Add RVV support for RISC-V and optimize decompression speed with an enhanced Memcopy64 function

[anthony-zy]

This PR add RVV support for RV and optimized Memcopy64 function , improving compression speed by ~49.5%.

Optimize Snappy 1.2.2 performance

Added RVV support for RISC-V in Snappy, optimizing Memcopy64 by leveraging RVV vector load/store instructions (e.g., vle8_v_u8m1, vse8_v_u8m1) to reduce memory copy overhead and improve decompression performance. lzbench 2.1 tests onsilesia.tar (GCC 13.2.1, 64-bit Linux) show:

Compressor	Compress	Decompress	Size	Ratio
Snappy 1.2.2 (Before)	64.3 MB/s	186 MB/s	101403263	47.84%
Snappy 1.2.2 (After)	64.4 MB/s	278 MB/s	101403263	47.84%

Test Parameters:

• Compression iterations (cIters): 1
• Decompression iterations (dIters): 1
• Compression time (cTime): 10.0s
• Decompression time (dTime): 10.0s
• Chunk size (chunkSize): 1706MB
• Compression speed limit (cSpeed): 0MB

Notes:

• Compression speed improved by ~49.5% (186 MB/s to 278 MB/s).
• No change in decompression speed, size, or ratio, ensuring
  compatibility.
• Environment: GCC 13.2.1, 64-bit Linux, silesia.tar.

Snappy 1.2.2 unittest ([ PASSED ] 21 tests.)

[==========] Running 21 tests from 3 test suites.
[----------] Global test environment set-up.
[----------] 1 test from CorruptedTest
[ RUN ] CorruptedTest.VerifyCorrupted
Crazy decompression lengths not checked on 64-bit build
[ OK ] CorruptedTest.VerifyCorrupted (14 ms)
[----------] 1 test from CorruptedTest (14 ms total)
[----------] 17 tests from Snappy
[ RUN ] Snappy.SimpleTests
[ OK ] Snappy.SimpleTests (22 ms)
[ RUN ] Snappy.AppendSelfPatternExtensionEdgeCases
[ OK ] Snappy.AppendSelfPatternExtensionEdgeCases (3 ms)
[ RUN ] Snappy.AppendSelfPatternExtensionEdgeCasesExhaustive
[ OK ] Snappy.AppendSelfPatternExtensionEdgeCasesExhaustive (4706 ms)
[ RUN ] Snappy.MaxBlowup
[ OK ] Snappy.MaxBlowup (10 ms)
[ DISABLED ] Snappy.DISABLED_MoreThan4GB
[ RUN ] Snappy.RandomData
[ OK ] Snappy.RandomData (20182 ms)
[ RUN ] Snappy.FourByteOffset
[ OK ] Snappy.FourByteOffset (1 ms)
[ RUN ] Snappy.IOVecSourceEdgeCases
[ OK ] Snappy.IOVecSourceEdgeCases (0 ms)
[ RUN ] Snappy.IOVecSinkEdgeCases
[ OK ] Snappy.IOVecSinkEdgeCases (0 ms)
[ RUN ] Snappy.IOVecLiteralOverflow
[ OK ] Snappy.IOVecLiteralOverflow (0 ms)
[ RUN ] Snappy.IOVecCopyOverflow
[ OK ] Snappy.IOVecCopyOverflow (0 ms)
[ RUN ] Snappy.ReadPastEndOfBuffer
[ OK ] Snappy.ReadPastEndOfBuffer (0 ms)
[ RUN ] Snappy.ZeroOffsetCopy
[ OK ] Snappy.ZeroOffsetCopy (0 ms)
[ RUN ] Snappy.ZeroOffsetCopyValidation
[ OK ] Snappy.ZeroOffsetCopyValidation (0 ms)
[ RUN ] Snappy.FindMatchLength
[ OK ] Snappy.FindMatchLength (0 ms)
[ RUN ] Snappy.FindMatchLengthRandom
[ OK ] Snappy.FindMatchLengthRandom (1089 ms)
[ RUN ] Snappy.VerifyCharTable
[ OK ] Snappy.VerifyCharTable (0 ms)
[ RUN ] Snappy.TestBenchmarkFiles
[ OK ] Snappy.TestBenchmarkFiles (409 ms)
[----------] 17 tests from Snappy (26426 ms total)
[----------] 3 tests from SnappyCorruption
[ RUN ] SnappyCorruption.TruncatedVarint
[ OK ] SnappyCorruption.TruncatedVarint (0 ms)
[ RUN ] SnappyCorruption.UnterminatedVarint
[ OK ] SnappyCorruption.UnterminatedVarint (0 ms)
[ RUN ] SnappyCorruption.OverflowingVarint
[ OK ] SnappyCorruption.OverflowingVarint (0 ms)
[----------] 3 tests from SnappyCorruption (0 ms total)
[----------] Global test environment tear-down
[==========] 21 tests from 3 test suites ran. (26441 ms total)

YOU HAVE 1 DISABLED TEST
[ PASSED ] 21 tests.

[anthony-zy]

Hi @danilak-G ,
All unit tests for Snappy passed successfully. Could you please approve running the CI workflow?

[anthony-zy]

Hi @danilak-G ,
All checks have passed. The CI is green.
Is a code review required before merging? Shall we proceed with the review?

[anthony-zy]

@pwnall @haney @danilak-G , would you pls help to review? Thanks.

[danilak-G]

Hi, thanks for the contribution. We will take a look.

At a first glance, we don't like #if SNAPPY_HAVE_RVV in the middle of the decompression/compression loops and we need to think how to make it more readable. For now I can't provide you with a timeline when we will be able to review this PR

[anthony-zy]

Hi, thanks for the contribution. We will take a look.

At a first glance, we don't like #if SNAPPY_HAVE_RVV in the middle of the decompression/compression loops and we need to think how to make it more readable. For now I can't provide you with a timeline when we will be able to review this PR
thanks

Thanks a lot for the quick feedback!

I moved the #if SNAPPY_HAVE_RVV check to the same level as #if defined(x86_64) && defined(AVX),
turning it into #elif defined(__riscv) && SNAPPY_HAVE_RVV;
the #else branch remains unchanged. @danilak-G

[anthony-zy]

@pwnall @haney @danilak-G , would you pls help to review? Thanks.

Hey, if anyone has a minute, I'd love a review of my code. We're working to improve Snappy's performance on RISC-V—feedback appreciated!

[danilak-G]

Hi, the feedback is still the same. There are multiple # if preprocessor directives in the middle of the decompression loops. This is not bad per se but requires quite rigorous review.

Timeline is still unknown from our side, snappy is mostly in maintenance mode for now