Using BoringSSL's SHA1 implementation
The portable implementation in base is nicely written and tested but could be made way faster by using SIMD and crypto instructions. Enters BoringSSL: its SHA1 implementation is highly optimized for many architectures, actively maintained and tested. It will make SHA1 *way* faster on Intel (6x), ARMv7 (8x) and ARMv8 (between 16x to 21x faster). Plus adds a performance benchmark to measure the speed (MB/s) of the hash implementation. Unfortunately, thanks to NaCL we still got keep the portable implementation around for the time being (until it is finally removed). Change-Id: Ifc4b6a82a0d36e5157b1fa6d94760c32197b570c Bug: 974341 Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1661110Reviewed-by:Adenilson Cavalcanti <cavalcantii@chromium.org> Reviewed-by:
Daniel Cheng <dcheng@chromium.org> Reviewed-by:
David Benjamin <davidben@chromium.org> Commit-Queue: Adenilson Cavalcanti <cavalcantii@chromium.org> Cr-Commit-Position: refs/heads/master@{#678434}
Showing
base/hash/sha1_boringssl.cc
0 → 100644
base/hash/sha1_perftest.cc
0 → 100644
Please register or sign in to comment