performance: sha2 is slower than OpenSSL provided one

[oech3]

Problem:

I used aws-lc-sys via openssl crate's aws-lc feature on uutils/coreutils. It uses statically linked latest OpenSSL by default, but it can be replaced with aws-lc by editing
src/uucore/Cargo.toml . I built cksum with cargo build --release -p uu_cksum.

As a result, aws-lc was faster than OpenSSL with md5 and sha1. However, sha2 was slower than OpenSSL.
uutils/coreutils#14563 (comment)

[torben-hansen]

Thanks for the data.

I don't see the same in our benchmark data where we measuring directly on the public api for the sha-algorithm primitive. Which hardware is it?

[oech3]

Model name:                Intel(R) Core(TM) i5-4260U CPU @ 1.40GHz

    Flags:                   fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca c
                             mov pat pse36 clflush dts acpi mmx fxsr sse sse2 ss ht tm 
                             pbe syscall nx pdpe1gb rdtscp lm constant_tsc arch_perfmon
                              pebs bts rep_good nopl xtopology nonstop_tsc cpuid aperfm
                             perf pni pclmulqdq dtes64 monitor ds_cpl vmx est tm2 ssse3
                              sdbg fma cx16 xtpr pdcm pcid sse4_1 sse4_2 x2apic movbe p
                             opcnt tsc_deadline_timer aes xsave avx f16c rdrand lahf_lm
                              abm cpuid_fault epb pti ssbd ibrs ibpb stibp tpr_shadow f
                             lexpriority ept vpid ept_ad fsgsbase tsc_adjust bmi1 avx2 
                             smep bmi2 erms invpcid xsaveopt dtherm ida arat pln pts vn
                             mi md_clear flush_l1d

[oech3]

with newer CPU:

  Model name:                AMD Ryzen 5 5600G with Radeon Graphics
    Flags:                   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 xtopology nonstop_tsc cpuid extd_apicid ape
                             rfmperf rapl pni pclmulqdq monitor ssse3 fma cx16 sse4_1 sse4_2 x2apic movbe popcnt aes xsave avx f16c rdran
                             d lahf_lm cmp_legacy svm extapic cr8_legacy abm sse4a misalignsse 3dnowprefetch osvw ibs skinit wdt tce topo
                             ext perfctr_core perfctr_nb bpext perfctr_llc mwaitx cpb cat_l3 cdp_l3 hw_pstate ssbd mba ibrs ibpb stibp vm
                             mcall fsgsbase bmi1 avx2 smep bmi2 erms invpcid cqm rdt_a rdseed adx smap clflushopt clwb sha_ni xsaveopt xs
                             avec xgetbv1 xsaves cqm_llc cqm_occup_llc cqm_mbm_total cqm_mbm_local user_shstk clzero irperf xsaveerptr rd
                             pru wbnoinvd cppc arat npt lbrv svm_lock nrip_save tsc_scale vmcb_clean flushbyasid decodeassists pausefilte
                             r pfthreshold avic v_vmsave_vmload vgif v_spec_ctrl umip pku ospke vaes vpclmulqdq rdpid overflow_recov succ
                             or smca fsrm debug_swap

I still see large performance drop

> hyperfine -N "target/release/cksum-ossl-static -a sha512 /lib/chromium/chromium" "target/release/cksum-
aws-lc -a sha512 /lib/chromium/chromium"
Benchmark 1: target/release/cksum-ossl-static -a sha512 /lib/chromium/chromium
  Time (mean ± σ):     347.1 ms ±   0.9 ms    [User: 326.7 ms, System: 19.2 ms]
  Range (min … max):   346.3 ms … 349.4 ms    10 runs
 
Benchmark 2: target/release/cksum-aws-lc -a sha512 /lib/chromium/chromium
  Time (mean ± σ):     478.8 ms ±   1.1 ms    [User: 458.1 ms, System: 19.1 ms]
  Range (min … max):   477.4 ms … 480.8 ms    10 runs
 
Summary
  target/release/cksum-ossl-static -a sha512 /lib/chromium/chromium ran
    1.38 ± 0.00 times faster than target/release/cksum-aws-lc -a sha512 /lib/chromium/chromium