Accelerate SHA256 computations in pure Go for both Intel (AVX2, AVX, SSE) as well as ARM (arm64) platforms.
This package is designed as a drop-in replacement for crypto/sha256. For Intel CPUs it has three flavors for AVX2, AVX and SSE whereby the fastest method is automatically chosen depending on CPU capabilities. For ARM CPUs with the Cryptography Extensions advantage is taken of the SHA2 instructions resulting in a massive performance improvement.
This package uses Golang assembly and as such does not depend on cgo. The Intel versions are based on the implementations as described in "Fast SHA-256 Implementations on Intel Architecture Processors" by J. Guilford et al.
Below is the speed in MB/s for a single core (ranked fast to slow) as well as the factor of improvement over crypto/sha256 (when applicable).
| Processor | Package | Speed | Improvement |
|---|---|---|---|
| 1.2 GHz ARM Cortex-A53 | minio/sha256-simd (ARM64) | 638.2 MB/s | 105x |
| 2.4 GHz Intel Xeon CPU E5-2620 v3 | minio/sha256-simd (AVX2) (*) | 355.0 MB/s | 1.88x |
| 2.4 GHz Intel Xeon CPU E5-2620 v3 | minio/sha256-simd (AVX) | 306.0 MB/s | 1.62x |
| 2.4 GHz Intel Xeon CPU E5-2620 v3 | minio/sha256-simd (SSE) | 298.7 MB/s | 1.58x |
| 2.4 GHz Intel Xeon CPU E5-2620 v3 | crypto/sha256 | 189.2 MB/s | |
| 1.2 GHz ARM Cortex-A53 | crypto/sha256 | 6.1 MB/s |
(*) Measured with the "unrolled"/"demacro-ed" AVX2 version. Due to some Golang assembly restrictions the AVX2 version that uses defines loses about 15% performance. The optimized version is contained in the git history so for maximum speed you want to do this after getting: git cat-file blob 586b6e > sha256blockAvx2_amd64.s (or vendor it for your project; see here to view it in its full glory).
See further down for detailed performance.
As measured on Intel Xeon (same as above) with AVX2 version:
| Method | Package | Speed |
|---|---|---|
| BLAKE2B | minio/blake2b-simd | 851 MB/s |
| MD5 | crypto/md5 | 607 MB/s |
| SHA1 | crypto/sha1 | 522 MB/s |
| SHA256 | minio/sha256-simd | 355 MB/s |
| SHA512 | crypto/sha512 | 306 MB/s |
In order to be able to work more easily with AVX2/AVX instructions, a separate tool was developed to convert AVX2/AVX instructions into the corresponding BYTE sequence as accepted by Go assembly. See asm2plan9s for more information.
One of the most performance sensitive parts of Minio server (object storage server compatible with Amazon S3) is related to SHA256 hash sums calculations. For instance during multi part uploads each part that is uploaded needs to be verified for data integrity by the server. Likewise in order to generated pre-signed URLs check sums must be calculated to ensure their validity.
Other applications that can benefit from enhanced SHA256 performance are deduplication in storage systems, intrusion detection, version control systems, integrity checking, etc.
The 64-bit ARMv8 core has introduced new instructions for SHA1 and SHA2 acceleration as part of the Cryptography Extensions. Below you can see a small excerpt highlighting one of the rounds as is done for the SHA256 calculation process (for full code see sha256block_arm64.s).
sha256h q2, q3, v9.4s
sha256h2 q3, q4, v9.4s
sha256su0 v5.4s, v6.4s
rev32 v8.16b, v8.16b
add v9.4s, v7.4s, v18.4s
mov v4.16b, v2.16b
sha256h q2, q3, v10.4s
sha256h2 q3, q4, v10.4s
sha256su0 v6.4s, v7.4s
sha256su1 v5.4s, v7.4s, v8.4s
Benchmarks generated on a 1.2 Ghz Quad-Core ARM Cortex A53 equipped Pine64.
minio@minio-arm:~/gopath/src/github.com/sha256-simd$ benchcmp golang.txt arm64.txt
benchmark old ns/op new ns/op delta
BenchmarkHash8Bytes-4 11836 1403 -88.15%
BenchmarkHash1K-4 181143 3138 -98.27%
BenchmarkHash8K-4 1365652 14356 -98.95%
BenchmarkHash1M-4 173192200 1642954 -99.05%
benchmark old MB/s new MB/s speedup
BenchmarkHash8Bytes-4 0.68 5.70 8.38x
BenchmarkHash1K-4 5.65 326.30 57.75x
BenchmarkHash8K-4 6.00 570.63 95.11x
BenchmarkHash1M-4 6.05 638.23 105.49x
Example performance metrics were generated on Intel(R) Xeon(R) CPU E5-2620 v3 @ 2.40GHz - 6 physical cores, 12 logical cores running Ubuntu GNU/Linux with kernel version 4.4.0-24-generic (vanilla with no optimizations).
$ benchcmp go.txt avx2.txt
benchmark old ns/op new ns/op delta
BenchmarkHash8Bytes-12 446 364 -18.39%
BenchmarkHash1K-12 5919 3279 -44.60%
BenchmarkHash8K-12 43791 23655 -45.98%
BenchmarkHash1M-12 5544989 2969305 -46.45%
benchmark old MB/s new MB/s speedup
BenchmarkHash8Bytes-12 17.93 21.96 1.22x
BenchmarkHash1K-12 172.98 312.27 1.81x
BenchmarkHash8K-12 187.07 346.31 1.85x
BenchmarkHash1M-12 189.10 353.14 1.87x
$ benchcmp go.txt avx.txt
benchmark old ns/op new ns/op delta
BenchmarkHash8Bytes-12 446 346 -22.42%
BenchmarkHash1K-12 5919 3701 -37.47%
BenchmarkHash8K-12 43791 27222 -37.84%
BenchmarkHash1M-12 5544989 3426938 -38.20%
benchmark old MB/s new MB/s speedup
BenchmarkHash8Bytes-12 17.93 23.06 1.29x
BenchmarkHash1K-12 172.98 276.64 1.60x
BenchmarkHash8K-12 187.07 300.93 1.61x
BenchmarkHash1M-12 189.10 305.98 1.62x
$ benchcmp go.txt sse.txt
benchmark old ns/op new ns/op delta
BenchmarkHash8Bytes-12 446 362 -18.83%
BenchmarkHash1K-12 5919 3751 -36.63%
BenchmarkHash8K-12 43791 27396 -37.44%
BenchmarkHash1M-12 5544989 3444623 -37.88%
benchmark old MB/s new MB/s speedup
BenchmarkHash8Bytes-12 17.93 22.05 1.23x
BenchmarkHash1K-12 172.98 272.92 1.58x
BenchmarkHash8K-12 187.07 299.01 1.60x
BenchmarkHash1M-12 189.10 304.41 1.61x
Released under the Apache License v2.0. You can find the complete text in the file LICENSE.
Contributions are welcome, please send PRs for any enhancements.