Zig mimalloc-bench

calc = standard malloc
jdzw = JdzGlobalAllocator, with a c wrapper adding deinitThread() calls on pthread destructor
jdzl = JdzAllocator
rpz = rpmalloc-zig, with a c wrapper adding initThread() and deinitThread() calls on pthread_create/destructor
zmlc = zimalloc

rpz currently seems to segfault on crossthread frees, relevant issue here https://github.com/InKryption/rpmalloc-zig-port/issues/6 - this corresponds to the signal 11 errors in the benchmark results

zig code compiled into a shared library which uses threadlocal variables may incur a significant _tls_get_addr overhead due to inability to use initial-exec tls model. On cfrac, this overhead appears to be ~.35s, but it does not seem to be consistent across benchmarks. This affects jdzw and rpz, but not jdzl, calc and zmlc. Relevant reading here: https://maskray.me/blog/2021-02-14-all-about-thread-local-storage

#------------------------------------------------------------------
# test    alloc   time  rss    user  sys  page-faults page-reclaims
cfrac       calc  03.58 3072 3.57 0.00 0 432
cfrac       jdzw  03.55 3712 3.55 0.00 0 382
cfrac       jdzl  04.42 3840 4.42 0.00 0 606
cfrac       rpz   03.82 4096 3.81 0.00 0 618
cfrac       zmlc  06.19 3584 6.19 0.00 0 530
espresso    calc  03.94 2432 3.93 0.00 0 466
espresso    jdzw  03.68 6272 3.68 0.00 0 1125
espresso    jdzl  04.00 6656 4.00 0.00 0 1347
espresso    rpz   03.76 7936 3.74 0.01 0 1668
espresso    zmlc  04.67 44288 4.65 0.01 0 40221
barnes      calc  01.86 58368 1.84 0.02 0 16563
barnes      jdzw  01.84 59136 1.84 0.00 0 16591
barnes      jdzl  01.85 59264 1.84 0.00 0 16809
barnes      rpz   01.91 59392 1.88 0.02 0 16806
barnes      zmlc  01.85 58368 1.85 0.00 0 16609
redis       calc  3.459 8064 0.15 0.02 0 1226
redis       jdzw  2.899 8960 0.12 0.03 0 1296
redis       jdzl  3.099 9472 0.15 0.02 0 1515
redis       rpz   3.140 9984 0.13 0.04 0 1722
redis       zmlc  4.360 14592 0.21 0.02 0 2849
leanN       calc  10.99 246680 25.56 0.41 0 60925
leanN       jdzw  09.64 251648 22.38 0.41 0 60181
leanN       jdzl  10.66 237392 25.64 0.34 0 57387
Command terminated by signal 11
leanN       rpz   00.07 19968 0.01 0.01 0 2592
leanN       zmlc  18.75 860576 58.39 0.54 0 239853
larsonN-sized calc  4.374 111744 79.05 0.32 0 27537
larsonN-sized jdzw  4.056 204360 78.92 0.25 0 53356
larsonN-sized jdzl  17.180 193408 76.43 0.12 0 47992
Command terminated by signal 11
larsonN-sized rpz    109952 8.14 0.04 0 27097
larsonN-sized zmlc  121.954 418304 79.26 0.13 0 104122
rptestN     calc  1.628 47888 4.63 1.45 4 23932
rptestN     jdzw  2.356 156636 7.09 1.10 1 39526
rptestN     jdzl  2.344 110460 7.21 0.95 2 27601
Command terminated by signal 11
rptestN     rpz    52864 0.04 0.09 0 13083
rptestN     zmlc  3.003 930720 8.13 2.43 0 319584
gs          calc  00.71 41100 0.71 0.00 0 7509
gs          jdzw  00.72 44416 0.71 0.00 0 7352
gs          jdzl  00.72 44288 0.71 0.01 0 7553
Command terminated by signal 11
gs          rpz   01.07 15488 0.00 0.00 0 1755
gs          zmlc  00.75 79744 0.73 0.02 0 16422
lua         calc  03.55 99584 3.26 0.26 0 159134
lua         jdzw  03.51 100608 3.21 0.26 0 167373
lua         jdzl  03.56 100864 3.25 0.28 0 190293
lua         rpz   03.56 101632 3.30 0.23 0 201916
lua         zmlc  04.00 135928 3.44 0.53 0 464067
alloc-test1 calc  03.10 13568 3.09 0.00 0 2925
alloc-test1 jdzw  02.82 13824 2.81 0.00 0 2772
alloc-test1 jdzl  03.06 13696 3.05 0.00 0 2886
alloc-test1 rpz   03.46 13568 3.46 0.00 0 2865
alloc-test1 zmlc  04.26 12800 4.26 0.00 0 2734
alloc-testN calc  04.46 15488 69.81 0.06 0 3799
alloc-testN jdzw  04.49 22656 70.27 0.03 0 5394
alloc-testN jdzl  13.24 21344 169.39 0.01 0 4948
alloc-testN rpz   06.39 23168 99.22 0.01 0 5337
alloc-testN zmlc  2:01.82 34812 1889.75 0.23 0 8156
sh6benchN   calc  00.61 340224 8.49 0.39 0 85053
sh6benchN   jdzw  00.30 258760 4.27 0.18 0 65326
sh6benchN   jdzl  03.04 253548 44.97 0.12 0 63297
sh6benchN   rpz   00.42 255232 6.20 0.14 0 63704
sh6benchN   zmlc  25.38 231040 394.26 0.12 0 57595
sh8benchN   calc  02.20 248340 32.89 0.23 0 81345
sh8benchN   jdzw  00.81 156160 11.80 0.07 0 38826
sh8benchN   jdzl  07.16 151168 104.71 0.11 0 37756
Command terminated by signal 11
sh8benchN   rpz   00.09 89728 0.00 0.03 0 22403
sh8benchN   zmlc  1:00.16 350388 939.36 2.84 0 1361511
xmalloc-testN calc  1.906 65564 73.52 0.92 0 37802
xmalloc-testN jdzw  0.428 399232 68.69 2.31 0 99628
xmalloc-testN jdzl  1.387 43520 78.90 0.48 5 10835
Command terminated by signal 11
xmalloc-testN rpz    33920 0.31 0.00 0 8448
xmalloc-testN zmlc  15.287 153592 78.87 0.36 0 175544
cache-scratch1 calc  00.96 3456 0.96 0.00 0 216
cache-scratch1 jdzw  01.00 4352 1.00 0.00 0 240
cache-scratch1 jdzl  01.01 4736 1.01 0.00 0 458
cache-scratch1 rpz   00.93 4480 0.92 0.00 0 441
cache-scratch1 zmlc  00.98 4096 0.98 0.00 0 363
cache-scratchN calc  00.14 3584 2.12 0.00 0 263
cache-scratchN jdzw  00.14 4224 2.21 0.01 0 270
cache-scratchN jdzl  00.14 4736 2.25 0.00 0 489
cache-scratchN rpz   00.15 4608 2.26 0.00 0 474
cache-scratchN zmlc  00.15 4224 2.28 0.00 0 412
glibc-simple calc  03.16 1792 3.15 0.00 0 209
glibc-simple jdzw  02.06 3200 2.05 0.00 0 476
glibc-simple jdzl  02.86 2944 2.85 0.00 0 500
glibc-simple rpz   02.30 2816 2.29 0.00 0 487
glibc-simple zmlc  04.96 4224 4.89 0.05 0 79602
glibc-thread calc  1.697 4096 31.82 0.01 14 965
glibc-thread jdzw  1.698 35200 31.69 0.02 12 8607
glibc-thread jdzl  7.373 33536 31.17 0.02 14 8369
glibc-thread rpz   2.052 18432 31.79 0.01 13 4737
glibc-thread zmlc  61.586 187136 31.37 0.13 15 46777
rocksdb     calc  04.37 91644 3.72 0.93 0 37101
rocksdb     jdzw  04.01 93696 3.48 0.78 0 20480
rocksdb     jdzl  04.32 91776 3.70 0.87 0 20077
Command terminated by signal 11
rocksdb     rpz   00.06 8704 0.00 0.00 0 604
rocksdb     zmlc  04.75 129012 4.16 0.88 0 46085