| Age | Commit message (Collapse) | Author | Files | Lines |
|---|
|
* cipher/sha512-armv7-neon.S
(_gcry_sha512_transform_armv7_neon): Byte-swap RW67q and RW1011q
correctly in multi-block loop.
* tests/basic.c (check_digests): Add large test vector for SHA512.
--
Patch fixes bug introduced to multi-block processing by commit df629ba53a6,
"Improve performance of SHA-512/ARM/NEON implementation". Patch also adds
multi-block test vector for SHA-512.
Signed-off-by: Jussi Kivilinna <jussi.kivilinna@iki.fi>
|
|
cipher/blowfish-amd64.S: Change utf-8 encoded copyright character to
'(C)'.
cipher/blowfish-arm.S: Ditto.
cipher/bufhelp.h: Ditto.
cipher/camellia-aesni-avx-amd64.S: Ditto.
cipher/camellia-aesni-avx2-amd64.S: Ditto.
cipher/camellia-arm.S: Ditto.
cipher/cast5-amd64.S: Ditto.
cipher/cast5-arm.S: Ditto.
cipher/cipher-ccm.c: Ditto.
cipher/cipher-cmac.c: Ditto.
cipher/cipher-gcm.c: Ditto.
cipher/cipher-selftest.c: Ditto.
cipher/cipher-selftest.h: Ditto.
cipher/mac-cmac.c: Ditto.
cipher/mac-gmac.c: Ditto.
cipher/mac-hmac.c: Ditto.
cipher/mac-internal.h: Ditto.
cipher/mac.c: Ditto.
cipher/rijndael-amd64.S: Ditto.
cipher/rijndael-arm.S: Ditto.
cipher/salsa20-amd64.S: Ditto.
cipher/salsa20-armv7-neon.S: Ditto.
cipher/serpent-armv7-neon.S: Ditto.
cipher/serpent-avx2-amd64.S: Ditto.
cipher/serpent-sse2-amd64.S: Ditto.
--
Avoid use of '©' for easier parsing of source for copyright information.
Signed-off-by: Jussi Kivilinna <jussi.kivilinna@iki.fi>
|
|
* cipher/sha512-armv7-neon.S (RT01q, RT23q, RT45q, RT67q): New.
(round_0_63, round_64_79): Remove.
(rounds2_0_63, rounds2_64_79): New.
(_gcry_sha512_transform_armv7_neon): Add 'nblks' input; Handle multiple
input blocks; Use new round macros.
* cipher/sha512.c [USE_ARM_NEON_ASM]
(_gcry_sha512_transform_armv7_neon): Add 'num_blks'.
(transform) [USE_ARM_NEON_ASM]: Pass nblks to assembly.
--
Benchmarks on ARM Cortex-A8:
C-language: 139.1 c/B
Old ARM/NEON: 34.30 c/B
New ARM/NEON: 24.46 c/B
New vs C: 5.68x
New vs Old: 1.40x
Signed-off-by: Jussi Kivilinna <jussi.kivilinna@iki.fi>
|
|
* cipher/Makefile.am: Add 'sha512-armv7-neon.S'.
* cipher/sha512-armv7-neon.S: New file.
* cipher/sha512.c (USE_ARM_NEON_ASM): New macro.
(SHA512_CONTEXT) [USE_ARM_NEON_ASM]: Add 'use_neon'.
(sha512_init, sha384_init) [USE_ARM_NEON_ASM]: Enable 'use_neon' if
CPU support NEON instructions.
(k): Round constant array moved outside of 'transform' function.
(__transform): Renamed from 'tranform' function.
[USE_ARM_NEON_ASM] (_gcry_sha512_transform_armv7_neon): New prototype.
(transform): New wrapper function for different transform versions.
(sha512_write, sha512_final): Burn stack by the amount returned by
transform function.
* configure.ac (sha512) [neonsupport]: Add 'sha512-armv7-neon.lo'.
--
Add NEON assembly for transform function for faster SHA512 on ARM. Major speed
up thanks to 64-bit integer registers and large register file that can hold
full input buffer.
Benchmark results on Cortex-A8, 1Ghz:
Old:
$ tests/benchmark --hash-repetitions 100 md sha512 sha384
SHA512 17050ms 18780ms 29120ms 18040ms 17190ms
SHA384 17130ms 18720ms 29160ms 18090ms 17280ms
New:
$ tests/benchmark --hash-repetitions 100 md sha512 sha384
SHA512 3600ms 5070ms 15330ms 4510ms 3480ms
SHA384 3590ms 5060ms 15350ms 4510ms 3520ms
New vs old:
SHA512 4.74x 3.70x 1.90x 4.00x 4.94x
SHA384 4.77x 3.70x 1.90x 4.01x 4.91x
Signed-off-by: Jussi Kivilinna <jussi.kivilinna@iki.fi>
|
