./llama-bench -m /home/user/.lmstudio/models/unsloth/Qwen3.6-27B-GGUF/Qwen3.6-27B-Q4_K_M.gguf -fa 1 -r 3 -p 512,8192 -n 128,1024 -d 0,16384,32768 -ngl 999 ggml_vulkan: Found 1 Vulkan devices: ggml_vulkan: 0 = Radeon 8060S Graphics (RADV STRIX_HALO) (radv) | uma: 1 | fp16: 1 | bf16: 0 | warp size: 64 | shared memory: 65536 | int dot: 1 | matrix cores: KHR_coopmat | model | size | params | backend | ngl | fa | test | t/s | | ------------------------------ | ---------: | ---------: | ---------- | --: | -: | --------------: | -------------------: | | qwen35 27B Q4_K - Medium | 15.65 GiB | 26.90 B | Vulkan | 999 | 1 | pp512 | 339.20 ± 1.18 | | qwen35 27B Q4_K - Medium | 15.65 GiB | 26.90 B | Vulkan | 999 | 1 | pp8192 | 315.63 ± 0.90 | | qwen35 27B Q4_K - Medium | 15.65 GiB | 26.90 B | Vulkan | 999 | 1 | tg128 | 12.55 ± 0.00 | | qwen35 27B Q4_K - Medium | 15.65 GiB | 26.90 B | Vulkan | 999 | 1 | tg1024 | 12.50 ± 0.00 | | qwen35 27B Q4_K - Medium | 15.65 GiB | 26.90 B | Vulkan | 999 | 1 | pp512 @ d16384 | 239.20 ± 1.29 | | qwen35 27B Q4_K - Medium | 15.65 GiB | 26.90 B | Vulkan | 999 | 1 | pp8192 @ d16384 | 209.30 ± 0.62 | | qwen35 27B Q4_K - Medium | 15.65 GiB | 26.90 B | Vulkan | 999 | 1 | tg128 @ d16384 | 11.75 ± 0.00 | | qwen35 27B Q4_K - Medium | 15.65 GiB | 26.90 B | Vulkan | 999 | 1 | tg1024 @ d16384 | 11.71 ± 0.00 | | qwen35 27B Q4_K - Medium | 15.65 GiB | 26.90 B | Vulkan | 999 | 1 | pp512 @ d32768 | 131.50 ± 1.19 | | qwen35 27B Q4_K - Medium | 15.65 GiB | 26.90 B | Vulkan | 999 | 1 | pp8192 @ d32768 | 116.64 ± 0.28 | | qwen35 27B Q4_K - Medium | 15.65 GiB | 26.90 B | Vulkan | 999 | 1 | tg128 @ d32768 | 11.10 ± 0.00 | | qwen35 27B Q4_K - Medium | 15.65 GiB | 26.90 B | Vulkan | 999 | 1 | tg1024 @ d32768 | 11.08 ± 0.00 | build: 0d0764dfd (8892) ./llama-bench -m /home/user/.lmstudio/models/unsloth/Qwen3.6-27B-GGUF/Qwen3.6-27B-Q4_K_M.gguf -fa 1 -r 3 -p 512,8192 -n 128,1024 -d 0,16384,32768 -ngl 999 -ctk q8_0 -ctv q8_0 ggml_vulkan: Found 1 Vulkan devices: ggml_vulkan: 0 = Radeon 8060S Graphics (RADV STRIX_HALO) (radv) | uma: 1 | fp16: 1 | bf16: 0 | warp size: 64 | shared memory: 65536 | int dot: 1 | matrix cores: KHR_coopmat | model | size | params | backend | ngl | type_k | type_v | fa | test | t/s | | ------------------------------ | ---------: | ---------: | ---------- | --: | -----: | -----: | -: | --------------: | -------------------: | | qwen35 27B Q4_K - Medium | 15.65 GiB | 26.90 B | Vulkan | 999 | q8_0 | q8_0 | 1 | pp512 | 344.90 ± 0.50 | | qwen35 27B Q4_K - Medium | 15.65 GiB | 26.90 B | Vulkan | 999 | q8_0 | q8_0 | 1 | pp8192 | 323.26 ± 0.40 | | qwen35 27B Q4_K - Medium | 15.65 GiB | 26.90 B | Vulkan | 999 | q8_0 | q8_0 | 1 | tg128 | 12.67 ± 0.00 | | qwen35 27B Q4_K - Medium | 15.65 GiB | 26.90 B | Vulkan | 999 | q8_0 | q8_0 | 1 | tg1024 | 12.65 ± 0.00 | | qwen35 27B Q4_K - Medium | 15.65 GiB | 26.90 B | Vulkan | 999 | q8_0 | q8_0 | 1 | pp512 @ d16384 | 268.90 ± 0.06 | | qwen35 27B Q4_K - Medium | 15.65 GiB | 26.90 B | Vulkan | 999 | q8_0 | q8_0 | 1 | pp8192 @ d16384 | 255.61 ± 0.02 | | qwen35 27B Q4_K - Medium | 15.65 GiB | 26.90 B | Vulkan | 999 | q8_0 | q8_0 | 1 | tg128 @ d16384 | 12.11 ± 0.01 | | qwen35 27B Q4_K - Medium | 15.65 GiB | 26.90 B | Vulkan | 999 | q8_0 | q8_0 | 1 | tg1024 @ d16384 | 12.12 ± 0.02 | | qwen35 27B Q4_K - Medium | 15.65 GiB | 26.90 B | Vulkan | 999 | q8_0 | q8_0 | 1 | pp512 @ d32768 | 219.13 ± 0.16 | | qwen35 27B Q4_K - Medium | 15.65 GiB | 26.90 B | Vulkan | 999 | q8_0 | q8_0 | 1 | pp8192 @ d32768 | 209.17 ± 0.03 | | qwen35 27B Q4_K - Medium | 15.65 GiB | 26.90 B | Vulkan | 999 | q8_0 | q8_0 | 1 | tg128 @ d32768 | 11.73 ± 0.00 | | qwen35 27B Q4_K - Medium | 15.65 GiB | 26.90 B | Vulkan | 999 | q8_0 | q8_0 | 1 | tg1024 @ d32768 | 11.71 ± 0.00 | build: 0d0764dfd (8892)