diff options
| author | R?my Oudompheng <oudomphe@phare.normalesup.org> | 2013-08-09 06:43:17 +0200 |
|---|---|---|
| committer | R?my Oudompheng <oudomphe@phare.normalesup.org> | 2013-08-09 06:43:17 +0200 |
| commit | b6d637e7221a0f6d2d636b9ae021522b06764fec (patch) | |
| tree | 23b8bf8713bbfecab98a67251e8161d69e5a765a /src/cmd/5l | |
| parent | ef6b8ab4219898596b935884bf14c46507b1ea2e (diff) | |
| download | go-b6d637e7221a0f6d2d636b9ae021522b06764fec.tar.gz | |
cmd/5c, cmd/5g, cmd/5l: turn MOVB, MOVH into plain moves, optimize short arithmetic.
Pseudo-instructions MOVBS and MOVHS are used to clarify
the semantics of short integers vs. registers:
* 8-bit and 16-bit values in registers are assumed to always
be zero-extended or sign-extended depending on their type.
* MOVB is truncation or move of an already extended value
between registers.
* MOVBU enforces zero-extension at the destination (register).
* MOVBS enforces sign-extension at the destination (register).
And similarly for MOVH/MOVS/MOVHU.
The linker is adapted to assemble MOVB and MOVH to an ordinary
mov. Also a peephole pass in 5g that aims at eliminating
redundant zero/sign extensions is improved.
encoding/binary:
benchmark old ns/op new ns/op delta
BenchmarkReadSlice1000Int32s 220387 217185 -1.45%
BenchmarkReadStruct 12839 12910 +0.55%
BenchmarkReadInts 5692 5534 -2.78%
BenchmarkWriteInts 6137 6016 -1.97%
BenchmarkPutUvarint32 257 241 -6.23%
BenchmarkPutUvarint64 812 754 -7.14%
benchmark old MB/s new MB/s speedup
BenchmarkReadSlice1000Int32s 18.15 18.42 1.01x
BenchmarkReadStruct 5.45 5.42 0.99x
BenchmarkReadInts 5.27 5.42 1.03x
BenchmarkWriteInts 4.89 4.99 1.02x
BenchmarkPutUvarint32 15.56 16.57 1.06x
BenchmarkPutUvarint64 9.85 10.60 1.08x
crypto/des:
benchmark old ns/op new ns/op delta
BenchmarkEncrypt 7002 5169 -26.18%
BenchmarkDecrypt 7015 5195 -25.94%
benchmark old MB/s new MB/s speedup
BenchmarkEncrypt 1.14 1.55 1.36x
BenchmarkDecrypt 1.14 1.54 1.35x
strconv:
benchmark old ns/op new ns/op delta
BenchmarkAtof64Decimal 457 385 -15.75%
BenchmarkAtof64Float 574 479 -16.55%
BenchmarkAtof64FloatExp 1035 906 -12.46%
BenchmarkAtof64Big 1793 1457 -18.74%
BenchmarkAtof64RandomBits 2267 2066 -8.87%
BenchmarkAtof64RandomFloats 1416 1194 -15.68%
BenchmarkAtof32Decimal 451 379 -15.96%
BenchmarkAtof32Float 547 435 -20.48%
BenchmarkAtof32FloatExp 1095 986 -9.95%
BenchmarkAtof32Random 1154 1006 -12.82%
BenchmarkAtoi 1415 1380 -2.47%
BenchmarkAtoiNeg 1414 1401 -0.92%
BenchmarkAtoi64 1744 1671 -4.19%
BenchmarkAtoi64Neg 1737 1662 -4.32%
Fixes issue 1837.
R=rsc, dave, bradfitz
CC=golang-dev
https://codereview.appspot.com/12424043
Diffstat (limited to 'src/cmd/5l')
| -rw-r--r-- | src/cmd/5l/asm.c | 2 | ||||
| -rw-r--r-- | src/cmd/5l/optab.c | 4 |
2 files changed, 4 insertions, 2 deletions
diff --git a/src/cmd/5l/asm.c b/src/cmd/5l/asm.c index 92296b5bc..28bb40682 100644 --- a/src/cmd/5l/asm.c +++ b/src/cmd/5l/asm.c @@ -1639,6 +1639,8 @@ oprrr(int a, int sc) case ACMP: return o | (0xa<<21) | (1<<20); case ACMN: return o | (0xb<<21) | (1<<20); case AORR: return o | (0xc<<21); + case AMOVB: + case AMOVH: case AMOVW: return o | (0xd<<21); case ABIC: return o | (0xe<<21); case AMVN: return o | (0xf<<21); diff --git a/src/cmd/5l/optab.c b/src/cmd/5l/optab.c index dfd93f31d..dc9e5e99f 100644 --- a/src/cmd/5l/optab.c +++ b/src/cmd/5l/optab.c @@ -94,10 +94,10 @@ Optab optab[] = { AMVN, C_LCON, C_NONE, C_REG, 13, 8, 0, LFROM }, { ACMP, C_LCON, C_REG, C_NONE, 13, 8, 0, LFROM }, - { AMOVB, C_REG, C_NONE, C_REG, 14, 8, 0 }, + { AMOVB, C_REG, C_NONE, C_REG, 1, 4, 0 }, { AMOVBS, C_REG, C_NONE, C_REG, 14, 8, 0 }, { AMOVBU, C_REG, C_NONE, C_REG, 58, 4, 0 }, - { AMOVH, C_REG, C_NONE, C_REG, 14, 8, 0 }, + { AMOVH, C_REG, C_NONE, C_REG, 1, 4, 0 }, { AMOVHS, C_REG, C_NONE, C_REG, 14, 8, 0 }, { AMOVHU, C_REG, C_NONE, C_REG, 14, 8, 0 }, |