Minimizing run time

[ivanstepanovftw]

I have just for fun tried to minimize run time (prompt eval + eval) using Bayesian optimization using different compilation flags and run arguments and got this

CPU: 8 cores, 16 threads, flags:

Details

fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush mmx fxsr sse sse2 ht syscall nx mmxext fxsr_opt pdpe1gb rdtscp lm constant_tsc rep_good nopl nonstop_tsc cpuid extd_apicid aperfmperf rapl pni pclmulqdq monitor ssse3 fma cx16 sse4_1 sse4_2 movbe popcnt aes xsave avx f16c rdrand lahf_lm cmp_legacy svm extapic cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw ibs skinit wdt tce topoext perfctr_core perfctr_nb bpext perfctr_llc mwaitx cpb cat_l3 cdp_l3 hw_pstate ssbd mba ibrs ibpb stibp vmmcall fsgsbase bmi1 avx2 smep bmi2 erms invpcid cqm rdt_a rdseed adx smap clflushopt clwb sha_ni xsaveopt xsavec xgetbv1 xsaves cqm_llc cqm_occup_llc cqm_mbm_total cqm_mbm_local clzero irperf xsaveerptr rdpru wbnoinvd cppc arat npt lbrv svm_lock nrip_save tsc_scale vmcb_clean flushbyasid decodeassists pausefilter pfthreshold avic v_vmsave_vmload vgif v_spec_ctrl umip pku ospke vaes vpclmulqdq rdpid overflow_recov succor smca fsrm

Script: rob.py.txt

The script is 90%+10% mix of GPT-4 and Copilot, so bear with me ;)

My output:

Best options found: ['-DLLAMA_LTO=OFF', '-DLLAMA_AVX=OFF', '-DLLAMA_AVX2=ON', '-DLLAMA_FMA=ON', '-DLLAMA_OPENBLAS=OFF', '', '', '--batch_size 32', '', '--threads 8']
Minimum runtime: 47.2552 seconds
Expected minimum: ['-DLLAMA_LTO=OFF', '-DLLAMA_AVX=OFF', '-DLLAMA_AVX2=OFF', '-DLLAMA_FMA=OFF', '-DLLAMA_OPENBLAS=OFF', '', '--no-mmap', '--batch_size 32', '', '--threads 8']

If you wanted to try it, modify n_calls=0, n_initial_points=0 to something like 1 and 1, then 20 and 10 resp. Note that results are saved after full minimization.

See also Inference quality discussion, where I tried to get best quality predictions.

[ivanstepanovftw]

Seems like 4 threads is more optimal than specified in search space