 with 136 new tokens
  298. 33.42.162.145 I slot print_timing: id 0 | task 1666 | prompt processing, n_tokens = 4232, progress = 0.79, t = 3.72 s / 1137.27 tokens per second
  299. 33.43.955.749 I slot print_timing: id 0 | task 1666 | prompt processing, n_tokens = 6280, progress = 0.84, t = 5.51 s / 1138.76 tokens per second
  300. 33.45.789.629 I slot print_timing: id 0 | task 1666 | prompt processing, n_tokens = 8328, progress = 0.89, t = 7.35 s / 1133.27 tokens per second
  301. 33.47.693.354 I slot print_timing: id 0 | task 1666 | prompt processing, n_tokens = 10376, progress = 0.94, t = 9.25 s / 1121.44 tokens per second
  302. 33.49.315.939 I slot print_timing: id 0 | task 1666 | prompt processing, n_tokens = 12037, progress = 0.99, t = 10.87 s / 1106.85 tokens per second
  303. 33.49.439.262 I slot create_check: id 0 | task 1666 | created context checkpoint 6 of 32 (pos_min = 39029, pos_max = 39029, n_tokens = 39030, size = 249.626 MiB)
  304. 33.49.945.921 I slot print_timing: id 0 | task 1666 | prompt processing, n_tokens = 12549, progress = 1.00, t = 11.50 s / 1090.75 tokens per second
  305. 33.50.617.778 I reasoning-budget: deactivated (natural end)
  306. 33.52.250.972 I slot print_timing: id 0 | task 1666 | n_decoded = 103, tg = 45.92 t/s
  307. 33.53.460.649 I slot print_timing: id 0 | task 1666 | prompt eval time = 11566.04 ms / 12553 tokens ( 0.92 ms per token, 1085.33 tokens per second)
  308. 33.53.460.656 I slot print_timing: id 0 | task 1666 | eval time = 3452.75 ms / 166 tokens ( 20.80 ms per token, 48.08 tokens per second)
  309. 33.53.460.657 I slot print_timing: id 0 | task 1666 | total time = 15018.79 ms / 12719 tokens
  310. 33.53.460.659 I slot print_timing: id 0 | task 1666 | graphs reused = 1507
  311. 33.53.460.660 I slot print_timing: id 0 | task 1666 | draft acceptance = 0.12090 ( 107 accepted / 885 generated)
  312. 33.53.460.662 I slot print_timing: id 0 | task 1666 | adaptive dm: fringe=0.00 n_max=15
  313. 33.53.460.686 I statistics dflash: #calls(b,g,a) = 10 1669 1669, #gen drafts = 1669, #acc drafts = 1276, #gen tokens = 24708, #acc tokens = 4529, dur(b,g,a) = 0.006, 12773.483, 0.518 ms
  314. 33.53.462.384 I slot release: id 0 | task 1666 | stop processing: n_tokens = 39713, truncated = 0
  315. 33.53.462.566 I srv update_slots: all slots are idle
  316. 34.53.882.635 I srv params_from_: Chat format: peg-native
  317. 34.53.934.312 I slot get_availabl: id 0 | task -1 | selected slot by LCP similarity, sim_best = 0.615 (> 0.100 thold), f_keep = 0.608
  318. 34.53.934.320 I slot get_availabl: id 0 | task -1 | adaptive dm: reset state for prompt change (sim=0.615, keep=0.608, preserve_policy=0, n_max=15)
  319. 34.53.935.140 I reasoning-budget: activated, budget=2147483647 tokens
  320. 34.53.935.578 I slot launch_slot_: id 0 | task 1736 | processing task, is_child = 0
  321. 34.53.935.610 I slot update_slots: id 0 | task 1736 | Checking checkpoint with [39029, 39029] against 24156...
  322. 34.53.935.611 I slot update_slots: id 0 | task 1736 | Checking checkpoint with [26992, 26992] against 24156...
  323. 34.53.935.612 I slot update_slots: id 0 | task 1736 | Checking checkpoint with [25940, 25940] against 24156...
  324. 34.53.935.613 I slot update_slots: id 0 | task 1736 | Checking checkpoint with [23642, 23642] against 24156...
  325. 34.54.164.299 W slot update_slots: id 0 | task 1736 | restored context checkpoint (pos_min = 23642, pos_max = 23642, n_tokens = 23643, n_past = 23643, size = 249.626 MiB)
  326. 34.54.164.307 W slot update_slots: id 0 | task 1736 | erased invalidated context checkpoint (pos_min = 25940, pos_max = 25940, n_tokens = 25941, n_swa = 0, pos_next = 23643, size = 199.236 MiB)
  327. 34.54.176.837 W slot update_slots: id 0 | task 1736 | erased invalidated context checkpoint (pos_min = 26992, pos_max = 26992, n_tokens = 26993, n_swa = 0, pos_next = 23643, size = 249.626 MiB)
  328. 34.54.195.178 W slot update_slots: id 0 | task 1736 | erased invalidated context checkpoint (pos_min = 39029, pos_max = 39029, n_tokens = 39030, n_swa = 0, pos_next = 23643, size = 249.626 MiB)
  329. 34.54.222.915 W find_slot: non-consecutive token position 24154 after 23645 for sequence 0 with 512 new tokens
  330. 34.54.224.532 W find_slot: non-consecutive token position 24154 after 23645 for sequence 0 with 512 new tokens
  331. 34.57.811.757 I slot print_timing: id 0 | task 1736 | prompt processing, n_tokens = 4096, progress = 0.71, t = 3.88 s / 1056.72 tokens per second
  332. 34.59.561.952 I slot print_timing: id 0 | task 1736 | prompt processing, n_tokens = 6144, progress = 0.76, t = 5.63 s / 1092.00 tokens per second
  333. 35.01.368.154 I slot print_timing: id 0 | task 1736 | prompt processing, n_tokens = 8192, progress = 0.81, t = 7.43 s / 1102.18 tokens per second
  334. 35.03.224.613 I slot print_timing: id 0 | task 1736 | prompt processing, n_tokens = 10240, progress = 0.86, t = 9.29 s / 1102.38 tokens per second
  335. 35.05.120.820 I slot print_timing: id 0 | task 1736 | prompt processing, n_tokens = 12288, progress = 0.91, t = 11.19 s / 1098.59 tokens per second
  336. 35.07.056.694 I slot print_timing: id 0 | task 1736 | prompt processing, n_tokens = 14336, progress = 0.97, t = 13.12 s / 1092.59 tokens per second
  337. 35.07.804.347 I slot print_timing: id 0 | task 1736 | prompt processing, n_tokens = 15116, progress = 0.99, t = 13.87 s / 1089.93 tokens per second
  338. 35.08.235.109 I slot print_timing: id 0 | task 1736 | prompt processing, n_tokens = 15616, progress = 1.00, t = 14.30 s / 1092.07 tokens per second
  339. 35.08.328.184 I slot create_check: id 0 | task 1736 | created context checkpoint 4 of 32 (pos_min = 39258, pos_max = 39258, n_tokens = 39259, size = 249.626 MiB)
  340. 35.08.384.379 I slot print_timing: id 0 | task 1736 | prompt processing, n_tokens = 15628, progress = 1.00, t = 14.45 s / 1081.61 tokens per second
  341. 35.08.826.104 I reasoning-budget: deactivated (natural end)
  342. 35.09.054.134 W slot process_toke: id 0 | task 1736 | raw tool marker observed while lazy grammar is enabled; keeping DFlash governed by active grammar boundary in_reasoning=0 n_decoded=33 reasoning_tokens=23 visible_tokens=10
  343. 35.10.604.574 I slot print_timing: id 0 | task 1736 | prompt eval time = 14511.60 ms / 15632 tokens ( 0.93 ms per token, 1077.21 tokens per second)
  344. 35.10.604.580 I slot print_timing: id 0 | task 1736 | eval time = 2155.32 ms / 97 tokens ( 22.22 ms per token, 45.00 tokens per second)
  345. 35.10.604.581 I slot print_timing: id 0 | task 1736 | total time = 16666.92 ms / 15729 tokens
  346. 35.10.604.582 I slot print_timing: id 0 | task 1736 | graphs reused = 1537
  347. 35.10.604.584 I slot print_timing: id 0 | task 1736 | draft acceptance = 0.11686 ( 61 accepted / 522 generated)
  348. 35.10.604.585 I slot print_timing: id 0 | task 1736 | adaptive dm: fringe=0.00 n_max=15
  349. 35.10.604.606 I statistics dflash: #calls(b,g,a) = 11 1704 1704, #gen drafts = 1704, #acc drafts = 1301, #gen tokens = 25230, #acc tokens = 4590, dur(b,g,a) = 0.007, 13039.567, 0.529 ms
  350. 35.10.606.424 I slot release: id 0 | task 1736 | stop processing: n_tokens = 39371, truncated = 0
  351. 35.10.606.599 I srv update_slots: all slots are idle
  352. 35.26.811.437 I srv params_from_: Chat format: peg-native
  353. 35.26.869.210 I slot get_availabl: id 0 | task -1 | selected slot by LCP similarity, sim_best = 0.840 (> 0.100 thold), f_keep = 1.000
  354. 35.26.869.219 I slot get_availabl: id 0 | task -1 | adaptive dm: reset state for prompt change (sim=0.840, keep=1.000, preserve_policy=0, n_max=15)
  355. 35.26.869.968 I reasoning-budget: activated, budget=2147483647 tokens
  356. 35.26.870.122 I slot launch_slot_: id 0 | task 1783 | processing task, is_child = 0
  357. 35.31.235.523 I slot print_timing: id 0 | task 1783 | prompt processing, n_tokens = 4098, progress = 0.93, t = 4.37 s / 938.75 tokens per second
  358. 35.33.347.323 I slot print_timing: id 0 | task 1783 | prompt processing, n_tokens = 6146, progress = 0.97, t = 6.48 s / 948.87 tokens per second
  359. 35.34.277.623 I slot print_timing: id 0 | task 1783 | prompt processing, n_tokens = 7001, progress = 0.99, t = 7.41 s / 945.13 tokens per second
  360. 35.34.349.675 I slot create_check: id 0 | task 1783 | created context checkpoint 5 of 32 (pos_min = 46371, pos_max = 46371, n_tokens = 46372, size = 199.236 MiB)
  361. 35.34.900.534 I slot print_timing: id 0 | task 1783 | prompt processing, n_tokens = 7513, progress = 1.00, t = 8.03 s / 935.57 tokens per second
  362. 35.37.230.967 I slot print_timing: id 0 | task 1783 | n_decoded = 104, tg = 45.92 t/s
  363. 35.40.258.922 I reasoning-budget: deactivated (natural end)
  364. 35.40.260.265 I slot print_timing: id 0 | task 1783 | n_decoded = 224, tg = 42.30 t/s
  365. 35.41.760.399 W slot process_toke: id 0 | task 1783 | raw tool marker observed while lazy grammar is enabled; keeping DFlash governed by active grammar boundary in_reasoning=0 n_decoded=318 reasoning_tokens=222 visible_tokens=96
  366. 35.43.264.999 I slot print_timing: id 0 | task 1783 | n_decoded = 361, tg = 43.49 t/s
  367. 35.44.223.545 I slot print_timing: id 0 | task 1783 | prompt eval time = 8094.03 ms / 7517 tokens ( 1.08 ms per token, 928.71 tokens per second)
  368. 35.44.223.551 I slot print_timing: id 0 | task 1783 | eval time = 9257.46 ms / 398 tokens ( 23.26 ms per token, 42.99 tokens per second)
  369. 35.44.223.553 I slot print_timing: id 0 | task 1783 | total time = 17351.49 ms / 7915 tokens
  370. 35.44.223.554 I slot print_timing: id 0 | task 1783 | graphs reused = 1662
  371. 35.44.223.555 I slot print_timing: id 0 | task 1783 | draft acceptance = 0.10977 ( 245 accepted / 2232 generated)
  372. 35.44.223.556 I slot print_timing: id 0 | task 1783 | adaptive dm: fringe=0.00 n_max=15
  373. 35.44.223.580 I statistics dflash: #calls(b,g,a) = 12 1856 1856, #gen drafts = 1856, #acc drafts = 1405, #gen tokens = 27462, #acc tokens = 4835, dur(b,g,a) = 0.008, 14183.459, 0.564 ms
  374. 35.44.225.531 I slot release: id 0 | task 1783 | stop processing: n_tokens = 47285, truncated = 0
  375. 35.44.225.693 I srv update_slots: all slots are idle
  376. 35.45.695.321 I srv params_from_: Chat format: peg-native
  377. 35.45.752.585 I slot get_availabl: id 0 | task -1 | selected slot by LCP similarity, sim_best = 0.998 (> 0.100 thold), f_keep = 1.000
  378. 35.45.752.591 I slot get_availabl: id 0 | task -1 | adaptive dm: preserving state for continuation (sim=0.998, keep=1.000, n_max=15)
  379. 35.45.753.370 I reasoning-budget: activated, budget=2147483647 tokens
  380. 35.45.753.550 I slot launch_slot_: id 0 | task 1943 | processing task, is_child = 0
  381. 35.46.829.821 I reasoning-budget: deactivated (natural end)
  382. 35.46.889.662 W slot process_toke: id 0 | task 1943 | raw tool marker observed while lazy grammar is enabled; keeping DFlash governed by active grammar boundary in_reasoning=0 n_decoded=40 reasoning_tokens=38 visible_tokens=2
  383. 35.48.920.515 I slot print_timing: id 0 | task 1943 | n_decoded = 111, tg = 38.42 t/s
  384. 35.49.143.530 I slot print_timing: id 0 | task 1943 | prompt eval time = 276.96 ms / 105 tokens ( 2.64 ms per token, 379.12 tokens per second)
  385. 35.49.143.536 I slot print_timing: id 0 | task 1943 | eval time = 3112.04 ms / 124 tokens ( 25.10 ms per token, 39.85 tokens per second)
  386. 35.49.143.538 I slot print_timing: id 0 | task 1943 | total time = 3389.00 ms / 229 tokens
  387. 35.49.143.539 I slot print_timing: id 0 | task 1943 | graphs reused = 1704
  388. 35.49.143.540 I slot print_timing: id 0 | task 1943 | draft acceptance = 0.09677 ( 72 accepted / 744 generated)
  389. 35.49.143.542 I slot print_timing: id 0 | task 1943 | adaptive dm: fringe=0.00 n_max=15
  390. 35.49.143.563 I statistics dflash: #calls(b,g,a) = 13 1907 1907, #gen drafts = 1907, #acc drafts = 1432, #gen tokens = 28206, #acc tokens = 4907, dur(b,g,a) = 0.009, 14539.396, 0.576 ms
  391. 35.49.145.533 I slot release: id 0 | task 1943 | stop processing: n_tokens = 47513, truncated = 0
  392. 35.49.145.724 I srv update_slots: all slots are idle
  393. 35.49.580.917 I srv params_from_: Chat format: peg-native
  394. 35.49.638.826 I slot get_availabl: id 0 | task -1 | selected slot by LCP similarity, sim_best = 0.994 (> 0.100 thold), f_keep = 1.000
  395. 35.49.638.833 I slot get_availabl: id 0 | task -1 | adaptive dm: preserving state for continuation (sim=0.994, keep=1.000, n_max=15)
  396. 35.49.639.535 I reasoning-budget: activated, budget=2147483647 tokens
  397. 35.49.639.703 I slot launch_slot_: id 0 | task 1998 | processing task, is_child = 0
  398. 35.49.766.288 I slot create_check: id 0 | task 1998 | created context checkpoint 6 of 32 (pos_min = 47514, pos_max = 47514, n_tokens = 47515, size = 149.821 MiB)
  399. 35.52.630.444 I reasoning-budget: deactivated (natural end)
  400. 35.52.631.895 W slot process_toke: id 0 | task 1998 | raw tool marker observed while lazy grammar is enabled; keeping DFlash governed by active grammar boundary in_reasoning=0 n_decoded=102 reasoning_tokens=100 visible_tokens=2
  401. 35.52.631.904 I slot print_timing: id 0 | task 1998 | n_decoded = 102, tg = 40.28 t/s
  402. 35.54.334.758 I slot print_timing: id 0 | task 1998 | prompt eval time = 458.93 ms / 281 tokens ( 1.63 ms per token, 612.29 tokens per second)
  403. 35.54.334.764 I slot print_timing: id 0 | task 1998 | eval time = 4235.26 ms / 162 tokens ( 26.14 ms per token, 38.25 tokens per second)
  404. 35.54.334.765 I slot print_timing: id 0 | task 1998 | total time = 4694.19 ms / 443 tokens
  405. 35.54.334.766 I slot print_timing: id 0 | task 1998 | graphs reused = 1760
  406. 35.54.334.767 I slot print_timing: id 0 | task 1998 | draft acceptance = 0.09028 ( 92 accepted / 1019 generated)
  407. 35.54.334.769 I slot print_timing: id 0 | task 1998 | adaptive dm: fringe=0.00 n_max=15
  408. 35.54.334.791 I statistics dflash: #calls(b,g,a) = 14 1977 1977, #gen drafts = 1977, #acc drafts = 1476, #gen tokens = 29225, #acc tokens = 4999, dur(b,g,a) = 0.010, 15037.959, 0.601 ms
  409. 35.54.336.762 I slot release: id 0 | task 1998 | stop processing: n_tokens = 47956, truncated = 0
  410. 35.54.336.939 I srv update_slots: all slots are idle
  411. 35.55.102.066 I srv params_from_: Chat format: peg-native
  412. 35.55.158.562 I slot get_availabl: id 0 | task -1 | selected slot by LCP similarity, sim_best = 0.996 (> 0.100 thold), f_keep = 0.999
  413. 35.55.158.567 I slot get_availabl: id 0 | task -1 | adaptive dm: preserving state for continuation (sim=0.996, keep=0.999, n_max=15)
  414. 35.55.159.264 I reasoning-budget: activated, budget=2147483647 tokens
  415. 35.55.159.575 I slot launch_slot_: id 0 | task 2072 | processing task, is_child = 0
  416. 35.55.159.629 I slot update_slots: id 0 | task 2072 | Checking checkpoint with [47514, 47514] against 47896...
  417. 35.55.206.089 W slot update_slots: id 0 | task 2072 | restored context checkpoint (pos_min = 47514, pos_max = 47514, n_tokens = 47515, n_past = 47515, size = 149.821 MiB)
  418. 35.55.207.510 W find_slot: non-consecutive token position 47561 after 47515 for sequence 0 with 47 new tokens
  419. 35.55.207.606 W find_slot: non-consecutive token position 47561 after 47515 for sequence 0 with 47 new tokens
  420. 35.56.332.842 I reasoning-budget: deactivated (natural end)
  421. 35.56.388.381 W slot process_toke: id 0 | task 2072 | raw tool marker observed while lazy grammar is enabled; keeping DFlash governed by active grammar boundary in_reasoning=0 n_decoded=17 reasoning_tokens=15 visible_tokens=2
  422. 35.58.102.761 I slot print_timing: id 0 | task 2072 | prompt eval time = 738.35 ms / 563 tokens ( 1.31 ms per token, 762.51 tokens per second)
  423. 35.58.102.767 I slot print_timing: id 0 | task 2072 | eval time = 2203.25 ms / 61 tokens ( 36.12 ms per token, 27.69 tokens per second)
  424. 35.58.102.768 I slot print_timing: id 0 | task 2072 | total time = 2941.60 ms / 624 tokens
  425. 35.58.102.771 I slot print_timing: id 0 | task 2072 | graphs reused = 1778
  426. 35.58.102.773 I slot print_timing: id 0 | task 2072 | draft acceptance = 0.10270 ( 38 accepted / 370 generated)
  427. 35.58.102.774 I slot print_timing: id 0 | task 2072 | adaptive dm: fringe=0.00 n_max=15
  428. 35.58.102.804 I statistics dflash: #calls(b,g,a) = 15 2002 2002, #gen drafts = 2002, #acc drafts = 1493, #gen tokens = 29595, #acc tokens = 5037, dur(b,g,a) = 0.010, 15218.660, 0.605 ms
  429. 35.58.104.986 I slot release: id 0 | task 2072 | stop processing: n_tokens = 48141, truncated = 0
  430. 35.58.105.144 I srv update_slots: all slots are idle
  431. 35.58.450.034 I srv params_from_: Chat format: peg-native
  432. 35.58.515.126 I slot get_availabl: id 0 | task -1 | selected slot by LCP similarity, sim_best = 0.999 (> 0.100 thold), f_keep = 0.999
  433. 35.58.515.131 I slot get_availabl: id 0 | task -1 | adaptive dm: preserving state for continuation (sim=0.999, keep=0.999, n_max=15)
  434. 35.58.516.298 I reasoning-budget: activated, budget=2147483647 tokens
  435. 35.58.516.478 I slot launch_slot_: id 0 | task 2102 | processing task, is_child = 0
  436. 35.58.516.518 I slot update_slots: id 0 | task 2102 | Checking checkpoint with [47514, 47514] against 48095...
  437. 35.58.564.123 W slot update_slots: id 0 | task 2102 | restored context checkpoint (pos_min = 47514, pos_max = 47514, n_tokens = 47515, n_past = 47515, size = 149.821 MiB)
  438. 35.58.566.325 W find_slot: non-consecutive token position 47647 after 47517 for sequence 0 with 133 new tokens
  439. 35.58.566.521 W find_slot: non-consecutive token position 47647 after 47517 for sequence 0 with 133 new tokens
  440. 36.01.536.027 I slot print_timing: id 0 | task 2102 | n_decoded = 108, tg = 50.92 t/s
  441. 36.01.647.828 I reasoning-budget: deactivated (natural end)
  442. 36.02.771.712 W slot process_toke: id 0 | task 2102 | raw tool marker observed while lazy grammar is enabled; keeping DFlash governed by active grammar boundary in_reasoning=0 n_decoded=180 reasoning_tokens=112 visible_tokens=68
  443. 36.04.588.509 I slot print_timing: id 0 | task 2102 | n_decoded = 223, tg = 43.10 t/s
  444. 36.05.677.516 I slot print_timing: id 0 | task 2102 | prompt eval time = 897.03 ms / 649 tokens ( 1.38 ms per token, 723.50 tokens per second)
  445. 36.05.677.523 I slot print_timing: id 0 | task 2102 | eval time = 6263.08 ms / 274 tokens ( 22.86 ms per token, 43.75 tokens per second)
  446. 36.05.677.524 I slot print_timing: id 0 | task 2102 | total time = 7160.11 ms / 923 tokens
  447. 36.05.677.525 I slot print_timing: id 0 | task 2102 | graphs reused = 1854
  448. 36.05.677.527 I slot print_timing: id 0 | task 2102 | draft acceptance = 0.12896 ( 179 accepted / 1388 generated)
  449. 36.05.677.528 I slot print_timing: id 0 | task 2102 | adaptive dm: fringe=0.00 n_max=15
  450. 36.05.677.552 I statistics dflash: #calls(b,g,a) = 16 2097 2097, #gen drafts = 2097, #acc drafts = 1560, #gen tokens = 30983, #acc tokens = 5216, dur(b,g,a) = 0.010, 15924.873, 0.632 ms
  451. 36.05.679.468 I slot release: id 0 | task 2102 | stop processing: n_tokens = 48438, truncated = 0
  452. 36.05.679.655 I srv update_slots: all slots are idle
  453. 36.06.026.311 I srv params_from_: Chat format: peg-native
  454. 36.06.087.111 I slot get_availabl: id 0 | task -1 | selected slot by LCP similarity, sim_best = 0.997 (> 0.100 thold), f_keep = 0.998
  455. 36.06.087.116 I slot get_availabl: id 0 | task -1 | adaptive dm: preserving state for continuation (sim=0.997, keep=0.998, n_max=15)
  456. 36.06.088.137 I reasoning-budget: activated, budget=2147483647 tokens
  457. 36.06.088.336 I slot launch_slot_: id 0 | task 2202 | processing task, is_child = 0
  458. 36.06.088.403 I slot update_slots: id 0 | task 2202 | Checking checkpoint with [47514, 47514] against 48344...
  459. 36.06.138.059 W slot update_slots: id 0 | task 2202 | restored context checkpoint (pos_min = 47514, pos_max = 47514, n_tokens = 47515, n_past = 47515, size = 149.821 MiB)
  460. 36.06.139.981 W find_slot: non-consecutive token position 47953 after 47515 for sequence 0 with 439 new tokens
  461. 36.06.140.348 W find_slot: non-consecutive token position 47953 after 47515 for sequence 0 with 439 new tokens
  462. 36.07.849.134 I reasoning-budget: deactivated (natural end)
  463. 36.08.126.536 W slot process_toke: id 0 | task 2202 | raw tool marker observed while lazy grammar is enabled; keeping DFlash governed by active grammar boundary in_reasoning=0 n_decoded=31 reasoning_tokens=19 visible_tokens=12
  464. 36.10.060.121 I slot print_timing: id 0 | task 2202 | n_decoded = 100, tg = 36.27 t/s
  465. 36.10.143.681 I slot operator (): id 0 | task 2202 | adaptive dm profit: cur=15 recommended=14 score=46.7 action=apply
  466. 36.10.568.985 I slot print_timing: id 0 | task 2202 | prompt eval time = 1214.08 ms / 955 tokens ( 1.27 ms per token, 786.60 tokens per second)
  467. 36.10.568.992 I slot print_timing: id 0 | task 2202 | eval time = 3265.72 ms / 148 tokens ( 22.07 ms per token, 45.32 tokens per second)
  468. 36.10.568.993 I slot print_timing: id 0 | task 2202 | total time = 4479.80 ms / 1103 tokens
  469. 36.10.568.994 I slot print_timing: id 0 | task 2202 | graphs reused = 1892
  470. 36.10.568.995 I slot print_timing: id 0 | task 2202 | draft acceptance = 0.15478 ( 102 accepted / 659 generated)
  471. 36.10.568.996 I slot print_timing: id 0 | task 2202 | adaptive dm: fringe=0.00 n_max=14
  472. 36.10.569.058 I statistics dflash: #calls(b,g,a) = 17 2143 2143, #gen drafts = 2143, #acc drafts = 1593, #gen tokens = 31642, #acc tokens = 5318, dur(b,g,a) = 0.011, 16260.282, 0.646 ms
  473. 36.10.571.611 I slot release: id 0 | task 2202 | stop processing: n_tokens = 48618, truncated = 0
  474. 36.10.571.784 I srv update_slots: all slots are idle
  475. 36.11.049.923 I srv params_from_: Chat format: peg-native
  476. 36.11.106.037 I slot get_availabl: id 0 | task -1 | selected slot by LCP similarity, sim_best = 0.996 (> 0.100 thold), f_keep = 0.998
  477. 36.11.106.042 I slot get_availabl: id 0 | task -1 | adaptive dm: preserving state for continuation (sim=0.996, keep=0.998, n_max=14)
  478. 36.11.106.744 I reasoning-budget: activated, budget=2147483647 tokens
  479. 36.11.106.935 I slot launch_slot_: id 0 | task 2253 | processing task, is_child = 0
  480. 36.11.106.987 I slot update_slots: id 0 | task 2253 | Checking checkpoint with [47514, 47514] against 48501...
  481. 36.11.152.762 W slot update_slots: id 0 | task 2253 | restored context checkpoint (pos_min = 47514, pos_max = 47514, n_tokens = 47515, n_past = 47515, size = 149.821 MiB)
  482. 36.11.155.020 W find_slot: non-consecutive token position 48026 after 47515 for sequence 0 with 512 new tokens
  483. 36.11.155.512 W find_slot: non-consecutive token position 48026 after 47515 for sequence 0 with 512 new tokens
  484. 36.12.500.998 I slot create_check: id 0 | task 2253 | created context checkpoint 7 of 32 (pos_min = 48618, pos_max = 48618, n_tokens = 48619, size = 240.251 MiB)
  485. 36.13.453.938 I reasoning-budget: deactivated (natural end)
  486. 36.13.626.290 W slot process_toke: id 0 | task 2253 | raw tool marker observed while lazy grammar is enabled; keeping DFlash governed by active grammar boundary in_reasoning=0 n_decoded=46 reasoning_tokens=32 visible_tokens=14
  487. 36.15.185.538 I slot print_timing: id 0 | task 2253 | prompt eval time = 1596.24 ms / 1202 tokens ( 1.33 ms per token, 753.02 tokens per second)
  488. 36.15.185.544 I slot print_timing: id 0 | task 2253 | eval time = 2481.51 ms / 100 tokens ( 24.82 ms per token, 40.30 tokens per second)
  489. 36.15.185.545 I slot print_timing: id 0 | task 2253 | total time = 4077.75 ms / 1302 tokens
  490. 36.15.185.546 I slot print_timing: id 0 | task 2253 | graphs reused = 1925
  491. 36.15.185.547 I slot print_timing: id 0 | task 2253 | draft acceptance = 0.10889 ( 60 accepted / 551 generated)
  492. 36.15.185.549 I slot print_timing: id 0 | task 2253 | adaptive dm: fringe=0.00 n_max=14
  493. 36.15.185.570 I statistics dflash: #calls(b,g,a) = 18 2183 2183, #gen drafts = 2183, #acc drafts = 1623, #gen tokens = 32193, #acc tokens = 5378, dur(b,g,a) = 0.012, 16558.076, 0.657 ms
  494. 36.15.187.490 I slot release: id 0 | task 2253 | stop processing: n_tokens = 48817, truncated = 0
  495. 36.15.187.641 I slot print_timing: id 0 | task -1 | n_decoded = 100, tg = 40.25 t/s
  496. 36.15.187.651 I srv update_slots: all slots are idle
  497. 36.15.519.426 I srv params_from_: Chat format: peg-native
  498. 36.15.575.561 I slot get_availabl: id 0 | task -1 | selected slot by LCP similarity, sim_best = 0.965 (> 0.100 thold), f_keep = 1.000
  499. 36.15.575.567 I slot get_availabl: id 0 | task -1 | adaptive dm: preserving state for continuation (sim=0.965, keep=1.000, n_max=14)
  500. 36.15.576.586 I reasoning-budget: activated, budget=2147483647 tokens
  501. 36.15.576.757 I slot launch_slot_: id 0 | task 2298 | processing task, is_child = 0
  502. 36.17.091.618 I slot create_check: id 0 | task 2298 | created context checkpoint 8 of 32 (pos_min = 50073, pos_max = 50073, n_tokens = 50074, size = 199.236 MiB)
  503. 36.18.609.078 I reasoning-budget: deactivated (natural end)
  504. 36.18.674.181 W slot process_toke: id 0 | task 2298 | raw tool marker observed while lazy grammar is enabled; keeping DFlash governed by active grammar boundary in_reasoning=0 n_decoded=34 reasoning_tokens=32 visible_tokens=2
  505. 36.19.059.721 I slot operator (): id 0 | task 2298 | adaptive dm profit: cur=14 recommended=15 score=42.9 action=apply
  506. 36.21.248.869 I slot print_timing: id 0 | task 2298 | n_decoded = 100, tg = 28.39 t/s
  507. 36.22.695.383 I slot print_timing: id 0 | task 2298 | prompt eval time = 2149.16 ms / 1773 tokens ( 1.21 ms per token, 824.97 tokens per second)
  508. 36.22.695.389 I slot print_timing: id 0 | task 2298 | eval time = 4968.36 ms / 228 tokens ( 21.79 ms per token, 45.89 tokens per second)
  509. 36.22.695.391 I slot print_timing: id 0 | task 2298 | total time = 7117.52 ms / 2001 tokens
  510. 36.22.695.392 I slot print_timing: id 0 | task 2298 | graphs reused = 1987
  511. 36.22.695.393 I slot print_timing: id 0 | task 2298 | draft acceptance = 0.14286 ( 154 accepted / 1078 generated)
  512. 36.22.695.395 I slot print_timing: id 0 | task 2298 | adaptive dm: fringe=0.00 n_max=15
  513. 36.22.695.418 I statistics dflash: #calls(b,g,a) = 19 2258 2258, #gen drafts = 2258, #acc drafts = 1672, #gen tokens = 33271, #acc tokens = 5532, dur(b,g,a) = 0.013, 17115.458, 0.687 ms
  514. 36.22.697.499 I slot release: id 0 | task 2298 | stop processing: n_tokens = 50819, truncated = 0
  515. 36.22.697.675 I srv update_slots: all slots are idle
  516. 36.23.133.710 I srv params_from_: Chat format: peg-native
  517. 36.23.221.392 I slot get_availabl: id 0 | task -1 | selected slot by LCP similarity, sim_best = 0.993 (> 0.100 thold), f_keep = 0.996
  518. 36.23.221.399 I slot get_availabl: id 0 | task -1 | adaptive dm: preserving state for continuation (sim=0.993, keep=0.996, n_max=15)
  519. 36.23.222.136 I reasoning-budget: activated, budget=2147483647 tokens
  520. 36.23.222.320 I slot launch_slot_: id 0 | task 2378 | processing task, is_child = 0
  521. 36.23.222.369 I slot update_slots: id 0 | task 2378 | Checking checkpoint with [50073, 50073] against 50624...
  522. 36.23.290.475 W slot update_slots: id 0 | task 2378 | restored context checkpoint (pos_min = 50073, pos_max = 50073, n_tokens = 50074, n_past = 50074, size = 199.236 MiB)
  523. 36.23.293.464 W find_slot: non-consecutive token position 50475 after 50075 for sequence 0 with 402 new tokens
  524. 36.23.293.772 W find_slot: non-consecutive token position 50475 after 50075 for sequence 0 with 402 new tokens
  525. 36.24.998.554 I reasoning-budget: deactivated (natural end)
  526. 36.25.230.270 W slot process_toke: id 0 | task 2378 | raw tool marker observed while lazy grammar is enabled; keeping DFlash governed by active grammar boundary in_reasoning=0 n_decoded=27 reasoning_tokens=19 visible_tokens=8
  527. 36.27.389.970 I slot print_timing: id 0 | task 2378 | prompt eval time = 1230.58 ms / 918 tokens ( 1.34 ms per token, 745.99 tokens per second)
  528. 36.27.389.977 I slot print_timing: id 0 | task 2378 | eval time = 2935.01 ms / 93 tokens ( 31.56 ms per token, 31.69 tokens per second)
  529. 36.27.389.978 I slot print_timing: id 0 | task 2378 | total time = 4165.59 ms / 1011 tokens
  530. 36.27.389.979 I slot print_timing: id 0 | task 2378 | graphs reused = 2025
  531. 36.27.389.981 I slot print_timing: id 0 | task 2378 | draft acceptance = 0.09167 ( 55 accepted / 600 generated)
  532. 36.27.389.982 I slot print_timing: id 0 | task 2378 | adaptive dm: fringe=0.00 n_max=15
  533. 36.27.390.008 I statistics dflash: #calls(b,g,a) = 20 2298 2298, #gen drafts = 2298, #acc drafts = 1700, #gen tokens = 33871, #acc tokens = 5587, dur(b,g,a) = 0.014, 17409.870, 0.701 ms
  534. 36.27.392.110 I slot release: id 0 | task 2378 | stop processing: n_tokens = 51087, truncated = 0
  535. 36.27.392.319 I srv update_slots: all slots are idle
  536. 36.54.717.954 I srv params_from_: Chat format: peg-native
  537. 36.54.826.858 I slot get_availabl: id 0 | task -1 | selected slot by LCP similarity, sim_best = 0.996 (> 0.100 thold), f_keep = 0.999
  538. 36.54.826.865 I slot get_availabl: id 0 | task -1 | adaptive dm: preserving state for continuation (sim=0.996, keep=0.999, n_max=15)
  539. 36.54.828.223 I reasoning-budget: activated, budget=2147483647 tokens
  540. 36.54.828.452 I slot launch_slot_: id 0 | task 2423 | processing task, is_child = 0
  541. 36.54.828.508 I slot update_slots: id 0 | task 2423 | Checking checkpoint with [50073, 50073] against 51019...
  542. 36.55.032.144 W slot update_slots: id 0 | task 2423 | restored context checkpoint (pos_min = 50073, pos_max = 50073, n_tokens = 50074, n_past = 50074, size = 199.236 MiB)
  543. 36.55.060.101 W find_slot: non-consecutive token position 50585 after 50076 for sequence 0 with 512 new tokens
  544. 36.55.060.841 W find_slot: non-consecutive token position 50585 after 50076 for sequence 0 with 512 new tokens
  545. 36.56.866.105 I slot create_check: id 0 | task 2423 | created context checkpoint 9 of 32 (pos_min = 51239, pos_max = 51239, n_tokens = 51240, size = 249.626 MiB)
  546. 36.57.740.445 I reasoning-budget: deactivated (natural end)
  547. 36.58.098.486 W slot process_toke: id 0 | task 2423 | raw tool marker observed while lazy grammar is enabled; keeping DFlash governed by active grammar boundary in_reasoning=0 n_decoded=45 reasoning_tokens=30 visible_tokens=15
  548. 37.00.430.114 I slot print_timing: id 0 | task 2423 | n_decoded = 100, tg = 28.63 t/s
  549. 37.00.608.862 I slot print_timing: id 0 | task 2423 | prompt eval time = 2106.93 ms / 1170 tokens ( 1.80 ms per token, 555.31 tokens per second)
  550. 37.00.608.896 I slot print_timing: id 0 | task 2423 | eval time = 3670.73 ms / 105 tokens ( 34.96 ms per token, 28.60 tokens per second)
  551. 37.00.608.897 I slot print_timing: id 0 | task 2423 | total time = 5777.66 ms / 1275 tokens
  552. 37.00.608.899 I slot print_timing: id 0 | task 2423 | graphs reused = 2069
  553. 37.00.608.901 I slot print_timing: id 0 | task 2423 | draft acceptance = 0.08551 ( 59 accepted / 690 generated)
  554. 37.00.608.902 I slot print_timing: id 0 | task 2423 | adaptive dm: fringe=0.00 n_max=15
  555. 37.00.608.928 I statistics dflash: #calls(b,g,a) = 21 2344 2344, #gen drafts = 2344, #acc drafts = 1731, #gen tokens = 34561, #acc tokens = 5646, dur(b,g,a) = 0.014, 17772.205, 0.715 ms
  556. 37.00.613.130 I slot release: id 0 | task 2423 | stop processing: n_tokens = 51349, truncated = 0
  557. 37.00.613.360 I srv update_slots: all slots are idle
  558. 37.02.399.979 I srv params_from_: Chat format: peg-native
  559. 37.02.507.517 I slot get_availabl: id 0 | task -1 | selected slot by LCP similarity, sim_best = 0.997 (> 0.100 thold), f_keep = 0.999
  560. 37.02.507.524 I slot get_availabl: id 0 | task -1 | adaptive dm: preserving state for continuation (sim=0.997, keep=0.999, n_max=15)
  561. 37.02.508.649 I reasoning-budget: activated, budget=2147483647 tokens
  562. 37.02.508.982 I slot launch_slot_: id 0 | task 2474 | processing task, is_child = 0
  563. 37.02.509.044 I slot update_slots: id 0 | task 2474 | Checking checkpoint with [51239, 51239] against 51289...
  564. 37.02.607.274 W slot update_slots: id 0 | task 2474 | restored context checkpoint (pos_min = 51239, pos_max = 51239, n_tokens = 51240, n_past = 51240, size = 249.626 MiB)
  565. 37.02.612.458 W find_slot: non-consecutive token position 51349 after 51240 for sequence 0 with 110 new tokens
  566. 37.02.612.651 W find_slot: non-consecutive token position 51349 after 51240 for sequence 0 with 110 new tokens
  567. 37.03.317.390 I reasoning-budget: deactivated (natural end)
  568. 37.03.496.832 W slot process_toke: id 0 | task 2474 | raw tool marker observed while lazy grammar is enabled; keeping DFlash governed by active grammar boundary in_reasoning=0 n_decoded=20 reasoning_tokens=12 visible_tokens=8
  569. 37.05.068.733 I slot print_timing: id 0 | task 2474 | prompt eval time = 489.75 ms / 202 tokens ( 2.42 ms per token, 412.46 tokens per second)
  570. 37.05.068.740 I slot print_timing: id 0 | task 2474 | eval time = 2068.83 ms / 85 tokens ( 24.34 ms per token, 41.09 tokens per second)
  571. 37.05.068.742 I slot print_timing: id 0 | task 2474 | total time = 2558.57 ms / 287 tokens
  572. 37.05.068.743 I slot print_timing: id 0 | task 2474 | graphs reused = 2098
  573. 37.05.068.745 I slot print_timing: id 0 | task 2474 | draft acceptance = 0.11828 ( 55 accepted / 465 generated)
  574. 37.05.068.746 I slot print_timing: id 0 | task 2474 | adaptive dm: fringe=0.00 n_max=15
  575. 37.05.068.768 I statistics dflash: #calls(b,g,a) = 22 2375 2375, #gen drafts = 2375, #acc drafts = 1758, #gen tokens = 35026, #acc tokens = 5701, dur(b,g,a) = 0.014, 18027.425, 0.732 ms
  576. 37.05.072.744 I slot release: id 0 | task 2474 | stop processing: n_tokens = 51528, truncated = 0
  577. 37.05.072.952 I srv update_slots: all slots are idle
  578. 37.35.209.777 I srv params_from_: Chat format: peg-native
  579. 37.35.325.550 I slot get_availabl: id 0 | task -1 | selected slot by LCP similarity, sim_best = 0.898 (> 0.100 thold), f_keep = 0.999
  580. 37.35.325.561 I slot get_availabl: id 0 | task -1 | adaptive dm: reset state for prompt change (sim=0.898, keep=0.999, preserve_policy=0, n_max=15)
  581. 37.35.326.650 I reasoning-budget: activated, budget=2147483647 tokens
  582. 37.35.326.891 I slot launch_slot_: id 0 | task 2509 | processing task, is_child = 0
  583. 37.35.326.977 I slot update_slots: id 0 | task 2509 | Checking checkpoint with [51239, 51239] against 51462...
  584. 37.35.527.791 W slot update_slots: id 0 | task 2509 | restored context checkpoint (pos_min = 51239, pos_max = 51239, n_tokens = 51240, n_past = 51240, size = 249.626 MiB)
  585. 37.35.564.054 W find_slot: non-consecutive token position 51527 after 51241 for sequence 0 with 288 new tokens
  586. 37.35.564.443 W find_slot: non-consecutive token position 51527 after 51241 for sequence 0 with 288 new tokens
  587. 37.38.534.791 I slot print_timing: id 0 | task 2509 | prompt processing, n_tokens = 2336, progress = 0.93, t = 3.21 s / 728.23 tokens per second
  588. 37.40.969.926 I slot print_timing: id 0 | task 2509 | prompt processing, n_tokens = 4384, progress = 0.97, t = 5.64 s / 776.89 tokens per second
  589. 37.42.396.582 I slot print_timing: id 0 | task 2509 | prompt processing, n_tokens = 5565, progress = 0.99, t = 7.07 s / 787.17 tokens per second
  590. 37.42.486.864 I slot create_check: id 0 | task 2509 | created context checkpoint 10 of 32 (pos_min = 56804, pos_max = 56804, n_tokens = 56805, size = 249.626 MiB)
  591. 37.43.083.750 I slot print_timing: id 0 | task 2509 | prompt processing, n_tokens = 6077, progress = 1.00, t = 7.76 s / 783.44 tokens per second
  592. 37.45.661.544 I slot print_timing: id 0 | task 2509 | n_decoded = 103, tg = 41.12 t/s
  593. 37.48.672.828 I slot print_timing: id 0 | task 2509 | n_decoded = 194, tg = 35.15 t/s
  594. 37.49.620.912 I reasoning-budget: deactivated (natural end)
  595. 37.51.596.322 W slot process_toke: id 0 | task 2509 | raw tool marker observed while lazy grammar is enabled; keeping DFlash governed by active grammar boundary in_reasoning=0 n_decoded=323 reasoning_tokens=232 visible_tokens=91
  596. 37.51.763.869 I slot print_timing: id 0 | task 2509 | n_decoded = 325, tg = 37.75 t/s
  597. 37.54.788.654 I slot print_timing: id 0 | task 2509 | n_decoded = 431, tg = 37.05 t/s
  598. 37.55.481.644 I slot print_timing: id 0 | task 2509 | prompt eval time = 7826.33 ms / 6081 tokens ( 1.29 ms per token, 776.99 tokens per second)
  599. 37.55.481.651 I slot print_timing: id 0 | task 2509 | eval time = 12325.95 ms / 461 tokens ( 26.74 ms per token, 37.40 tokens per second)
  600. 37.55.481.652 I slot print_timing: id 0 | task 2509 | total time = 20152.28 ms / 6542 tokens
  601. 37.55.481.654 I slot print_timing: id 0 | task 2509 | graphs reused = 2245
  602. 37.55.481.655 I slot print_timing: id 0 | task 2509 | draft acceptance = 0.10634 ( 280 accepted / 2633 generated)
  603. 37.55.481.657 I slot print_timing: id 0 | task 2509 | adaptive dm: fringe=0.00 n_max=15
  604. 37.55.481.679 I statistics dflash: #calls(b,g,a) = 23 2555 2555, #gen drafts = 2555, #acc drafts = 1876, #gen tokens = 37659, #acc tokens = 5981, dur(b,g,a) = 0.015, 19457.483, 0.790 ms
  605. 37.55.484.542 I slot release: id 0 | task 2509 | stop processing: n_tokens = 57781, truncated = 0
  606. 37.55.484.722 I srv update_slots: all slots are idle
  607. 37.58.245.153 I srv params_from_: Chat format: peg-native
  608. 37.58.318.888 I slot get_availabl: id 0 | task -1 | selected slot by LCP similarity, sim_best = 0.997 (> 0.100 thold), f_keep = 1.000
  609. 37.58.318.894 I slot get_availabl: id 0 | task -1 | adaptive dm: preserving state for continuation (sim=0.997, keep=1.000, n_max=15)
  610. 37.58.319.822 I reasoning-budget: activated, budget=2147483647 tokens
  611. 37.58.320.043 I slot launch_slot_: id 0 | task 2696 | processing task, is_child = 0
  612. 37.58.707.397 I slot create_check: id 0 | task 2696 | created context checkpoint 11 of 32 (pos_min = 57940, pos_max = 57940, n_tokens = 57941, size = 165.251 MiB)
  613. 37.59.773.370 I reasoning-budget: deactivated (natural end)
  614. 38.00.667.606 W slot process_toke: id 0 | task 2696 | raw tool marker observed while lazy grammar is enabled; keeping DFlash governed by active grammar boundary in_reasoning=0 n_decoded=81 reasoning_tokens=39 visible_tokens=42
  615. 38.01.831.596 I slot print_timing: id 0 | task 2696 | n_decoded = 100, tg = 32.83 t/s
  616. 38.03.792.738 I slot print_timing: id 0 | task 2696 | prompt eval time = 464.63 ms / 164 tokens ( 2.83 ms per token, 352.97 tokens per second)
  617. 38.03.792.744 I slot print_timing: id 0 | task 2696 | eval time = 5005.52 ms / 156 tokens ( 32.09 ms per token, 31.17 tokens per second)
  618. 38.03.792.746 I slot print_timing: id 0 | task 2696 | total time = 5470.14 ms / 320 tokens
  619. 38.03.792.747 I slot print_timing: id 0 | task 2696 | graphs reused = 2302
  620. 38.03.792.749 I slot print_timing: id 0 | task 2696 | draft acceptance = 0.08358 ( 86 accepted / 1029 generated)
  621. 38.03.792.750 I slot print_timing: id 0 | task 2696 | adaptive dm: fringe=0.00 n_max=15
  622. 38.03.792.778 I statistics dflash: #calls(b,g,a) = 24 2625 2625, #gen drafts = 2625, #acc drafts = 1918, #gen tokens = 38688, #acc tokens = 6067, dur(b,g,a) = 0.016, 19963.710, 0.803 ms
  623. 38.03.795.393 I slot release: id 0 | task 2696 | stop processing: n_tokens = 58101, truncated = 0
  624. 38.03.795.589 I srv update_slots: all slots are idle
  625. 38.52.377.296 I srv params_from_: Chat format: peg-native
  626. 38.52.480.522 I slot get_availabl: id 0 | task -1 | selected slot by LCP similarity, sim_best = 0.686 (> 0.100 thold), f_keep = 0.676
  627. 38.52.480.529 I slot get_availabl: id 0 | task -1 | adaptive dm: reset state for prompt change (sim=0.686, keep=0.676, preserve_policy=0, n_max=15)
  628. 38.52.481.597 I reasoning-budget: activated, budget=2147483647 tokens
  629. 38.52.481.838 I slot launch_slot_: id 0 | task 2770 | processing task, is_child = 0
  630. 38.52.481.891 I slot update_slots: id 0 | task 2770 | Checking checkpoint with [57940, 57940] against 39272...
  631. 38.52.481.893 I slot update_slots: id 0 | task 2770 | Checking checkpoint with [56804, 56804] against 39272...
  632. 38.52.481.893 I slot update_slots: id 0 | task 2770 | Checking checkpoint with [51239, 51239] against 39272...
  633. 38.52.481.895 I slot update_slots: id 0 | task 2770 | Checking checkpoint with [50073, 50073] against 39272...
  634. 38.52.481.895 I slot update_slots: id 0 | task 2770 | Checking checkpoint with [48618, 48618] against 39272...
  635. 38.52.481.896 I slot update_slots: id 0 | task 2770 | Checking checkpoint with [47514, 47514] against 39272...
  636. 38.52.481.897 I slot update_slots: id 0 | task 2770 | Checking checkpoint with [46371, 46371] against 39272...
  637. 38.52.481.899 I slot update_slots: id 0 | task 2770 | Checking checkpoint with [39258, 39258] against 39272...
  638. 38.52.715.348 W slot update_slots: id 0 | task 2770 | restored context checkpoint (pos_min = 39258, pos_max = 39258, n_tokens = 39259, n_past = 39259, size = 249.626 MiB)
  639. 38.52.715.356 W slot update_slots: id 0 | task 2770 | erased invalidated context checkpoint (pos_min = 46371, pos_max = 46371, n_tokens = 46372, n_swa = 0, pos_next = 39259, size = 199.236 MiB)
  640. 38.52.723.472 W slot update_slots: id 0 | task 2770 | erased invalidated context checkpoint (pos_min = 47514, pos_max = 47514, n_tokens = 47515, n_swa = 0, pos_next = 39259, size = 149.821 MiB)
  641. 38.52.729.425 W slot update_slots: id 0 | task 2770 | erased invalidated context checkpoint (pos_min = 48618, pos_max = 48618, n_tokens = 48619, n_swa = 0, pos_next = 39259, size = 240.251 MiB)
  642. 38.52.739.086 W slot update_slots: id 0 | task 2770 | erased invalidated context checkpoint (pos_min = 50073, pos_max = 50073, n_tokens = 50074, n_swa = 0, pos_next = 39259, size = 199.236 MiB)
  643. 38.52.746.922 W slot update_slots: id 0 | task 2770 | erased invalidated context checkpoint (pos_min = 51239, pos_max = 51239, n_tokens = 51240, n_swa = 0, pos_next = 39259, size = 249.626 MiB)
  644. 38.52.762.123 W slot update_slots: id 0 | task 2770 | erased invalidated context checkpoint (pos_min = 56804, pos_max = 56804, n_tokens = 56805, n_swa = 0, pos_next = 39259, size = 249.626 MiB)
  645. 38.52.770.744 W slot update_slots: id 0 | task 2770 | erased invalidated context checkpoint (pos_min = 57940, pos_max = 57940, n_tokens = 57941, n_swa = 0, pos_next = 39259, size = 165.251 MiB)
  646. 38.52.791.492 W find_slot: non-consecutive token position 39770 after 39259 for sequence 0 with 512 new tokens
  647. 38.52.792.705 W find_slot: non-consecutive token position 39770 after 39259 for sequence 0 with 512 new tokens
  648. 38.57.255.405 I slot print_timing: id 0 | task 2770 | prompt processing, n_tokens = 4096, progress = 0.76, t = 4.77 s / 858.06 tokens per second
  649. 38.59.383.132 I slot print_timing: id 0 | task 2770 | prompt processing, n_tokens = 6144, progress = 0.79, t = 6.90 s / 890.27 tokens per second
  650. 39.01.556.047 I slot print_timing: id 0 | task 2770 | prompt processing, n_tokens = 8192, progress = 0.83, t = 9.07 s / 902.78 tokens per second
  651. 39.03.775.005 I slot print_timing: id 0 | task 2770 | prompt processing, n_tokens = 10240, progress = 0.86, t = 11.29 s / 906.75 tokens per second
  652. 39.06.023.810 I slot print_timing: id 0 | task 2770 | prompt processing, n_tokens = 12288, progress = 0.90, t = 13.54 s / 907.40 tokens per second
  653. 39.08.303.965 I slot print_timing: id 0 | task 2770 | prompt processing, n_tokens = 14336, progress = 0.94, t = 15.82 s / 906.07 tokens per second
  654. 39.10.666.135 I slot print_timing: id 0 | task 2770 | prompt processing, n_tokens = 16384, progress = 0.97, t = 18.18 s / 901.00 tokens per second
  655. 39.11.980.913 I slot print_timing: id 0 | task 2770 | prompt processing, n_tokens = 17469, progress = 0.99, t = 19.50 s / 895.89 tokens per second
  656. 39.12.512.834 I slot print_timing: id 0 | task 2770 | prompt processing, n_tokens = 17897, progress = 1.00, t = 20.03 s / 893.47 tokens per second
  657. 39.12.596.228 I slot create_check: id 0 | task 2770 | created context checkpoint 5 of 32 (pos_min = 57155, pos_max = 57155, n_tokens = 57156, size = 249.626 MiB)
  658. 39.12.719.520 I slot print_timing: id 0 | task 2770 | prompt processing, n_tokens = 17981, progress = 1.00, t = 20.24 s / 888.49 tokens per second
  659. 39.15.392.964 I slot print_timing: id 0 | task 2770 | n_decoded = 102, tg = 39.22 t/s
  660. 39.16.082.528 I reasoning-budget: deactivated (natural end)
  661. 39.17.373.545 W slot process_toke: id 0 | task 2770 | raw tool marker observed while lazy grammar is enabled; keeping DFlash governed by active grammar boundary in_reasoning=0 n_decoded=206 reasoning_tokens=132 visible_tokens=74
  662. 39.18.428.021 I slot print_timing: id 0 | task 2770 | n_decoded = 241, tg = 42.75 t/s
  663. 39.20.928.102 I slot print_timing: id 0 | task 2770 | prompt eval time = 20308.06 ms / 17985 tokens ( 1.13 ms per token, 885.61 tokens per second)
  664. 39.20.928.109 I slot print_timing: id 0 | task 2770 | eval time = 8134.96 ms / 365 tokens ( 22.29 ms per token, 44.87 tokens per second)
  665. 39.20.928.110 I slot print_timing: id 0 | task 2770 | total time = 28443.02 ms / 18350 tokens
  666. 39.20.928.111 I slot print_timing: id 0 | task 2770 | graphs reused = 2401
  667. 39.20.928.113 I slot print_timing: id 0 | task 2770 | draft acceptance = 0.13568 ( 243 accepted / 1791 generated)
  668. 39.20.928.114 I slot print_timing: id 0 | task 2770 | adaptive dm: fringe=0.00 n_max=15
  669. 39.20.928.138 I statistics dflash: #calls(b,g,a) = 25 2747 2747, #gen drafts = 2747, #acc drafts = 2007, #gen tokens = 40479, #acc tokens = 6310, dur(b,g,a) = 0.017, 20891.784, 0.838 ms
  670. 39.20.931.471 I slot release: id 0 | task 2770 | stop processing: n_tokens = 57609, truncated = 0
  671. 39.20.931.647 I srv update_slots: all slots are idle
  672. 39.23.629.502 I srv params_from_: Chat format: peg-native
  673. 39.23.698.508 I slot get_availabl: id 0 | task -1 | selected slot by LCP similarity, sim_best = 0.994 (> 0.100 thold), f_keep = 0.997
  674. 39.23.698.513 I slot get_availabl: id 0 | task -1 | adaptive dm: preserving state for continuation (sim=0.994, keep=0.997, n_max=15)
  675. 39.23.699.265 I reasoning-budget: activated, budget=2147483647 tokens
  676. 39.23.699.438 I slot launch_slot_: id 0 | task 2905 | processing task, is_child = 0
  677. 39.23.699.483 I slot update_slots: id 0 | task 2905 | Checking checkpoint with [57155, 57155] against 57450...
  678. 39.23.783.250 W slot update_slots: id 0 | task 2905 | restored context checkpoint (pos_min = 57155, pos_max = 57155, n_tokens = 57156, n_past = 57156, size = 249.626 MiB)
  679. 39.23.788.370 W find_slot: non-consecutive token position 57270 after 57156 for sequence 0 with 115 new tokens
  680. 39.23.788.509 W find_slot: non-consecutive token position 57270 after 57156 for sequence 0 with 115 new tokens
  681. 39.25.630.868 I reasoning-budget: deactivated (natural end)
  682. 39.25.694.465 W slot process_toke: id 0 | task 2905 | raw tool marker observed while lazy grammar is enabled; keeping DFlash governed by active grammar boundary in_reasoning=0 n_decoded=35 reasoning_tokens=33 visible_tokens=2
  683. 39.27.210.247 I slot print_timing: id 0 | task 2905 | prompt eval time = 1080.64 ms / 631 tokens ( 1.71 ms per token, 583.92 tokens per second)
  684. 39.27.210.253 I slot print_timing: id 0 | task 2905 | eval time = 2429.22 ms / 88 tokens ( 27.60 ms per token, 36.23 tokens per second)
  685. 39.27.210.255 I slot print_timing: id 0 | task 2905 | total time = 3509.86 ms / 719 tokens
  686. 39.27.210.258 I slot print_timing: id 0 | task 2905 | graphs reused = 2429
  687. 39.27.210.259 I slot print_timing: id 0 | task 2905 | draft acceptance = 0.09462 ( 51 accepted / 539 generated)
  688. 39.27.210.261 I slot print_timing: id 0 | task 2905 | adaptive dm: fringe=0.00 n_max=15
  689. 39.27.210.282 I statistics dflash: #calls(b,g,a) = 26 2784 2784, #gen drafts = 2784, #acc drafts = 2028, #gen tokens = 41018, #acc tokens = 6361, dur(b,g,a) = 0.018, 21167.438, 0.852 ms
  690. 39.27.212.534 I slot release: id 0 | task 2905 | stop processing: n_tokens = 57875, truncated = 0
  691. 39.27.212.713 I srv update_slots: all slots are idle
  692. 39.28.666.008 I srv params_from_: Chat format: peg-native
  693. 39.28.740.338 I slot get_availabl: id 0 | task -1 | selected slot by LCP similarity, sim_best = 0.998 (> 0.100 thold), f_keep = 0.999
  694. 39.28.740.344 I slot get_availabl: id 0 | task -1 | adaptive dm: preserving state for continuation (sim=0.998, keep=0.999, n_max=15)
  695. 39.28.741.236 I reasoning-budget: activated, budget=2147483647 tokens
  696. 39.28.741.480 I slot launch_slot_: id 0 | task 2947 | processing task, is_child = 0
  697. 39.28.741.536 I slot update_slots: id 0 | task 2947 | Checking checkpoint with [57155, 57155] against 57822...
  698. 39.28.825.829 W slot update_slots: id 0 | task 2947 | restored context checkpoint (pos_min = 57155, pos_max = 57155, n_tokens = 57156, n_past = 57156, size = 249.626 MiB)
  699. 39.28.830.597 W find_slot: non-consecutive token position 57427 after 57156 for sequence 0 with 272 new tokens
  700. 39.28.830.829 W find_slot: non-consecutive token position 57427 after 57156 for sequence 0 with 272 new tokens
  701. 39.30.640.716 I reasoning-budget: deactivated (natural end)
  702. 39.31.280.570 W slot process_toke: id 0 | task 2947 | raw tool marker observed while lazy grammar is enabled; keeping DFlash governed by active grammar boundary in_reasoning=0 n_decoded=48 reasoning_tokens=27 visible_tokens=21
  703. 39.32.984.852 I slot print_timing: id 0 | task 2947 | prompt eval time = 1052.48 ms / 788 tokens ( 1.34 ms per token, 748.70 tokens per second)
  704. 39.32.984.858 I slot print_timing: id 0 | task 2947 | eval time = 3190.00 ms / 100 tokens ( 31.90 ms per token, 31.35 tokens per second)
  705. 39.32.984.860 I slot print_timing: id 0 | task 2947 | total time = 4242.49 ms / 888 tokens
  706. 39.32.984.861 I slot print_timing: id 0 | task 2947 | graphs reused = 2465
  707. 39.32.984.862 I slot print_timing: id 0 | task 2947 | draft acceptance = 0.08511 ( 56 accepted / 658 generated)
  708. 39.32.984.864 I slot print_timing: id 0 | task 2947 | adaptive dm: fringe=0.00 n_max=15
  709. 39.32.984.885 I statistics dflash: #calls(b,g,a) = 27 2829 2829, #gen drafts = 2829, #acc drafts = 2053, #gen tokens = 41676, #acc tokens = 6417, dur(b,g,a) = 0.019, 21504.106, 0.863 ms
  710. 39.32.987.335 I slot release: id 0 | task 2947 | stop processing: n_tokens = 58045, truncated = 0
  711. 39.32.987.541 I slot print_timing: id 0 | task -1 | n_decoded = 100, tg = 31.32 t/s
  712. 39.32.987.555 I srv update_slots: all slots are idle
  713. 39.35.466.575 I srv params_from_: Chat format: peg-native
  714. 39.35.538.643 I slot get_availabl: id 0 | task -1 | selected slot by LCP similarity, sim_best = 0.996 (> 0.100 thold), f_keep = 1.000
  715. 39.35.538.649 I slot get_availabl: id 0 | task -1 | adaptive dm: preserving state for continuation (sim=0.996, keep=1.000, n_max=15)
  716. 39.35.539.885 I reasoning-budget: activated, budget=2147483647 tokens
  717. 39.35.540.070 I slot launch_slot_: id 0 | task 2997 | processing task, is_child = 0
  718. 39.35.988.178 I slot create_check: id 0 | task 2997 | created context checkpoint 6 of 32 (pos_min = 58251, pos_max = 58251, n_tokens = 58252, size = 169.841 MiB)
  719. 39.37.811.678 I reasoning-budget: deactivated (natural end)
  720. 39.38.251.238 W slot process_toke: id 0 | task 2997 | raw tool marker observed while lazy grammar is enabled; keeping DFlash governed by active grammar boundary in_reasoning=0 n_decoded=74 reasoning_tokens=58 visible_tokens=16
  721. 39.39.410.584 I slot print_timing: id 0 | task 2997 | n_decoded = 102, tg = 30.39 t/s
  722. 39.40.634.763 I slot print_timing: id 0 | task 2997 | prompt eval time = 511.87 ms / 211 tokens ( 2.43 ms per token, 412.21 tokens per second)
  723. 39.40.634.769 I slot print_timing: id 0 | task 2997 | eval time = 4580.71 ms / 147 tokens ( 31.16 ms per token, 32.09 tokens per second)
  724. 39.40.634.770 I slot print_timing: id 0 | task 2997 | total time = 5092.58 ms / 358 tokens
  725. 39.40.634.771 I slot print_timing: id 0 | task 2997 | graphs reused = 2523
  726. 39.40.634.773 I slot print_timing: id 0 | task 2997 | draft acceptance = 0.07534 ( 77 accepted / 1022 generated)
  727. 39.40.634.774 I slot print_timing: id 0 | task 2997 | adaptive dm: fringe=0.00 n_max=15
  728. 39.40.634.796 I statistics dflash: #calls(b,g,a) = 28 2899 2899, #gen drafts = 2899, #acc drafts = 2091, #gen tokens = 42698, #acc tokens = 6494, dur(b,g,a) = 0.019, 21988.852, 0.886 ms
  729. 39.40.637.216 I slot release: id 0 | task 2997 | stop processing: n_tokens = 58403, truncated = 0
  730. 39.40.637.400 I srv update_slots: all slots are idle
  731. 40.06.066.462 I srv params_from_: Chat format: peg-native
  732. 40.06.165.397 I slot get_availabl: id 0 | task -1 | selected slot by LCP similarity, sim_best = 0.996 (> 0.100 thold), f_keep = 0.999
  733. 40.06.165.403 I slot get_availabl: id 0 | task -1 | adaptive dm: preserving state for continuation (sim=0.996, keep=0.999, n_max=15)
  734. 40.06.166.267 I reasoning-budget: activated, budget=2147483647 tokens
  735. 40.06.166.499 I slot launch_slot_: id 0 | task 3070 | processing task, is_child = 0
  736. 40.06.166.570 I slot update_slots: id 0 | task 3070 | Checking checkpoint with [58251, 58251] against 58330...
  737. 40.06.350.155 W slot update_slots: id 0 | task 3070 | restored context checkpoint (pos_min = 58251, pos_max = 58251, n_tokens = 58252, n_past = 58252, size = 169.841 MiB)
  738. 40.06.371.836 W find_slot: non-consecutive token position 58403 after 58252 for sequence 0 with 152 new tokens
  739. 40.06.372.059 W find_slot: non-consecutive token position 58403 after 58252 for sequence 0 with 152 new tokens
  740. 40.10.542.147 I slot print_timing: id 0 | task 3070 | n_decoded = 100, tg = 30.17 t/s
  741. 40.11.249.029 I reasoning-budget: deactivated (natural end)
  742. 40.11.629.346 W slot process_toke: id 0 | task 3070 | raw tool marker observed while lazy grammar is enabled; keeping DFlash governed by active grammar boundary in_reasoning=0 n_decoded=136 reasoning_tokens=120 visible_tokens=16
  743. 40.13.584.211 I slot print_timing: id 0 | task 3070 | n_decoded = 181, tg = 28.47 t/s
  744. 40.14.758.178 I slot print_timing: id 0 | task 3070 | prompt eval time = 1058.12 ms / 324 tokens ( 3.27 ms per token, 306.20 tokens per second)
  745. 40.14.758.185 I slot print_timing: id 0 | task 3070 | eval time = 7531.88 ms / 227 tokens ( 33.18 ms per token, 30.14 tokens per second)
  746. 40.14.758.187 I slot print_timing: id 0 | task 3070 | total time = 8589.99 ms / 551 tokens
  747. 40.14.758.188 I slot print_timing: id 0 | task 3070 | graphs reused = 2604
  748. 40.14.758.190 I slot print_timing: id 0 | task 3070 | draft acceptance = 0.08755 ( 128 accepted / 1462 generated)
  749. 40.14.758.191 I slot print_timing: id 0 | task 3070 | adaptive dm: fringe=0.00 n_max=15
  750. 40.14.758.214 I statistics dflash: #calls(b,g,a) = 29 2999 2999, #gen drafts = 2999, #acc drafts = 2155, #gen tokens = 44160, #acc tokens = 6622, dur(b,g,a) = 0.020, 22821.085, 0.936 ms
  751. 40.14.761.626 I slot release: id 0 | task 3070 | stop processing: n_tokens = 58804, truncated = 0
  752. 40.14.761.930 I srv update_slots: all slots are idle
  753. 40.21.296.053 I srv params_from_: Chat format: peg-native
  754. 40.21.426.852 I slot get_availabl: id 0 | task -1 | selected slot by LCP similarity, sim_best = 0.922 (> 0.100 thold), f_keep = 0.998
  755. 40.21.426.859 I slot get_availabl: id 0 | task -1 | adaptive dm: preserving state for continuation (sim=0.922, keep=0.998, n_max=15)
  756. 40.21.428.464 I reasoning-budget: activated, budget=2147483647 tokens
  757. 40.21.428.704 I slot launch_slot_: id 0 | task 3174 | processing task, is_child = 0
  758. 40.21.428.764 I slot update_slots: id 0 | task 3174 | Checking checkpoint with [58251, 58251] against 58712...
  759. 40.21.497.488 W slot update_slots: id 0 | task 3174 | restored context checkpoint (pos_min = 58251, pos_max = 58251, n_tokens = 58252, n_past = 58252, size = 169.841 MiB)
  760. 40.21.501.120 W find_slot: non-consecutive token position 58763 after 58253 for sequence 0 with 512 new tokens
  761. 40.21.501.757 W find_slot: non-consecutive token position 58763 after 58253 for sequence 0 with 512 new tokens
  762. 40.24.978.981 I slot print_timing: id 0 | task 3174 | prompt processing, n_tokens = 2600, progress = 0.96, t = 3.55 s / 732.34 tokens per second
  763. 40.27.496.802 I slot print_timing: id 0 | task 3174 | prompt processing, n_tokens = 4648, progress = 0.99, t = 6.07 s / 765.98 tokens per second
  764. 40.27.810.946 I slot print_timing: id 0 | task 3174 | prompt processing, n_tokens = 4920, progress = 0.99, t = 6.38 s / 770.89 tokens per second
  765. 40.27.900.753 I slot create_check: id 0 | task 3174 | created context checkpoint 7 of 32 (pos_min = 63171, pos_max = 63171, n_tokens = 63172, size = 219.450 MiB)
  766. 40.28.548.559 I slot print_timing: id 0 | task 3174 | prompt processing, n_tokens = 5432, progress = 1.00, t = 7.12 s / 762.94 tokens per second
  767. 40.30.347.945 I reasoning-budget: deactivated (natural end)
  768. 40.30.413.063 W slot process_toke: id 0 | task 3174 | raw tool marker observed while lazy grammar is enabled; keeping DFlash governed by active grammar boundary in_reasoning=0 n_decoded=55 reasoning_tokens=53 visible_tokens=2
  769. 40.32.300.032 I slot print_timing: id 0 | task 3174 | n_decoded = 102, tg = 27.76 t/s
  770. 40.33.577.546 I slot print_timing: id 0 | task 3174 | prompt eval time = 7195.76 ms / 5436 tokens ( 1.32 ms per token, 755.45 tokens per second)
  771. 40.33.577.554 I slot print_timing: id 0 | task 3174 | eval time = 4950.36 ms / 131 tokens ( 37.79 ms per token, 26.46 tokens per second)
  772. 40.33.577.555 I slot print_timing: id 0 | task 3174 | total time = 12146.11 ms / 5567 tokens
  773. 40.33.577.557 I slot print_timing: id 0 | task 3174 | graphs reused = 2655
  774. 40.33.577.559 I slot print_timing: id 0 | task 3174 | draft acceptance = 0.07634 ( 70 accepted / 917 generated)
  775. 40.33.577.560 I slot print_timing: id 0 | task 3174 | adaptive dm: fringe=0.00 n_max=15
  776. 40.33.577.586 I statistics dflash: #calls(b,g,a) = 30 3062 3062, #gen drafts = 3062, #acc drafts = 2191, #gen tokens = 45077, #acc tokens = 6692, dur(b,g,a) = 0.021, 23325.640, 0.964 ms
  777. 40.33.581.669 I slot release: id 0 | task 3174 | stop processing: n_tokens = 63821, truncated = 0
  778. 40.33.582.020 I srv update_slots: all slots are idle
  779. 40.35.323.798 I srv params_from_: Chat format: peg-native
  780. 40.35.454.268 I slot get_availabl: id 0 | task -1 | selected slot by LCP similarity, sim_best = 0.997 (> 0.100 thold), f_keep = 0.999
  781. 40.35.454.274 I slot get_availabl: id 0 | task -1 | adaptive dm: preserving state for continuation (sim=0.997, keep=0.999, n_max=15)
  782. 40.35.455.406 I reasoning-budget: activated, budget=2147483647 tokens
  783. 40.35.455.683 I slot launch_slot_: id 0 | task 3244 | processing task, is_child = 0
  784. 40.35.455.762 I slot update_slots: id 0 | task 3244 | Checking checkpoint with [63171, 63171] against 63743...
  785. 40.35.544.617 W slot update_slots: id 0 | task 3244 | restored context checkpoint (pos_min = 63171, pos_max = 63171, n_tokens = 63172, n_past = 63172, size = 219.450 MiB)
  786. 40.35.549.191 W find_slot: non-consecutive token position 63396 after 63174 for sequence 0 with 225 new tokens
  787. 40.35.549.532 W find_slot: non-consecutive token position 63396 after 63174 for sequence 0 with 225 new tokens
  788. 40.39.653.912 I reasoning-budget: deactivated (natural end)
  789. 40.39.734.383 W slot process_toke: id 0 | task 3244 | raw tool marker observed while lazy grammar is enabled; keeping DFlash governed by active grammar boundary in_reasoning=0 n_decoded=89 reasoning_tokens=87 visible_tokens=2
  790. 40.40.576.517 I slot print_timing: id 0 | task 3244 | n_decoded = 100, tg = 26.32 t/s
  791. 40.42.188.112 I slot print_timing: id 0 | task 3244 | prompt eval time = 1320.23 ms / 741 tokens ( 1.78 ms per token, 561.26 tokens per second)
  792. 40.42.188.117 I slot print_timing: id 0 | task 3244 | eval time = 5410.89 ms / 156 tokens ( 34.69 ms per token, 28.83 tokens per second)
  793. 40.42.188.118 I slot print_timing: id 0 | task 3244 | total time = 6731.13 ms / 897 tokens
  794. 40.42.188.119 I slot print_timing: id 0 | task 3244 | graphs reused = 2709
  795. 40.42.188.121 I slot print_timing: id 0 | task 3244 | draft acceptance = 0.09109 ( 90 accepted / 988 generated)
  796. 40.42.188.122 I slot print_timing: id 0 | task 3244 | adaptive dm: fringe=0.00 n_max=15
  797. 40.42.188.145 I statistics dflash: #calls(b,g,a) = 31 3130 3130, #gen drafts = 3130, #acc drafts = 2229, #gen tokens = 46065, #acc tokens = 6782, dur(b,g,a) = 0.021, 23862.765, 0.989 ms
  798. 40.42.190.557 I slot release: id 0 | task 3244 | stop processing: n_tokens = 64071, truncated = 0
  799. 40.42.190.692 I srv update_slots: all slots are idle
  800. 40.43.727.252 I srv params_from_: Chat format: peg-native
  801. 40.43.803.647 I slot get_availabl: id 0 | task -1 | selected slot by LCP similarity, sim_best = 0.998 (> 0.100 thold), f_keep = 0.999
  802. 40.43.803.654 I slot get_availabl: id 0 | task -1 | adaptive dm: preserving state for continuation (sim=0.998, keep=0.999, n_max=15)
  803. 40.43.804.500 I reasoning-budget: activated, budget=2147483647 tokens
  804. 40.43.804.743 I slot launch_slot_: id 0 | task 3317 | processing task, is_child = 0
  805. 40.43.804.800 I slot update_slots: id 0 | task 3317 | Checking checkpoint with [63171, 63171] against 64002...
  806. 40.43.875.883 W slot update_slots: id 0 | task 3317 | restored context checkpoint (pos_min = 63171, pos_max = 63171, n_tokens = 63172, n_past = 63172, size = 219.450 MiB)
  807. 40.43.880.662 W find_slot: non-consecutive token position 63643 after 63174 for sequence 0 with 472 new tokens
  808. 40.43.881.121 W find_slot: non-consecutive token position 63643 after 63174 for sequence 0 with 472 new tokens
  809. 40.46.679.068 I reasoning-budget: deactivated (natural end)
  810. 40.46.680.482 W slot process_toke: id 0 | task 3317 | raw tool marker observed while lazy grammar is enabled; keeping DFlash governed by active grammar boundary in_reasoning=0 n_decoded=54 reasoning_tokens=52 visible_tokens=2
  811. 40.48.150.444 I slot print_timing: id 0 | task 3317 | n_decoded = 101, tg = 35.77 t/s
  812. 40.48.591.955 I slot print_timing: id 0 | task 3317 | prompt eval time = 1519.20 ms / 988 tokens ( 1.54 ms per token, 650.34 tokens per second)
  813. 40.48.591.963 I slot print_timing: id 0 | task 3317 | eval time = 3266.72 ms / 113 tokens ( 28.91 ms per token, 34.59 tokens per second)
  814. 40.48.591.965 I slot print_timing: id 0 | task 3317 | total time = 4785.92 ms / 1101 tokens
  815. 40.48.591.966 I slot print_timing: id 0 | task 3317 | graphs reused = 2750
  816. 40.48.591.968 I slot print_timing: id 0 | task 3317 | draft acceptance = 0.09753 ( 67 accepted / 687 generated)
  817. 40.48.591.969 I slot print_timing: id 0 | task 3317 | adaptive dm: fringe=0.00 n_max=15
  818. 40.48.591.993 I statistics dflash: #calls(b,g,a) = 32 3176 3176, #gen drafts = 3176, #acc drafts = 2260, #gen tokens = 46752, #acc tokens = 6849, dur(b,g,a) = 0.022, 24203.298, 1.005 ms
  819. 40.48.595.570 I slot release: id 0 | task 3317 | stop processing: n_tokens = 64273, truncated = 0
  820. 40.48.595.763 I srv update_slots: all slots are idle
  821. 40.50.144.763 I srv params_from_: Chat format: peg-native
  822. 40.50.227.860 I slot get_availabl: id 0 | task -1 | selected slot by LCP similarity, sim_best = 0.998 (> 0.100 thold), f_keep = 0.999
  823. 40.50.227.866 I slot get_availabl: id 0 | task -1 | adaptive dm: preserving state for continuation (sim=0.998, keep=0.999, n_max=15)
  824. 40.50.228.796 I reasoning-budget: activated, budget=2147483647 tokens
  825. 40.50.228.984 I slot launch_slot_: id 0 | task 3368 | processing task, is_child = 0
  826. 40.50.229.058 I slot update_slots: id 0 | task 3368 | Checking checkpoint with [63171, 63171] against 64214...
  827. 40.50.301.063 W slot update_slots: id 0 | task 3368 | restored context checkpoint (pos_min = 63171, pos_max = 63171, n_tokens = 63172, n_past = 63172, size = 219.450 MiB)
  828. 40.50.304.973 W find_slot: non-consecutive token position 63683 after 63172 for sequence 0 with 512 new tokens
  829. 40.50.305.433 W find_slot: non-consecutive token position 63683 after 63172 for sequence 0 with 512 new tokens
  830. 40.51.821.131 I slot create_check: id 0 | task 3368 | created context checkpoint 8 of 32 (pos_min = 64273, pos_max = 64273, n_tokens = 64274, size = 249.626 MiB)
  831. 40.53.230.035 I reasoning-budget: deactivated (natural end)
  832. 40.53.769.554 W slot process_toke: id 0 | task 3368 | raw tool marker observed while lazy grammar is enabled; keeping DFlash governed by active grammar boundary in_reasoning=0 n_decoded=74 reasoning_tokens=42 visible_tokens=32
  833. 40.54.804.669 I slot print_timing: id 0 | task 3368 | n_decoded = 100, tg = 35.78 t/s
  834. 40.57.543.206 I slot print_timing: id 0 | task 3368 | prompt eval time = 1779.54 ms / 1184 tokens ( 1.50 ms per token, 665.34 tokens per second)
  835. 40.57.543.213 I slot print_timing: id 0 | task 3368 | eval time = 5531.49 ms / 193 tokens ( 28.66 ms per token, 34.89 tokens per second)
  836. 40.57.543.214 I slot print_timing: id 0 | task 3368 | total time = 7311.02 ms / 1377 tokens
  837. 40.57.543.215 I slot print_timing: id 0 | task 3368 | graphs reused = 2828
  838. 40.57.543.217 I slot print_timing: id 0 | task 3368 | draft acceptance = 0.09383 ( 114 accepted / 1215 generated)
  839. 40.57.543.218 I slot print_timing: id 0 | task 3368 | adaptive dm: fringe=0.00 n_max=15
  840. 40.57.543.243 I statistics dflash: #calls(b,g,a) = 33 3257 3257, #gen drafts = 3257, #acc drafts = 2312, #gen tokens = 47967, #acc tokens = 6963, dur(b,g,a) = 0.022, 24809.748, 1.026 ms
  841. 40.57.547.497 I slot release: id 0 | task 3368 | stop processing: n_tokens = 64551, truncated = 0
  842. 40.57.547.772 I srv update_slots: all slots are idle
  843. 40.58.000.371 I srv params_from_: Chat format: peg-native
  844. 40.58.079.896 I slot get_availabl: id 0 | task -1 | selected slot by LCP similarity, sim_best = 0.998 (> 0.100 thold), f_keep = 0.998
  845. 40.58.079.901 I slot get_availabl: id 0 | task -1 | adaptive dm: preserving state for continuation (sim=0.998, keep=0.998, n_max=15)
  846. 40.58.080.606 I reasoning-budget: activated, budget=2147483647 tokens
  847. 40.58.080.762 I slot launch_slot_: id 0 | task 3454 | processing task, is_child = 0
  848. 40.58.080.801 I slot update_slots: id 0 | task 3454 | Checking checkpoint with [64273, 64273] against 64430...
  849. 40.58.170.381 W slot update_slots: id 0 | task 3454 | restored context checkpoint (pos_min = 64273, pos_max = 64273, n_tokens = 64274, n_past = 64274, size = 249.626 MiB)
  850. 40.58.175.107 W find_slot: non-consecutive token position 64546 after 64276 for sequence 0 with 273 new tokens
  851. 40.58.175.485 W find_slot: non-consecutive token position 64546 after 64276 for sequence 0 with 273 new tokens
  852. 40.59.114.381 I reasoning-budget: deactivated (natural end)
  853. 40.59.176.949 W slot process_toke: id 0 | task 3454 | raw tool marker observed while lazy grammar is enabled; keeping DFlash governed by active grammar boundary in_reasoning=0 n_decoded=17 reasoning_tokens=15 visible_tokens=2
  854. 41.01.876.210 I slot print_timing: id 0 | task 3454 | n_decoded = 104, tg = 32.18 t/s
  855. 41.02.963.122 I slot print_timing: id 0 | task 3454 | prompt eval time = 560.67 ms / 304 tokens ( 1.84 ms per token, 542.21 tokens per second)
  856. 41.02.963.128 I slot print_timing: id 0 | task 3454 | eval time = 4320.71 ms / 186 tokens ( 23.23 ms per token, 43.05 tokens per second)
  857. 41.02.963.130 I slot print_timing: id 0 | task 3454 | total time = 4881.39 ms / 490 tokens
  858. 41.02.963.131 I slot print_timing: id 0 | task 3454 | graphs reused = 2876
  859. 41.02.963.133 I slot print_timing: id 0 | task 3454 | draft acceptance = 0.17908 ( 137 accepted / 765 generated)
  860. 41.02.963.134 I slot print_timing: id 0 | task 3454 | adaptive dm: fringe=0.00 n_max=15
  861. 41.02.963.160 I statistics dflash: #calls(b,g,a) = 34 3308 3308, #gen drafts = 3308, #acc drafts = 2342, #gen tokens = 48732, #acc tokens = 7100, dur(b,g,a) = 0.023, 25235.964, 1.047 ms
  862. 41.02.965.743 I slot release: id 0 | task 3454 | stop processing: n_tokens = 64766, truncated = 0
  863. 41.02.965.947 I srv update_slots: all slots are idle
  864. 41.03.582.219 I srv params_from_: Chat format: peg-native
  865. 41.03.678.272 I slot get_availabl: id 0 | task -1 | selected slot by LCP similarity, sim_best = 0.996 (> 0.100 thold), f_keep = 0.997
  866. 41.03.678.279 I slot get_availabl: id 0 | task -1 | adaptive dm: preserving state for continuation (sim=0.996, keep=0.997, n_max=15)
  867. 41.03.679.280 I reasoning-budget: activated, budget=2147483647 tokens
  868. 41.03.679.471 I slot launch_slot_: id 0 | task 3509 | processing task, is_child = 0
  869. 41.03.679.526 I slot update_slots: id 0 | task 3509 | Checking checkpoint with [64273, 64273] against 64595...
  870. 41.03.768.232 W slot update_slots: id 0 | task 3509 | restored context checkpoint (pos_min = 64273, pos_max = 64273, n_tokens = 64274, n_past = 64274, size = 249.626 MiB)
  871. 41.03.772.740 W find_slot: non-consecutive token position 64351 after 64276 for sequence 0 with 78 new tokens
  872. 41.03.772.863 W find_slot: non-consecutive token position 64351 after 64276 for sequence 0 with 78 new tokens
  873. 41.07.190.930 I slot print_timing: id 0 | task 3509 | n_decoded = 100, tg = 40.67 t/s
  874. 41.07.626.413 I reasoning-budget: deactivated (natural end)
  875. 41.07.824.987 W slot process_toke: id 0 | task 3509 | raw tool marker observed while lazy grammar is enabled; keeping DFlash governed by active grammar boundary in_reasoning=0 n_decoded=126 reasoning_tokens=115 visible_tokens=11
  876. 41.10.194.621 I slot print_timing: id 0 | task 3509 | prompt eval time = 1051.25 ms / 594 tokens ( 1.77 ms per token, 565.04 tokens per second)
  877. 41.10.194.627 I slot print_timing: id 0 | task 3509 | eval time = 5461.90 ms / 189 tokens ( 28.90 ms per token, 34.60 tokens per second)
  878. 41.10.194.630 I slot print_timing: id 0 | task 3509 | total time = 6513.15 ms / 783 tokens
  879. 41.10.194.631 I slot print_timing: id 0 | task 3509 | graphs reused = 2947
  880. 41.10.194.633 I slot print_timing: id 0 | task 3509 | draft acceptance = 0.10541 ( 117 accepted / 1110 generated)
  881. 41.10.194.634 I slot print_timing: id 0 | task 3509 | adaptive dm: fringe=0.00 n_max=15
  882. 41.10.194.661 I statistics dflash: #calls(b,g,a) = 35 3382 3382, #gen drafts = 3382, #acc drafts = 2399, #gen tokens = 49842, #acc tokens = 7217, dur(b,g,a) = 0.023, 25804.518, 1.073 ms
  883. 41.10.197.817 I slot release: id 0 | task 3509 | stop processing: n_tokens = 65059, truncated = 0
  884. 41.10.197.988 I slot print_timing: id 0 | task -1 | n_decoded = 189, tg = 34.57 t/s
  885. 41.10.198.002 I srv update_slots: all slots are idle
  886. 41.35.773.445 I srv params_from_: Chat format: peg-native
  887. 41.35.856.156 I slot get_availabl: id 0 | task -1 | selected slot by LCP similarity, sim_best = 0.987 (> 0.100 thold), f_keep = 0.999
  888. 41.35.856.161 I slot get_availabl: id 0 | task -1 | adaptive dm: preserving state for continuation (sim=0.987, keep=0.999, n_max=15)
  889. 41.35.856.994 I reasoning-budget: activated, budget=2147483647 tokens
  890. 41.35.857.187 I slot launch_slot_: id 0 | task 3588 | processing task, is_child = 0
  891. 41.35.857.236 I slot update_slots: id 0 | task 3588 | Checking checkpoint with [64273, 64273] against 64994...
  892. 41.36.095.968 W slot update_slots: id 0 | task 3588 | restored context checkpoint (pos_min = 64273, pos_max = 64273, n_tokens = 64274, n_past = 64274, size = 249.626 MiB)
  893. 41.36.130.805 W find_slot: non-consecutive token position 64785 after 64276 for sequence 0 with 512 new tokens
  894. 41.36.131.370 W find_slot: non-consecutive token position 64785 after 64276 for sequence 0 with 512 new tokens
  895. 41.37.743.401 I slot create_check: id 0 | task 3588 | created context checkpoint 9 of 32 (pos_min = 65344, pos_max = 65344, n_tokens = 65345, size = 249.626 MiB)
  896. 41.40.406.929 I reasoning-budget: deactivated (natural end)
  897. 41.40.653.797 I slot print_timing: id 0 | task 3588 | n_decoded = 101, tg = 46.25 t/s
  898. 41.41.276.939 W slot process_toke: id 0 | task 3588 | raw tool marker observed while lazy grammar is enabled; keeping DFlash governed by active grammar boundary in_reasoning=0 n_decoded=136 reasoning_tokens=86 visible_tokens=50
  899. 41.43.689.460 I slot print_timing: id 0 | task 3588 | prompt eval time = 2609.76 ms / 1587 tokens ( 1.64 ms per token, 608.10 tokens per second)
  900. 41.43.689.465 I slot print_timing: id 0 | task 3588 | eval time = 5221.63 ms / 202 tokens ( 25.85 ms per token, 38.69 tokens per second)
  901. 41.43.689.466 I slot print_timing: id 0 | task 3588 | total time = 7831.40 ms / 1789 tokens
  902. 41.43.689.467 I slot print_timing: id 0 | task 3588 | graphs reused = 3026
  903. 41.43.689.468 I slot print_timing: id 0 | task 3588 | draft acceptance = 0.09675 ( 119 accepted / 1230 generated)
  904. 41.43.689.469 I slot print_timing: id 0 | task 3588 | adaptive dm: fringe=0.00 n_max=15
  905. 41.43.689.488 I statistics dflash: #calls(b,g,a) = 36 3464 3464, #gen drafts = 3464, #acc drafts = 2449, #gen tokens = 51072, #acc tokens = 7336, dur(b,g,a) = 0.023, 26414.774, 1.089 ms
  906. 41.43.692.100 I slot release: id 0 | task 3588 | stop processing: n_tokens = 66062, truncated = 0
  907. 41.43.692.226 I slot print_timing: id 0 | task -1 | n_decoded = 202, tg = 38.66 t/s
  908. 41.43.692.234 I srv update_slots: all slots are idle
  909. 41.44.155.020 I srv params_from_: Chat format: peg-native
  910. 41.44.244.598 I slot get_availabl: id 0 | task -1 | selected slot by LCP similarity, sim_best = 0.995 (> 0.100 thold), f_keep = 1.000
  911. 41.44.244.605 I slot get_availabl: id 0 | task -1 | adaptive dm: preserving state for continuation (sim=0.995, keep=1.000, n_max=15)
  912. 41.44.245.497 I reasoning-budget: activated, budget=2147483647 tokens
  913. 41.44.245.676 I slot launch_slot_: id 0 | task 3675 | processing task, is_child = 0
  914. 41.44.865.608 I slot create_check: id 0 | task 3675 | created context checkpoint 10 of 32 (pos_min = 66389, pos_max = 66389, n_tokens = 66390, size = 181.657 MiB)
  915. 41.46.883.262 I slot print_timing: id 0 | task 3675 | n_decoded = 100, tg = 51.24 t/s
  916. 41.49.942.220 I slot print_timing: id 0 | task 3675 | n_decoded = 237, tg = 47.30 t/s
  917. 41.50.189.633 I reasoning-budget: deactivated (natural end)
  918. 41.51.181.934 W slot process_toke: id 0 | task 3675 | raw tool marker observed while lazy grammar is enabled; keeping DFlash governed by active grammar boundary in_reasoning=0 n_decoded=283 reasoning_tokens=245 visible_tokens=38
  919. 41.52.976.929 I slot print_timing: id 0 | task 3675 | n_decoded = 323, tg = 40.15 t/s
  920. 41.56.038.445 I slot print_timing: id 0 | task 3675 | n_decoded = 498, tg = 44.83 t/s
  921. 41.56.596.439 I slot print_timing: id 0 | task 3675 | prompt eval time = 684.32 ms / 332 tokens ( 2.06 ms per token, 485.15 tokens per second)
  922. 41.56.596.446 I slot print_timing: id 0 | task 3675 | eval time = 11664.38 ms / 517 tokens ( 22.56 ms per token, 44.32 tokens per second)
  923. 41.56.596.447 I slot print_timing: id 0 | task 3675 | total time = 12348.70 ms / 849 tokens
  924. 41.56.596.448 I slot print_timing: id 0 | task 3675 | graphs reused = 3206
  925. 41.56.596.450 I slot print_timing: id 0 | task 3675 | draft acceptance = 0.12029 ( 332 accepted / 2760 generated)
  926. 41.56.596.451 I slot print_timing: id 0 | task 3675 | adaptive dm: fringe=0.00 n_max=15
  927. 41.56.596.476 I statistics dflash: #calls(b,g,a) = 37 3648 3648, #gen drafts = 3648, #acc drafts = 2559, #gen tokens = 53832, #acc tokens = 7668, dur(b,g,a) = 0.024, 27741.969, 1.147 ms
  928. 41.56.599.841 I slot release: id 0 | task 3675 | stop processing: n_tokens = 66910, truncated = 0
  929. 41.56.600.043 I srv update_slots: all slots are idle
  930. 41.57.611.166 I srv params_from_: Chat format: peg-native
  931. 41.57.738.414 I slot get_availabl: id 0 | task -1 | selected slot by LCP similarity, sim_best = 0.997 (> 0.100 thold), f_keep = 1.000
  932. 41.57.738.421 I slot get_availabl: id 0 | task -1 | adaptive dm: preserving state for continuation (sim=0.997, keep=1.000, n_max=15)
  933. 41.57.739.731 I reasoning-budget: activated, budget=2147483647 tokens
  934. 41.57.739.957 I slot launch_slot_: id 0 | task 3863 | processing task, is_child = 0
  935. 41.58.616.424 I reasoning-budget: deactivated (natural end)
  936. 41.58.680.057 W slot process_toke: id 0 | task 3863 | raw tool marker observed while lazy grammar is enabled; keeping DFlash governed by active grammar boundary in_reasoning=0 n_decoded=14 reasoning_tokens=12 visible_tokens=2
  937. 41.59.845.241 I slot operator (): id 0 | task 3863 | adaptive dm profit: cur=0 recommended=15 score=40.4 action=apply
  938. 42.00.989.325 I slot print_timing: id 0 | task 3863 | prompt eval time = 468.79 ms / 219 tokens ( 2.14 ms per token, 467.16 tokens per second)
  939. 42.00.989.332 I slot print_timing: id 0 | task 3863 | eval time = 2779.70 ms / 55 tokens ( 50.54 ms per token, 19.79 tokens per second)
  940. 42.00.989.333 I slot print_timing: id 0 | task 3863 | total time = 3248.49 ms / 274 tokens
  941. 42.00.989.334 I slot print_timing: id 0 | task 3863 | graphs reused = 3231
  942. 42.00.989.336 I slot print_timing: id 0 | task 3863 | draft acceptance = 0.06667 ( 28 accepted / 420 generated)
  943. 42.00.989.338 I slot print_timing: id 0 | task 3863 | adaptive dm: fringe=0.00 n_max=15
  944. 42.00.989.363 I statistics dflash: #calls(b,g,a) = 38 3676 3676, #gen drafts = 3676, #acc drafts = 2578, #gen tokens = 54252, #acc tokens = 7696, dur(b,g,a) = 0.024, 27938.746, 1.157 ms
  945. 42.00.992.007 I slot release: id 0 | task 3863 | stop processing: n_tokens = 67186, truncated = 0
  946. 42.00.992.202 I srv update_slots: all slots are idle
  947. 42.01.523.439 I srv params_from_: Chat format: peg-native
  948. 42.01.605.011 I slot get_availabl: id 0 | task -1 | selected slot by LCP similarity, sim_best = 0.995 (> 0.100 thold), f_keep = 0.999
  949. 42.01.605.017 I slot get_availabl: id 0 | task -1 | adaptive dm: preserving state for continuation (sim=0.995, keep=0.999, n_max=15)
  950. 42.01.605.863 I reasoning-budget: activated, budget=2147483647 tokens
  951. 42.01.606.040 I slot launch_slot_: id 0 | task 3896 | processing task, is_child = 0
  952. 42.01.606.088 I slot update_slots: id 0 | task 3896 | Checking checkpoint with [66389, 66389] against 67143...
  953. 42.01.668.133 W slot update_slots: id 0 | task 3896 | restored context checkpoint (pos_min = 66389, pos_max = 66389, n_tokens = 66390, n_past = 66390, size = 181.657 MiB)
  954. 42.01.670.672 W find_slot: non-consecutive token position 66901 after 66392 for sequence 0 with 512 new tokens
  955. 42.01.671.104 W find_slot: non-consecutive token position 66901 after 66392 for sequence 0 with 512 new tokens
  956. 42.03.171.995 I slot create_check: id 0 | task 3896 | created context checkpoint 11 of 32 (pos_min = 67502, pos_max = 67502, n_tokens = 67503, size = 249.626 MiB)
  957. 42.04.147.347 I reasoning-budget: deactivated (natural end)
  958. 42.04.210.497 W slot process_toke: id 0 | task 3896 | raw tool marker observed while lazy grammar is enabled; keeping DFlash governed by active grammar boundary in_reasoning=0 n_decoded=94 reasoning_tokens=92 visible_tokens=2
  959. 42.04.547.750 I slot print_timing: id 0 | task 3896 | n_decoded = 100, tg = 76.30 t/s
  960. 42.06.084.010 I slot print_timing: id 0 | task 3896 | prompt eval time = 1630.25 ms / 1117 tokens ( 1.46 ms per token, 685.17 tokens per second)
  961. 42.06.084.017 I slot print_timing: id 0 | task 3896 | eval time = 2846.81 ms / 171 tokens ( 16.65 ms per token, 60.07 tokens per second)
  962. 42.06.084.018 I slot print_timing: id 0 | task 3896 | total time = 4477.06 ms / 1288 tokens
  963. 42.06.084.019 I slot print_timing: id 0 | task 3896 | graphs reused = 3269
  964. 42.06.084.021 I slot print_timing: id 0 | task 3896 | draft acceptance = 0.21463 ( 132 accepted / 615 generated)
  965. 42.06.084.022 I slot print_timing: id 0 | task 3896 | adaptive dm: fringe=0.00 n_max=15
  966. 42.06.084.046 I statistics dflash: #calls(b,g,a) = 39 3717 3717, #gen drafts = 3717, #acc drafts = 2611, #gen tokens = 54867, #acc tokens = 7828, dur(b,g,a) = 0.025, 28242.649, 1.168 ms
  967. 42.06.086.773 I slot release: id 0 | task 3896 | stop processing: n_tokens = 67680, truncated = 0
  968. 42.06.086.961 I srv update_slots: all slots are idle
  969. 42.06.535.783 I srv params_from_: Chat format: peg-native
  970. 42.06.612.408 I slot get_availabl: id 0 | task -1 | selected slot by LCP similarity, sim_best = 0.998 (> 0.100 thold), f_keep = 0.999
  971. 42.06.612.414 I slot get_availabl: id 0 | task -1 | adaptive dm: preserving state for continuation (sim=0.998, keep=0.999, n_max=15)
  972. 42.06.613.123 I reasoning-budget: activated, budget=2147483647 tokens
  973. 42.06.613.291 I slot launch_slot_: id 0 | task 3942 | processing task, is_child = 0
  974. 42.06.613.341 I slot update_slots: id 0 | task 3942 | Checking checkpoint with [67502, 67502] against 67601...
  975. 42.06.698.710 W slot update_slots: id 0 | task 3942 | restored context checkpoint (pos_min = 67502, pos_max = 67502, n_tokens = 67503, n_past = 67503, size = 249.626 MiB)
  976. 42.06.703.252 W find_slot: non-consecutive token position 67678 after 67505 for sequence 0 with 176 new tokens
  977. 42.06.703.430 W find_slot: non-consecutive token position 67678 after 67505 for sequence 0 with 176 new tokens
  978. 42.07.996.501 I reasoning-budget: deactivated (natural end)
  979. 42.08.058.497 W slot process_toke: id 0 | task 3942 | raw tool marker observed while lazy grammar is enabled; keeping DFlash governed by active grammar boundary in_reasoning=0 n_decoded=42 reasoning_tokens=40 visible_tokens=2
  980. 42.10.122.825 I slot print_timing: id 0 | task 3942 | n_decoded = 101, tg = 33.63 t/s
  981. 42.13.182.987 I slot print_timing: id 0 | task 3942 | n_decoded = 353, tg = 58.22 t/s
  982. 42.13.656.478 I slot print_timing: id 0 | task 3942 | prompt eval time = 505.23 ms / 207 tokens ( 2.44 ms per token, 409.71 tokens per second)
  983. 42.13.656.485 I slot print_timing: id 0 | task 3942 | eval time = 6536.73 ms / 389 tokens ( 16.80 ms per token, 59.51 tokens per second)
  984. 42.13.656.487 I slot print_timing: id 0 | task 3942 | total time = 7041.96 ms / 596 tokens
  985. 42.13.656.488 I slot print_timing: id 0 | task 3942 | graphs reused = 3364
  986. 42.13.656.490 I slot print_timing: id 0 | task 3942 | draft acceptance = 0.19529 ( 290 accepted / 1485 generated)
  987. 42.13.656.491 I slot print_timing: id 0 | task 3942 | adaptive dm: fringe=0.00 n_max=15
  988. 42.13.656.516 I statistics dflash: #calls(b,g,a) = 40 3816 3816, #gen drafts = 3816, #acc drafts = 2687, #gen tokens = 56352, #acc tokens = 8118, dur(b,g,a) = 0.025, 28975.181, 1.193 ms
  989. 42.13.659.447 I slot release: id 0 | task 3942 | stop processing: n_tokens = 68099, truncated = 0
  990. 42.13.659.662 I srv update_slots: all slots are idle
  991. 42.14.486.529 I srv params_from_: Chat format: peg-native
  992. 42.14.564.645 I slot get_availabl: id 0 | task -1 | selected slot by LCP similarity, sim_best = 0.970 (> 0.100 thold), f_keep = 0.995
  993. 42.14.564.652 I slot get_availabl: id 0 | task -1 | adaptive dm: preserving state for continuation (sim=0.970, keep=0.995, n_max=15)
  994. 42.14.565.398 I reasoning-budget: activated, budget=2147483647 tokens
  995. 42.14.565.596 I slot launch_slot_: id 0 | task 4045 | processing task, is_child = 0
  996. 42.14.565.654 I slot update_slots: id 0 | task 4045 | Checking checkpoint with [67502, 67502] against 67752...
  997. 42.14.649.674 W slot update_slots: id 0 | task 4045 | restored context checkpoint (pos_min = 67502, pos_max = 67502, n_tokens = 67503, n_past = 67503, size = 249.626 MiB)
  998. 42.14.654.079 W find_slot: non-consecutive token position 68014 after 67503 for sequence 0 with 512 new tokens
  999. 42.14.654.524 W find_slot: non-consecutive token position 68014 after 67503 for sequence 0 with 512 new tokens
  1000. 42.17.135.901 I slot create_check: id 0 | task 4045 | created context checkpoint 12 of 32 (pos_min = 69329, pos_max = 69329, n_tokens = 69330, size = 249.626 MiB)
  1001. 42.17.811.025 I slot print_timing: id 0 | task 4045 | prompt processing, n_tokens = 2339, progress = 1.00, t = 3.25 s / 720.71 tokens per second
  1002. 42.19.112.205 I reasoning-budget: deactivated (natural end)
  1003. 42.19.113.674 W slot process_toke: id 0 | task 4045 | raw tool marker observed while lazy grammar is enabled; keeping DFlash governed by active grammar boundary in_reasoning=0 n_decoded=86 reasoning_tokens=84 visible_tokens=2
  1004. 42.19.829.331 I slot print_timing: id 0 | task 4045 | n_decoded = 100, tg = 51.42 t/s
  1005. 42.21.154.565 I slot print_timing: id 0 | task 4045 | prompt eval time = 3317.89 ms / 2343 tokens ( 1.42 ms per token, 706.17 tokens per second)
  1006. 42.21.154.571 I slot print_timing: id 0 | task 4045 | eval time = 3270.15 ms / 163 tokens ( 20.06 ms per token, 49.84 tokens per second)
  1007. 42.21.154.572 I slot print_timing: id 0 | task 4045 | total time = 6588.04 ms / 2506 tokens
  1008. 42.21.154.573 I slot print_timing: id 0 | task 4045 | graphs reused = 3408
  1009. 42.21.154.575 I slot print_timing: id 0 | task 4045 | draft acceptance = 0.16738 ( 118 accepted / 705 generated)
  1010. 42.21.154.576 I slot print_timing: id 0 | task 4045 | adaptive dm: fringe=0.00 n_max=15
  1011. 42.21.154.598 I statistics dflash: #calls(b,g,a) = 41 3863 3863, #gen drafts = 3863, #acc drafts = 2724, #gen tokens = 57057, #acc tokens = 8236, dur(b,g,a) = 0.025, 29316.837, 1.205 ms
  1012. 42.21.157.068 I slot release: id 0 | task 4045 | stop processing: n_tokens = 70011, truncated = 0
  1013. 42.21.157.263 I srv update_slots: all slots are idle
  1014. 42.21.568.405 I srv params_from_: Chat format: peg-native
  1015. 42.21.666.981 I slot get_availabl: id 0 | task -1 | selected slot by LCP similarity, sim_best = 0.998 (> 0.100 thold), f_keep = 0.999
  1016. 42.21.666.987 I slot get_availabl: id 0 | task -1 | adaptive dm: preserving state for continuation (sim=0.998, keep=0.999, n_max=15)
  1017. 42.21.668.136 I reasoning-budget: activated, budget=2147483647 tokens
  1018. 42.21.668.352 I slot launch_slot_: id 0 | task 4097 | processing task, is_child = 0
  1019. 42.21.668.402 I slot update_slots: id 0 | task 4097 | Checking checkpoint with [69329, 69329] against 69932...
  1020. 42.21.753.550 W slot update_slots: id 0 | task 4097 | restored context checkpoint (pos_min = 69329, pos_max = 69329, n_tokens = 69330, n_past = 69330, size = 249.626 MiB)
  1021. 42.21.758.158 W find_slot: non-consecutive token position 69524 after 69332 for sequence 0 with 195 new tokens
  1022. 42.21.758.383 W find_slot: non-consecutive token position 69524 after 69332 for sequence 0 with 195 new tokens
  1023. 42.23.624.376 I reasoning-budget: deactivated (natural end)
  1024. 42.23.625.740 W slot process_toke: id 0 | task 4097 | raw tool marker observed while lazy grammar is enabled; keeping DFlash governed by active grammar boundary in_reasoning=0 n_decoded=36 reasoning_tokens=34 visible_tokens=2
  1025. 42.24.576.417 I slot print_timing: id 0 | task 4097 | n_decoded = 101, tg = 57.84 t/s
  1026. 42.25.036.345 I slot print_timing: id 0 | task 4097 | prompt eval time = 1160.80 ms / 711 tokens ( 1.63 ms per token, 612.51 tokens per second)
  1027. 42.25.036.351 I slot print_timing: id 0 | task 4097 | eval time = 2204.25 ms / 113 tokens ( 19.51 ms per token, 51.26 tokens per second)
  1028. 42.25.036.353 I slot print_timing: id 0 | task 4097 | total time = 3365.05 ms / 824 tokens
  1029. 42.25.036.354 I slot print_timing: id 0 | task 4097 | graphs reused = 3431
  1030. 42.25.036.356 I slot print_timing: id 0 | task 4097 | draft acceptance = 0.22821 ( 89 accepted / 390 generated)
  1031. 42.25.036.358 I slot print_timing: id 0 | task 4097 | adaptive dm: fringe=0.00 n_max=15
  1032. 42.25.036.380 I statistics dflash: #calls(b,g,a) = 42 3889 3889, #gen drafts = 3889, #acc drafts = 2747, #gen tokens = 57447, #acc tokens = 8325, dur(b,g,a) = 0.026, 29508.455, 1.214 ms
  1033. 42.25.038.832 I slot release: id 0 | task 4097 | stop processing: n_tokens = 70156, truncated = 0
  1034. 42.25.039.043 I srv update_slots: all slots are idle
  1035. 42.25.467.128 I srv params_from_: Chat format: peg-native
  1036. 42.25.549.641 I slot get_availabl: id 0 | task -1 | selected slot by LCP similarity, sim_best = 0.998 (> 0.100 thold), f_keep = 0.999
  1037. 42.25.549.648 I slot get_availabl: id 0 | task -1 | adaptive dm: preserving state for continuation (sim=0.998, keep=0.999, n_max=15)
  1038. 42.25.550.376 I reasoning-budget: activated, budget=2147483647 tokens
  1039. 42.25.550.524 I slot launch_slot_: id 0 | task 4128 | processing task, is_child = 0
  1040. 42.25.550.574 I slot update_slots: id 0 | task 4128 | Checking checkpoint with [69329, 69329] against 70077...
  1041. 42.25.632.239 W slot update_slots: id 0 | task 4128 | restored context checkpoint (pos_min = 69329, pos_max = 69329, n_tokens = 69330, n_past = 69330, size = 249.626 MiB)
  1042. 42.25.636.710 W find_slot: non-consecutive token position 69669 after 69332 for sequence 0 with 340 new tokens
  1043. 42.25.636.994 W find_slot: non-consecutive token position 69669 after 69332 for sequence 0 with 340 new tokens
  1044. 42.27.423.842 I reasoning-budget: deactivated (natural end)
  1045. 42.27.486.231 W slot process_toke: id 0 | task 4128 | raw tool marker observed while lazy grammar is enabled; keeping DFlash governed by active grammar boundary in_reasoning=0 n_decoded=24 reasoning_tokens=22 visible_tokens=2
  1046. 42.28.746.904 I slot print_timing: id 0 | task 4128 | n_decoded = 100, tg = 56.22 t/s
  1047. 42.29.084.446 I slot print_timing: id 0 | task 4128 | prompt eval time = 1416.72 ms / 856 tokens ( 1.66 ms per token, 604.21 tokens per second)
  1048. 42.29.084.451 I slot print_timing: id 0 | task 4128 | eval time = 2116.28 ms / 101 tokens ( 20.95 ms per token, 47.73 tokens per second)
  1049. 42.29.084.452 I slot print_timing: id 0 | task 4128 | total time = 3533.00 ms / 957 tokens
  1050. 42.29.084.453 I slot print_timing: id 0 | task 4128 | graphs reused = 3454
  1051. 42.29.084.454 I slot print_timing: id 0 | task 4128 | draft acceptance = 0.20800 ( 78 accepted / 375 generated)
  1052. 42.29.084.455 I slot print_timing: id 0 | task 4128 | adaptive dm: fringe=0.00 n_max=15
  1053. 42.29.084.476 I statistics dflash: #calls(b,g,a) = 43 3914 3914, #gen drafts = 3914, #acc drafts = 2769, #gen tokens = 57822, #acc tokens = 8403, dur(b,g,a) = 0.026, 29691.005, 1.224 ms
  1054. 42.29.087.089 I slot release: id 0 | task 4128 | stop processing: n_tokens = 70289, truncated = 0
  1055. 42.29.087.227 I srv update_slots: all slots are idle
  1056. 42.29.502.355 I srv params_from_: Chat format: peg-native
  1057. 42.29.583.825 I slot get_availabl: id 0 | task -1 | selected slot by LCP similarity, sim_best = 0.998 (> 0.100 thold), f_keep = 0.999
  1058. 42.29.583.831 I slot get_availabl: id 0 | task -1 | adaptive dm: preserving state for continuation (sim=0.998, keep=0.999, n_max=15)
  1059. 42.29.584.787 I reasoning-budget: activated, budget=2147483647 tokens
  1060. 42.29.584.972 I slot launch_slot_: id 0 | task 4158 | processing task, is_child = 0
  1061. 42.29.585.025 I slot update_slots: id 0 | task 4158 | Checking checkpoint with [69329, 69329] against 70210...
  1062. 42.29.668.208 W slot update_slots: id 0 | task 4158 | restored context checkpoint (pos_min = 69329, pos_max = 69329, n_tokens = 69330, n_past = 69330, size = 249.626 MiB)
  1063. 42.29.672.713 W find_slot: non-consecutive token position 69802 after 69332 for sequence 0 with 473 new tokens
  1064. 42.29.673.051 W find_slot: non-consecutive token position 69802 after 69332 for sequence 0 with 473 new tokens
  1065. 42.31.532.926 I reasoning-budget: deactivated (natural end)
  1066. 42.31.534.531 W slot process_toke: id 0 | task 4158 | raw tool marker observed while lazy grammar is enabled; keeping DFlash governed by active grammar boundary in_reasoning=0 n_decoded=27 reasoning_tokens=25 visible_tokens=2
  1067. 42.32.971.439 I slot print_timing: id 0 | task 4158 | n_decoded = 100, tg = 50.92 t/s
  1068. 42.33.366.938 I slot print_timing: id 0 | task 4158 | prompt eval time = 1421.72 ms / 989 tokens ( 1.44 ms per token, 695.63 tokens per second)
  1069. 42.33.366.943 I slot print_timing: id 0 | task 4158 | eval time = 2359.36 ms / 104 tokens ( 22.69 ms per token, 44.08 tokens per second)
  1070. 42.33.366.944 I slot print_timing: id 0 | task 4158 | total time = 3781.08 ms / 1093 tokens
  1071. 42.33.366.945 I slot print_timing: id 0 | task 4158 | graphs reused = 3477
  1072. 42.33.366.946 I slot print_timing: id 0 | task 4158 | draft acceptance = 0.20513 ( 80 accepted / 390 generated)
  1073. 42.33.366.947 I slot print_timing: id 0 | task 4158 | adaptive dm: fringe=0.00 n_max=15
  1074. 42.33.366.965 I statistics dflash: #calls(b,g,a) = 44 3940 3940, #gen drafts = 3940, #acc drafts = 2792, #gen tokens = 58212, #acc tokens = 8483, dur(b,g,a) = 0.027, 29884.000, 1.233 ms
  1075. 42.33.369.323 I slot release: id 0 | task 4158 | stop processing: n_tokens = 70425, truncated = 0
  1076. 42.33.369.466 I srv update_slots: all slots are idle
  1077. 42.33.783.083 I srv params_from_: Chat format: peg-native
  1078. 42.33.867.020 I slot get_availabl: id 0 | task -1 | selected slot by LCP similarity, sim_best = 0.998 (> 0.100 thold), f_keep = 0.999
  1079. 42.33.867.026 I slot get_availabl: id 0 | task -1 | adaptive dm: preserving state for continuation (sim=0.998, keep=0.999, n_max=15)
  1080. 42.33.868.012 I reasoning-budget: activated, budget=2147483647 tokens
  1081. 42.33.868.165 I slot launch_slot_: id 0 | task 4189 | processing task, is_child = 0
  1082. 42.33.868.204 I slot update_slots: id 0 | task 4189 | Checking checkpoint with [69329, 69329] against 70346...
  1083. 42.33.951.753 W slot update_slots: id 0 | task 4189 | restored context checkpoint (pos_min = 69329, pos_max = 69329, n_tokens = 69330, n_past = 69330, size = 249.626 MiB)
  1084. 42.33.956.248 W find_slot: non-consecutive token position 69841 after 69333 for sequence 0 with 512 new tokens
  1085. 42.33.956.677 W find_slot: non-consecutive token position 69841 after 69333 for sequence 0 with 512 new tokens
  1086. 42.35.534.590 I slot create_check: id 0 | task 4189 | created context checkpoint 13 of 32 (pos_min = 70423, pos_max = 70423, n_tokens = 70424, size = 249.626 MiB)
  1087. 42.36.302.479 I reasoning-budget: deactivated (natural end)
  1088. 42.36.364.137 W slot process_toke: id 0 | task 4189 | raw tool marker observed while lazy grammar is enabled; keeping DFlash governed by active grammar boundary in_reasoning=0 n_decoded=27 reasoning_tokens=25 visible_tokens=2
  1089. 42.37.753.336 I slot print_timing: id 0 | task 4189 | n_decoded = 100, tg = 48.70 t/s
  1090. 42.38.204.610 I slot print_timing: id 0 | task 4189 | prompt eval time = 1830.98 ms / 1125 tokens ( 1.63 ms per token, 614.43 tokens per second)
  1091. 42.38.204.619 I slot print_timing: id 0 | task 4189 | eval time = 2504.16 ms / 104 tokens ( 24.08 ms per token, 41.53 tokens per second)
  1092. 42.38.204.620 I slot print_timing: id 0 | task 4189 | total time = 4335.14 ms / 1229 tokens
  1093. 42.38.204.622 I slot print_timing: id 0 | task 4189 | graphs reused = 3501
  1094. 42.38.204.624 I slot print_timing: id 0 | task 4189 | draft acceptance = 0.20513 ( 80 accepted / 390 generated)
  1095. 42.38.204.626 I slot print_timing: id 0 | task 4189 | adaptive dm: fringe=0.00 n_max=15
  1096. 42.38.204.651 I statistics dflash: #calls(b,g,a) = 45 3966 3966, #gen drafts = 3966, #acc drafts = 2813, #gen tokens = 58602, #acc tokens = 8563, dur(b,g,a) = 0.028, 30075.798, 1.241 ms
  1097. 42.38.207.721 I slot release: id 0 | task 4189 | stop processing: n_tokens = 70561, truncated = 0
  1098. 42.38.207.925 I srv update_slots: all slots are idle
  1099. 42.38.711.081 I srv params_from_: Chat format: peg-native
  1100. 42.38.804.744 I slot get_availabl: id 0 | task -1 | selected slot by LCP similarity, sim_best = 0.998 (> 0.100 thold), f_keep = 0.999
  1101. 42.38.804.750 I slot get_availabl: id 0 | task -1 | adaptive dm: preserving state for continuation (sim=0.998, keep=0.999, n_max=15)
  1102. 42.38.805.571 I reasoning-budget: activated, budget=2147483647 tokens
  1103. 42.38.805.877 I slot launch_slot_: id 0 | task 4220 | processing task, is_child = 0
  1104. 42.38.805.925 I slot update_slots: id 0 | task 4220 | Checking checkpoint with [70423, 70423] against 70482...
  1105. 42.38.890.027 W slot update_slots: id 0 | task 4220 | restored context checkpoint (pos_min = 70423, pos_max = 70423, n_tokens = 70424, n_past = 70424, size = 249.626 MiB)
  1106. 42.38.895.802 W find_slot: non-consecutive token position 70559 after 70426 for sequence 0 with 136 new tokens
  1107. 42.38.895.957 W find_slot: non-consecutive token position 70559 after 70426 for sequence 0 with 136 new tokens
  1108. 42.39.912.180 I reasoning-budget: deactivated (natural end)
  1109. 42.39.914.912 W slot process_toke: id 0 | task 4220 | raw tool marker observed while lazy grammar is enabled; keeping DFlash governed by active grammar boundary in_reasoning=0 n_decoded=35 reasoning_tokens=33 visible_tokens=3
  1110. 42.40.765.595 I slot print_timing: id 0 | task 4220 | n_decoded = 111, tg = 73.24 t/s
  1111. 42.41.133.112 I slot print_timing: id 0 | task 4220 | prompt eval time = 441.49 ms / 167 tokens ( 2.64 ms per token, 378.27 tokens per second)
  1112. 42.41.133.119 I slot print_timing: id 0 | task 4220 | eval time = 1883.86 ms / 112 tokens ( 16.82 ms per token, 59.45 tokens per second)
  1113. 42.41.133.121 I slot print_timing: id 0 | task 4220 | total time = 2325.35 ms / 279 tokens
  1114. 42.41.133.122 I slot print_timing: id 0 | task 4220 | graphs reused = 3521
  1115. 42.41.133.124 I slot print_timing: id 0 | task 4220 | draft acceptance = 0.26087 ( 90 accepted / 345 generated)
  1116. 42.41.133.125 I slot print_timing: id 0 | task 4220 | adaptive dm: fringe=0.00 n_max=15
  1117. 42.41.133.150 I statistics dflash: #calls(b,g,a) = 46 3989 3989, #gen drafts = 3989, #acc drafts = 2834, #gen tokens = 58947, #acc tokens = 8653, dur(b,g,a) = 0.029, 30243.868, 1.250 ms
  1118. 42.41.135.985 I slot release: id 0 | task 4220 | stop processing: n_tokens = 70704, truncated = 0
  1119. 42.41.136.153 I srv update_slots: all slots are idle
  1120. 42.41.571.257 I srv params_from_: Chat format: peg-native
  1121. 42.41.660.175 I slot get_availabl: id 0 | task -1 | selected slot by LCP similarity, sim_best = 1.000 (> 0.100 thold), f_keep = 1.000
  1122. 42.41.660.183 I slot get_availabl: id 0 | task -1 | adaptive dm: preserving state for continuation (sim=1.000, keep=1.000, n_max=15)
  1123. 42.41.660.958 I reasoning-budget: activated, budget=2147483647 tokens
  1124. 42.41.661.312 I slot launch_slot_: id 0 | task 4247 | processing task, is_child = 0
  1125. 42.42.433.860 I reasoning-budget: deactivated (natural end)
  1126. 42.42.437.144 W slot process_toke: id 0 | task 4247 | raw tool marker observed while lazy grammar is enabled; keeping DFlash governed by active grammar boundary in_reasoning=0 n_decoded=26 reasoning_tokens=24 visible_tokens=2
  1127. 42.45.276.097 I slot print_timing: id 0 | task 4247 | n_decoded = 102, tg = 29.57 t/s
  1128. 42.45.341.210 I slot print_timing: id 0 | task 4247 | prompt eval time = 163.48 ms / 31 tokens ( 5.27 ms per token, 189.63 tokens per second)
  1129. 42.45.341.218 I slot print_timing: id 0 | task 4247 | eval time = 3514.51 ms / 103 tokens ( 34.12 ms per token, 29.31 tokens per second)
  1130. 42.45.341.220 I slot print_timing: id 0 | task 4247 | total time = 3677.99 ms / 134 tokens
  1131. 42.45.341.221 I slot print_timing: id 0 | task 4247 | graphs reused = 3566
  1132. 42.45.341.223 I slot print_timing: id 0 | task 4247 | draft acceptance = 0.07943 ( 56 accepted / 705 generated)
  1133. 42.45.341.225 I slot print_timing: id 0 | task 4247 | adaptive dm: fringe=0.00 n_max=15
  1134. 42.45.341.248 I statistics dflash: #calls(b,g,a) = 47 4036 4036, #gen drafts = 4036, #acc drafts = 2864, #gen tokens = 59652, #acc tokens = 8709, dur(b,g,a) = 0.030, 30576.873, 1.260 ms
  1135. 42.45.343.913 I slot release: id 0 | task 4247 | stop processing: n_tokens = 70838, truncated = 0
  1136. 42.45.344.105 I srv update_slots: all slots are idle
  1137. 42.45.767.493 I srv params_from_: Chat format: peg-native
  1138. 42.45.854.649 I slot get_availabl: id 0 | task -1 | selected slot by LCP similarity, sim_best = 0.998 (> 0.100 thold), f_keep = 0.999
  1139. 42.45.854.655 I slot get_availabl: id 0 | task -1 | adaptive dm: preserving state for continuation (sim=0.998, keep=0.999, n_max=15)
  1140. 42.45.855.375 I reasoning-budget: activated, budget=2147483647 tokens
  1141. 42.45.855.675 I slot launch_slot_: id 0 | task 4297 | processing task, is_child = 0
  1142. 42.45.855.726 I slot update_slots: id 0 | task 4297 | Checking checkpoint with [70423, 70423] against 70761...
  1143. 42.45.938.019 W slot update_slots: id 0 | task 4297 | restored context checkpoint (pos_min = 70423, pos_max = 70423, n_tokens = 70424, n_past = 70424, size = 249.626 MiB)
  1144. 42.45.942.838 W find_slot: non-consecutive token position 70838 after 70424 for sequence 0 with 415 new tokens
  1145. 42.45.943.208 W find_slot: non-consecutive token position 70838 after 70424 for sequence 0 with 415 new tokens
  1146. 42.49.032.365 I slot print_timing: id 0 | task 4297 | n_decoded = 109, tg = 46.41 t/s
  1147. 42.52.080.698 I slot print_timing: id 0 | task 4297 | n_decoded = 616, tg = 114.12 t/s
  1148. 42.52.991.540 I reasoning-budget: deactivated (natural end)
  1149. 42.53.056.029 W slot process_toke: id 0 | task 4297 | raw tool marker observed while lazy grammar is enabled; keeping DFlash governed by active grammar boundary in_reasoning=0 n_decoded=693 reasoning_tokens=691 visible_tokens=2
  1150. 42.55.110.622 I slot print_timing: id 0 | task 4297 | prompt eval time = 827.09 ms / 446 tokens ( 1.85 ms per token, 539.24 tokens per second)
  1151. 42.55.110.629 I slot print_timing: id 0 | task 4297 | eval time = 8426.83 ms / 770 tokens ( 10.94 ms per token, 91.37 tokens per second)
  1152. 42.55.110.631 I slot print_timing: id 0 | task 4297 | total time = 9253.92 ms / 1216 tokens
  1153. 42.55.110.632 I slot print_timing: id 0 | task 4297 | graphs reused = 3687
  1154. 42.55.110.633 I slot print_timing: id 0 | task 4297 | draft acceptance = 0.34180 ( 646 accepted / 1890 generated)
  1155. 42.55.110.635 I slot print_timing: id 0 | task 4297 | adaptive dm: fringe=0.00 n_max=15
  1156. 42.55.110.659 I statistics dflash: #calls(b,g,a) = 48 4162 4162, #gen drafts = 4162, #acc drafts = 2966, #gen tokens = 61542, #acc tokens = 9355, dur(b,g,a) = 0.031, 31512.081, 1.290 ms
  1157. 42.55.113.168 I slot release: id 0 | task 4297 | stop processing: n_tokens = 71642, truncated = 0
  1158. 42.55.113.349 I slot print_timing: id 0 | task -1 | n_decoded = 770, tg = 91.34 t/s
  1159. 42.55.113.362 I srv update_slots: all slots are idle
  1160. 42.55.531.765 I srv params_from_: Chat format: peg-native
  1161. 42.55.614.778 I slot get_availabl: id 0 | task -1 | selected slot by LCP similarity, sim_best = 0.998 (> 0.100 thold), f_keep = 0.999
  1162. 42.55.614.784 I slot get_availabl: id 0 | task -1 | adaptive dm: preserving state for continuation (sim=0.998, keep=0.999, n_max=15)
  1163. 42.55.615.518 I reasoning-budget: activated, budget=2147483647 tokens
  1164. 42.55.615.812 I slot launch_slot_: id 0 | task 4427 | processing task, is_child = 0
  1165. 42.55.615.860 I slot update_slots: id 0 | task 4427 | Checking checkpoint with [70423, 70423] against 71563...
  1166. 42.55.696.080 W slot update_slots: id 0 | task 4427 | restored context checkpoint (pos_min = 70423, pos_max = 70423, n_tokens = 70424, n_past = 70424, size = 249.626 MiB)
  1167. 42.55.700.553 W find_slot: non-consecutive token position 70935 after 70426 for sequence 0 with 512 new tokens
  1168. 42.55.701.050 W find_slot: non-consecutive token position 70935 after 70426 for sequence 0 with 512 new tokens
  1169. 42.57.413.230 I slot create_check: id 0 | task 4427 | created context checkpoint 14 of 32 (pos_min = 71640, pos_max = 71640, n_tokens = 71641, size = 249.626 MiB)
  1170. 42.58.230.779 I reasoning-budget: deactivated (natural end)
  1171. 42.58.232.112 W slot process_toke: id 0 | task 4427 | raw tool marker observed while lazy grammar is enabled; keeping DFlash governed by active grammar boundary in_reasoning=0 n_decoded=34 reasoning_tokens=32 visible_tokens=2
  1172. 42.59.544.989 I slot print_timing: id 0 | task 4427 | n_decoded = 107, tg = 54.37 t/s
  1173. 43.00.000.376 I slot print_timing: id 0 | task 4427 | prompt eval time = 1960.28 ms / 1248 tokens ( 1.57 ms per token, 636.64 tokens per second)
  1174. 43.00.000.382 I slot print_timing: id 0 | task 4427 | eval time = 2421.61 ms / 111 tokens ( 21.82 ms per token, 45.84 tokens per second)
  1175. 43.00.000.383 I slot print_timing: id 0 | task 4427 | total time = 4381.90 ms / 1359 tokens
  1176. 43.00.000.384 I slot print_timing: id 0 | task 4427 | graphs reused = 3710
  1177. 43.00.000.386 I slot print_timing: id 0 | task 4427 | draft acceptance = 0.23467 ( 88 accepted / 375 generated)
  1178. 43.00.000.387 I slot print_timing: id 0 | task 4427 | adaptive dm: fringe=0.00 n_max=15
  1179. 43.00.000.410 I statistics dflash: #calls(b,g,a) = 49 4187 4187, #gen drafts = 4187, #acc drafts = 2988, #gen tokens = 61917, #acc tokens = 9443, dur(b,g,a) = 0.031, 31695.366, 1.301 ms
  1180. 43.00.003.227 I slot release: id 0 | task 4427 | stop processing: n_tokens = 71785, truncated = 0
  1181. 43.00.003.432 I srv update_slots: all slots are idle
  1182. 43.00.446.048 I srv params_from_: Chat format: peg-native
  1183. 43.00.535.736 I slot get_availabl: id 0 | task -1 | selected slot by LCP similarity, sim_best = 0.998 (> 0.100 thold), f_keep = 0.999
  1184. 43.00.535.742 I slot get_availabl: id 0 | task -1 | adaptive dm: preserving state for continuation (sim=0.998, keep=0.999, n_max=15)
  1185. 43.00.536.476 I reasoning-budget: activated, budget=2147483647 tokens
  1186. 43.00.536.806 I slot launch_slot_: id 0 | task 4457 | processing task, is_child = 0
  1187. 43.00.536.855 I slot update_slots: id 0 | task 4457 | Checking checkpoint with [71640, 71640] against 71706...
  1188. 43.00.622.414 W slot update_slots: id 0 | task 4457 | restored context checkpoint (pos_min = 71640, pos_max = 71640, n_tokens = 71641, n_past = 71641, size = 249.626 MiB)
  1189. 43.00.627.421 W find_slot: non-consecutive token position 71783 after 71643 for sequence 0 with 143 new tokens
  1190. 43.00.627.633 W find_slot: non-consecutive token position 71783 after 71643 for sequence 0 with 143 new tokens
  1191. 43.02.758.116 I slot print_timing: id 0 | task 4457 | n_decoded = 109, tg = 63.37 t/s
  1192. 43.04.338.201 I reasoning-budget: deactivated (natural end)
  1193. 43.04.401.683 W slot process_toke: id 0 | task 4457 | raw tool marker observed while lazy grammar is enabled; keeping DFlash governed by active grammar boundary in_reasoning=0 n_decoded=219 reasoning_tokens=217 visible_tokens=2
  1194. 43.05.787.429 I slot print_timing: id 0 | task 4457 | n_decoded = 291, tg = 61.28 t/s
  1195. 43.06.286.196 I slot print_timing: id 0 | task 4457 | prompt eval time = 499.87 ms / 174 tokens ( 2.87 ms per token, 348.09 tokens per second)
  1196. 43.06.286.203 I slot print_timing: id 0 | task 4457 | eval time = 5247.24 ms / 296 tokens ( 17.73 ms per token, 56.41 tokens per second)
  1197. 43.06.286.204 I slot print_timing: id 0 | task 4457 | total time = 5747.11 ms / 470 tokens
  1198. 43.06.286.205 I slot print_timing: id 0 | task 4457 | graphs reused = 3781
  1199. 43.06.286.207 I slot print_timing: id 0 | task 4457 | draft acceptance = 0.20180 ( 224 accepted / 1110 generated)
  1200. 43.06.286.208 I slot print_timing: id 0 | task 4457 | adaptive dm: fringe=0.00 n_max=15
  1201. 43.06.286.231 I statistics dflash: #calls(b,g,a) = 50 4261 4261, #gen drafts = 4261, #acc drafts = 3046, #gen tokens = 63027, #acc tokens = 9667, dur(b,g,a) = 0.031, 32248.655, 1.322 ms
  1202. 43.06.288.809 I slot release: id 0 | task 4457 | stop processing: n_tokens = 72113, truncated = 0
  1203. 43.06.288.997 I srv update_slots: all slots are idle
  1204. 43.06.786.451 I srv params_from_: Chat format: peg-native
  1205. 43.06.876.106 I slot get_availabl: id 0 | task -1 | selected slot by LCP similarity, sim_best = 0.999 (> 0.100 thold), f_keep = 0.999
  1206. 43.06.876.112 I slot get_availabl: id 0 | task -1 | adaptive dm: preserving state for continuation (sim=0.999, keep=0.999, n_max=15)
  1207. 43.06.876.842 I reasoning-budget: activated, budget=2147483647 tokens
  1208. 43.06.877.166 I slot launch_slot_: id 0 | task 4535 | processing task, is_child = 0
  1209. 43.06.877.224 I slot update_slots: id 0 | task 4535 | Checking checkpoint with [71640, 71640] against 72034...
  1210. 43.06.963.974 W slot update_slots: id 0 | task 4535 | restored context checkpoint (pos_min = 71640, pos_max = 71640, n_tokens = 71641, n_past = 71641, size = 249.626 MiB)
  1211. 43.06.968.481 W find_slot: non-consecutive token position 72111 after 71643 for sequence 0 with 471 new tokens
  1212. 43.06.968.846 W find_slot: non-consecutive token position 72111 after 71643 for sequence 0 with 471 new tokens
  1213. 43.08.848.602 I reasoning-budget: deactivated (natural end)
  1214. 43.08.851.166 W slot process_toke: id 0 | task 4535 | raw tool marker observed while lazy grammar is enabled; keeping DFlash governed by active grammar boundary in_reasoning=0 n_decoded=39 reasoning_tokens=37 visible_tokens=2
  1215. 43.10.065.317 I slot print_timing: id 0 | task 4535 | n_decoded = 104, tg = 45.15 t/s
  1216. 43.10.610.112 I slot print_timing: id 0 | task 4535 | prompt eval time = 883.74 ms / 502 tokens ( 1.76 ms per token, 568.04 tokens per second)
  1217. 43.10.610.121 I slot print_timing: id 0 | task 4535 | eval time = 2848.19 ms / 116 tokens ( 24.55 ms per token, 40.73 tokens per second)
  1218. 43.10.610.123 I slot print_timing: id 0 | task 4535 | total time = 3731.92 ms / 618 tokens
  1219. 43.10.610.124 I slot print_timing: id 0 | task 4535 | graphs reused = 3810
  1220. 43.10.610.126 I slot print_timing: id 0 | task 4535 | draft acceptance = 0.17917 ( 86 accepted / 480 generated)
  1221. 43.10.610.127 I slot print_timing: id 0 | task 4535 | adaptive dm: fringe=0.00 n_max=15
  1222. 43.10.610.157 I statistics dflash: #calls(b,g,a) = 51 4293 4293, #gen drafts = 4293, #acc drafts = 3071, #gen tokens = 63507, #acc tokens = 9753, dur(b,g,a) = 0.031, 32483.438, 1.336 ms
  1223. 43.10.613.326 I slot release: id 0 | task 4535 | stop processing: n_tokens = 72261, truncated = 0
  1224. 43.10.613.509 I srv update_slots: all slots are idle
  1225. 43.11.057.923 I srv params_from_: Chat format: peg-native
  1226. 43.11.141.349 I slot get_availabl: id 0 | task -1 | selected slot by LCP similarity, sim_best = 0.999 (> 0.100 thold), f_keep = 0.999
  1227. 43.11.141.356 I slot get_availabl: id 0 | task -1 | adaptive dm: preserving state for continuation (sim=0.999, keep=0.999, n_max=15)
  1228. 43.11.142.129 I reasoning-budget: activated, budget=2147483647 tokens
  1229. 43.11.142.463 I slot launch_slot_: id 0 | task 4571 | processing task, is_child = 0
  1230. 43.11.142.517 I slot update_slots: id 0 | task 4571 | Checking checkpoint with [71640, 71640] against 72182...
  1231. 43.11.225.667 W slot update_slots: id 0 | task 4571 | restored context checkpoint (pos_min = 71640, pos_max = 71640, n_tokens = 71641, n_past = 71641, size = 249.626 MiB)
  1232. 43.11.230.638 W find_slot: non-consecutive token position 71774 after 71643 for sequence 0 with 134 new tokens
  1233. 43.11.230.793 W find_slot: non-consecutive token position 71774 after 71643 for sequence 0 with 134 new tokens
  1234. 43.13.674.214 I reasoning-budget: deactivated (natural end)
  1235. 43.13.738.881 W slot process_toke: id 0 | task 4571 | raw tool marker observed while lazy grammar is enabled; keeping DFlash governed by active grammar boundary in_reasoning=0 n_decoded=55 reasoning_tokens=53 visible_tokens=2
  1236. 43.14.833.452 I slot print_timing: id 0 | task 4571 | n_decoded = 109, tg = 42.10 t/s
  1237. 43.15.477.490 I slot print_timing: id 0 | task 4571 | prompt eval time = 1101.38 ms / 650 tokens ( 1.69 ms per token, 590.17 tokens per second)
  1238. 43.15.477.499 I slot print_timing: id 0 | task 4571 | eval time = 3232.42 ms / 132 tokens ( 24.49 ms per token, 40.84 tokens per second)
  1239. 43.15.477.501 I slot print_timing: id 0 | task 4571 | total time = 4333.80 ms / 782 tokens
  1240. 43.15.477.503 I slot print_timing: id 0 | task 4571 | graphs reused = 3846
  1241. 43.15.477.505 I slot print_timing: id 0 | task 4571 | draft acceptance = 0.16842 ( 96 accepted / 570 generated)
  1242. 43.15.477.507 I slot print_timing: id 0 | task 4571 | adaptive dm: fringe=0.00 n_max=15
  1243. 43.15.477.537 I statistics dflash: #calls(b,g,a) = 52 4331 4331, #gen drafts = 4331, #acc drafts = 3101, #gen tokens = 64077, #acc tokens = 9849, dur(b,g,a) = 0.031, 32763.016, 1.348 ms
  1244. 43.15.480.697 I slot release: id 0 | task 4571 | stop processing: n_tokens = 72425, truncated = 0
  1245. 43.15.480.906 I srv update_slots: all slots are idle
  1246. 43.15.992.039 I srv params_from_: Chat format: peg-native
  1247. 43.16.078.797 I slot get_availabl: id 0 | task -1 | selected slot by LCP similarity, sim_best = 0.999 (> 0.100 thold), f_keep = 0.999
  1248. 43.16.078.805 I slot get_availabl: id 0 | task -1 | adaptive dm: preserving state for continuation (sim=0.999, keep=0.999, n_max=15)
  1249. 43.16.079.911 I reasoning-budget: activated, budget=2147483647 tokens
  1250. 43.16.080.396 I slot launch_slot_: id 0 | task 4614 | processing task, is_child = 0
  1251. 43.16.080.452 I slot update_slots: id 0 | task 4614 | Checking checkpoint with [71640, 71640] against 72346...
  1252. 43.16.162.593 W slot update_slots: id 0 | task 4614 | restored context checkpoint (pos_min = 71640, pos_max = 71640, n_tokens = 71641, n_past = 71641, size = 249.626 MiB)
  1253. 43.16.167.154 W find_slot: non-consecutive token position 71938 after 71643 for sequence 0 with 298 new tokens
  1254. 43.16.167.414 W find_slot: non-consecutive token position 71938 after 71643 for sequence 0 with 298 new tokens
  1255. 43.18.407.793 I reasoning-budget: deactivated (natural end)
  1256. 43.18.409.067 W slot process_toke: id 0 | task 4614 | raw tool marker observed while lazy grammar is enabled; keeping DFlash governed by active grammar boundary in_reasoning=0 n_decoded=58 reasoning_tokens=56 visible_tokens=2
  1257. 43.19.476.948 I slot print_timing: id 0 | task 4614 | n_decoded = 114, tg = 55.84 t/s
  1258. 43.20.119.444 I slot print_timing: id 0 | task 4614 | prompt eval time = 1354.81 ms / 814 tokens ( 1.66 ms per token, 600.82 tokens per second)
  1259. 43.20.119.452 I slot print_timing: id 0 | task 4614 | eval time = 2683.08 ms / 135 tokens ( 19.87 ms per token, 50.32 tokens per second)
  1260. 43.20.119.453 I slot print_timing: id 0 | task 4614 | total time = 4037.89 ms / 949 tokens
  1261. 43.20.119.455 I slot print_timing: id 0 | task 4614 | graphs reused = 3873
  1262. 43.20.119.457 I slot print_timing: id 0 | task 4614 | draft acceptance = 0.24828 ( 108 accepted / 435 generated)
  1263. 43.20.119.459 I slot print_timing: id 0 | task 4614 | adaptive dm: fringe=0.00 n_max=15
  1264. 43.20.119.483 I statistics dflash: #calls(b,g,a) = 53 4360 4360, #gen drafts = 4360, #acc drafts = 3129, #gen tokens = 64512, #acc tokens = 9957, dur(b,g,a) = 0.032, 32979.525, 1.358 ms
  1265. 43.20.122.049 I slot release: id 0 | task 4614 | stop processing: n_tokens = 72592, truncated = 0
  1266. 43.20.122.224 I srv update_slots: all slots are idle
  1267. 43.20.580.645 I srv params_from_: Chat format: peg-native
  1268. 43.20.665.936 I slot get_availabl: id 0 | task -1 | selected slot by LCP similarity, sim_best = 0.999 (> 0.100 thold), f_keep = 0.999
  1269. 43.20.665.942 I slot get_availabl: id 0 | task -1 | adaptive dm: preserving state for continuation (sim=0.999, keep=0.999, n_max=15)
  1270. 43.20.666.685 I reasoning-budget: activated, budget=2147483647 tokens
  1271. 43.20.666.991 I slot launch_slot_: id 0 | task 4648 | processing task, is_child = 0
  1272. 43.20.667.049 I slot update_slots: id 0 | task 4648 | Checking checkpoint with [71640, 71640] against 72513...
  1273. 43.20.749.958 W slot update_slots: id 0 | task 4648 | restored context checkpoint (pos_min = 71640, pos_max = 71640, n_tokens = 71641, n_past = 71641, size = 249.626 MiB)
  1274. 43.20.754.556 W find_slot: non-consecutive token position 72105 after 71643 for sequence 0 with 465 new tokens
  1275. 43.20.754.909 W find_slot: non-consecutive token position 72105 after 71643 for sequence 0 with 465 new tokens
  1276. 43.23.261.234 I reasoning-budget: deactivated (natural end)
  1277. 43.23.263.910 W slot process_toke: id 0 | task 4648 | raw tool marker observed while lazy grammar is enabled; keeping DFlash governed by active grammar boundary in_reasoning=0 n_decoded=43 reasoning_tokens=41 visible_tokens=2
  1278. 43.24.235.642 I slot print_timing: id 0 | task 4648 | n_decoded = 103, tg = 51.63 t/s
  1279. 43.24.882.864 I slot print_timing: id 0 | task 4648 | prompt eval time = 1571.59 ms / 981 tokens ( 1.60 ms per token, 624.21 tokens per second)
  1280. 43.24.882.870 I slot print_timing: id 0 | task 4648 | eval time = 2641.57 ms / 120 tokens ( 22.01 ms per token, 45.43 tokens per second)
  1281. 43.24.882.871 I slot print_timing: id 0 | task 4648 | total time = 4213.15 ms / 1101 tokens
  1282. 43.24.882.872 I slot print_timing: id 0 | task 4648 | graphs reused = 3902
  1283. 43.24.882.874 I slot print_timing: id 0 | task 4648 | draft acceptance = 0.18750 ( 90 accepted / 480 generated)
  1284. 43.24.882.875 I slot print_timing: id 0 | task 4648 | adaptive dm: fringe=0.00 n_max=15
  1285. 43.24.882.898 I statistics dflash: #calls(b,g,a) = 54 4392 4392, #gen drafts = 4392, #acc drafts = 3155, #gen tokens = 64992, #acc tokens = 10047, dur(b,g,a) = 0.033, 33212.319, 1.368 ms
  1286. 43.24.885.564 I slot release: id 0 | task 4648 | stop processing: n_tokens = 72744, truncated = 0
  1287. 43.24.885.731 I srv update_slots: all slots are idle
  1288. 43.25.329.738 I srv params_from_: Chat format: peg-native
  1289. 43.25.415.457 I slot get_availabl: id 0 | task -1 | selected slot by LCP similarity, sim_best = 0.999 (> 0.100 thold), f_keep = 0.999
  1290. 43.25.415.463 I slot get_availabl: id 0 | task -1 | adaptive dm: preserving state for continuation (sim=0.999, keep=0.999, n_max=15)
  1291. 43.25.416.379 I reasoning-budget: activated, budget=2147483647 tokens
  1292. 43.25.416.989 I slot launch_slot_: id 0 | task 4685 | processing task, is_child = 0
  1293. 43.25.417.044 I slot update_slots: id 0 | task 4685 | Checking checkpoint with [71640, 71640] against 72665...
  1294. 43.25.500.666 W slot update_slots: id 0 | task 4685 | restored context checkpoint (pos_min = 71640, pos_max = 71640, n_tokens = 71641, n_past = 71641, size = 249.626 MiB)
  1295. 43.25.505.596 W find_slot: non-consecutive token position 72152 after 71643 for sequence 0 with 512 new tokens
  1296. 43.25.506.100 W find_slot: non-consecutive token position 72152 after 71643 for sequence 0 with 512 new tokens
  1297. 43.27.101.949 I slot create_check: id 0 | task 4685 | created context checkpoint 15 of 32 (pos_min = 72742, pos_max = 72742, n_tokens = 72743, size = 249.626 MiB)
  1298. 43.27.264.684 W srv stop: cancel task, id_task = 4685
  1299. 43.27.346.166 I slot process_sing: id 0 | task 4685 | adaptive dm: reset state for canceled task
  1300. 43.27.346.172 I slot release: id 0 | task 4685 | stop processing: n_tokens = 72776, truncated = 0
  1301. 43.27.346.379 I srv update_slots: all slots are idle
  1302. 81.53.582.436 I srv operator (): operator (): cleaning up before exit...
Tags: log
Add Comment
Please, Sign In to add comment
Public Pastes
We use cookies for various purposes including analytics. By continuing to use Pastebin, you agree to our use of cookies as described in the Cookies Policy.  OK, I Understand
Not a member of Pastebin yet?
Sign Up, it unlocks many cool features!