mirror of https://github.com/golang/go.git
cmd/compile: optimize arm64's MADD and MSUB
This CL implements constant folding for MADD/MSUB on arm64. 1. The total size of pkg/android_arm64/ decreases about 4KB, excluding cmd/compile/ . 2. There is no regression in the go1 benchmark, excluding noise. name old time/op new time/op delta BinaryTree17-4 16.4s ± 1% 16.5s ± 1% +0.24% (p=0.008 n=29+29) Fannkuch11-4 8.73s ± 0% 8.71s ± 0% -0.15% (p=0.000 n=29+29) FmtFprintfEmpty-4 174ns ± 0% 174ns ± 0% ~ (all equal) FmtFprintfString-4 370ns ± 0% 372ns ± 2% +0.53% (p=0.007 n=24+30) FmtFprintfInt-4 419ns ± 0% 419ns ± 0% ~ (all equal) FmtFprintfIntInt-4 673ns ± 1% 661ns ± 1% -1.81% (p=0.000 n=30+27) FmtFprintfPrefixedInt-4 806ns ± 0% 805ns ± 0% ~ (p=0.957 n=28+27) FmtFprintfFloat-4 1.09µs ± 0% 1.09µs ± 0% -0.04% (p=0.001 n=22+30) FmtManyArgs-4 2.67µs ± 0% 2.68µs ± 0% +0.03% (p=0.045 n=29+28) GobDecode-4 33.2ms ± 1% 32.5ms ± 1% -2.11% (p=0.000 n=29+29) GobEncode-4 29.5ms ± 0% 29.2ms ± 0% -1.04% (p=0.000 n=28+28) Gzip-4 1.39s ± 2% 1.38s ± 1% -0.48% (p=0.023 n=30+30) Gunzip-4 139ms ± 0% 139ms ± 0% ~ (p=0.616 n=30+28) HTTPClientServer-4 766µs ± 4% 758µs ± 3% -1.03% (p=0.013 n=28+29) JSONEncode-4 49.7ms ± 0% 49.6ms ± 0% -0.24% (p=0.000 n=30+30) JSONDecode-4 266ms ± 0% 268ms ± 1% +1.07% (p=0.000 n=29+30) Mandelbrot200-4 16.6ms ± 0% 16.6ms ± 0% ~ (p=0.248 n=30+29) GoParse-4 15.9ms ± 0% 16.0ms ± 0% +0.76% (p=0.000 n=29+29) RegexpMatchEasy0_32-4 381ns ± 0% 380ns ± 0% -0.14% (p=0.000 n=30+30) RegexpMatchEasy0_1K-4 1.18µs ± 0% 1.19µs ± 1% +0.30% (p=0.000 n=29+30) RegexpMatchEasy1_32-4 357ns ± 0% 357ns ± 0% ~ (all equal) RegexpMatchEasy1_1K-4 2.04µs ± 0% 2.05µs ± 0% +0.50% (p=0.000 n=26+28) RegexpMatchMedium_32-4 590ns ± 0% 589ns ± 0% -0.12% (p=0.000 n=30+23) RegexpMatchMedium_1K-4 162µs ± 0% 162µs ± 0% ~ (p=0.318 n=28+25) RegexpMatchHard_32-4 9.56µs ± 0% 9.56µs ± 0% ~ (p=0.072 n=30+29) RegexpMatchHard_1K-4 287µs ± 0% 287µs ± 0% -0.02% (p=0.005 n=28+28) Revcomp-4 2.50s ± 0% 2.51s ± 0% ~ (p=0.246 n=29+29) Template-4 312ms ± 1% 313ms ± 1% +0.46% (p=0.002 n=30+30) TimeParse-4 1.68µs ± 0% 1.67µs ± 0% -0.31% (p=0.000 n=27+29) TimeFormat-4 1.66µs ± 0% 1.64µs ± 0% -0.92% (p=0.000 n=29+26) [Geo mean] 247µs 246µs -0.15% name old speed new speed delta GobDecode-4 23.1MB/s ± 1% 23.6MB/s ± 0% +2.17% (p=0.000 n=29+28) GobEncode-4 26.0MB/s ± 0% 26.3MB/s ± 0% +1.05% (p=0.000 n=28+28) Gzip-4 14.0MB/s ± 2% 14.1MB/s ± 1% +0.47% (p=0.026 n=30+30) Gunzip-4 139MB/s ± 0% 139MB/s ± 0% ~ (p=0.624 n=30+28) JSONEncode-4 39.1MB/s ± 0% 39.2MB/s ± 0% +0.24% (p=0.000 n=30+30) JSONDecode-4 7.31MB/s ± 0% 7.23MB/s ± 1% -1.07% (p=0.000 n=28+30) GoParse-4 3.65MB/s ± 0% 3.62MB/s ± 0% -0.77% (p=0.000 n=29+29) RegexpMatchEasy0_32-4 84.0MB/s ± 0% 84.1MB/s ± 0% +0.18% (p=0.000 n=28+30) RegexpMatchEasy0_1K-4 864MB/s ± 0% 861MB/s ± 1% -0.29% (p=0.000 n=29+30) RegexpMatchEasy1_32-4 89.5MB/s ± 0% 89.5MB/s ± 0% ~ (p=0.841 n=28+28) RegexpMatchEasy1_1K-4 502MB/s ± 0% 500MB/s ± 0% -0.51% (p=0.000 n=29+29) RegexpMatchMedium_32-4 1.69MB/s ± 0% 1.70MB/s ± 0% +0.41% (p=0.000 n=26+30) RegexpMatchMedium_1K-4 6.31MB/s ± 0% 6.30MB/s ± 0% ~ (p=0.129 n=30+25) RegexpMatchHard_32-4 3.35MB/s ± 0% 3.35MB/s ± 0% ~ (p=0.657 n=30+29) RegexpMatchHard_1K-4 3.57MB/s ± 0% 3.57MB/s ± 0% ~ (all equal) Revcomp-4 102MB/s ± 0% 101MB/s ± 0% ~ (p=0.213 n=29+29) Template-4 6.22MB/s ± 1% 6.19MB/s ± 1% -0.42% (p=0.005 n=30+29) [Geo mean] 24.1MB/s 24.2MB/s +0.08% Change-Id: I6c02d3c9975f6bd8bc215cb1fc14d29602b45649 Reviewed-on: https://go-review.googlesource.com/138095 Run-TryBot: Ben Shi <powerman1st@163.com> TryBot-Result: Gobot Gobot <gobot@golang.org> Reviewed-by: Cherry Zhang <cherryyz@google.com>
This commit is contained in:
Ben Shi
Cherry Zhang
parent
bf8e6b7027
commit
d60cf39f8e
|
|
@ -1154,15 +1154,15 @@
|
||||||
(MULW (NEG x) y) -> (MNEGW x y)
|
(MULW (NEG x) y) -> (MNEGW x y)
|
||||||
|
|
||||||
// madd/msub
|
// madd/msub
|
||||||
(ADD a l:(MUL x y)) && l.Uses==1 && x.Op!=OpARM64MOVDconst && y.Op!=OpARM64MOVDconst && a.Op!=OpARM64MOVDconst && clobber(l) -> (MADD a x y)
|
(ADD a l:(MUL x y)) && l.Uses==1 && clobber(l) -> (MADD a x y)
|
||||||
(SUB a l:(MUL x y)) && l.Uses==1 && x.Op!=OpARM64MOVDconst && y.Op!=OpARM64MOVDconst && a.Op!=OpARM64MOVDconst && clobber(l) -> (MSUB a x y)
|
(SUB a l:(MUL x y)) && l.Uses==1 && clobber(l) -> (MSUB a x y)
|
||||||
(ADD a l:(MNEG x y)) && l.Uses==1 && x.Op!=OpARM64MOVDconst && y.Op!=OpARM64MOVDconst && a.Op!=OpARM64MOVDconst && clobber(l) -> (MSUB a x y)
|
(ADD a l:(MNEG x y)) && l.Uses==1 && clobber(l) -> (MSUB a x y)
|
||||||
(SUB a l:(MNEG x y)) && l.Uses==1 && x.Op!=OpARM64MOVDconst && y.Op!=OpARM64MOVDconst && a.Op!=OpARM64MOVDconst && clobber(l) -> (MADD a x y)
|
(SUB a l:(MNEG x y)) && l.Uses==1 && clobber(l) -> (MADD a x y)
|
||||||
|
|
||||||
(ADD a l:(MULW x y)) && l.Uses==1 && a.Type.Size() != 8 && x.Op!=OpARM64MOVDconst && y.Op!=OpARM64MOVDconst && a.Op!=OpARM64MOVDconst && clobber(l) -> (MADDW a x y)
|
(ADD a l:(MULW x y)) && a.Type.Size() != 8 && l.Uses==1 && clobber(l) -> (MADDW a x y)
|
||||||
(SUB a l:(MULW x y)) && l.Uses==1 && a.Type.Size() != 8 && x.Op!=OpARM64MOVDconst && y.Op!=OpARM64MOVDconst && a.Op!=OpARM64MOVDconst && clobber(l) -> (MSUBW a x y)
|
(SUB a l:(MULW x y)) && a.Type.Size() != 8 && l.Uses==1 && clobber(l) -> (MSUBW a x y)
|
||||||
(ADD a l:(MNEGW x y)) && l.Uses==1 && a.Type.Size() != 8 && x.Op!=OpARM64MOVDconst && y.Op!=OpARM64MOVDconst && a.Op!=OpARM64MOVDconst && clobber(l) -> (MSUBW a x y)
|
(ADD a l:(MNEGW x y)) && a.Type.Size() != 8 && l.Uses==1 && clobber(l) -> (MSUBW a x y)
|
||||||
(SUB a l:(MNEGW x y)) && l.Uses==1 && a.Type.Size() != 8 && x.Op!=OpARM64MOVDconst && y.Op!=OpARM64MOVDconst && a.Op!=OpARM64MOVDconst && clobber(l) -> (MADDW a x y)
|
(SUB a l:(MNEGW x y)) && a.Type.Size() != 8 && l.Uses==1 && clobber(l) -> (MADDW a x y)
|
||||||
|
|
||||||
// mul by constant
|
// mul by constant
|
||||||
(MUL x (MOVDconst [-1])) -> (NEG x)
|
(MUL x (MOVDconst [-1])) -> (NEG x)
|
||||||
|
|
@ -1210,6 +1210,94 @@
|
||||||
(MNEGW x (MOVDconst [c])) && c%7 == 0 && isPowerOfTwo(c/7) && is32Bit(c) -> (SLLconst <x.Type> [log2(c/7)] (SUBshiftLL <x.Type> x x [3]))
|
(MNEGW x (MOVDconst [c])) && c%7 == 0 && isPowerOfTwo(c/7) && is32Bit(c) -> (SLLconst <x.Type> [log2(c/7)] (SUBshiftLL <x.Type> x x [3]))
|
||||||
(MNEGW x (MOVDconst [c])) && c%9 == 0 && isPowerOfTwo(c/9) && is32Bit(c) -> (NEG (SLLconst <x.Type> [log2(c/9)] (ADDshiftLL <x.Type> x x [3])))
|
(MNEGW x (MOVDconst [c])) && c%9 == 0 && isPowerOfTwo(c/9) && is32Bit(c) -> (NEG (SLLconst <x.Type> [log2(c/9)] (ADDshiftLL <x.Type> x x [3])))
|
||||||
|
|
||||||
|
(MADD a x (MOVDconst [-1])) -> (SUB a x)
|
||||||
|
(MADD a _ (MOVDconst [0])) -> a
|
||||||
|
(MADD a x (MOVDconst [1])) -> (ADD a x)
|
||||||
|
(MADD a x (MOVDconst [c])) && isPowerOfTwo(c) -> (ADDshiftLL a x [log2(c)])
|
||||||
|
(MADD a x (MOVDconst [c])) && isPowerOfTwo(c-1) && c>=3 -> (ADD a (ADDshiftLL <x.Type> x x [log2(c-1)]))
|
||||||
|
(MADD a x (MOVDconst [c])) && isPowerOfTwo(c+1) && c>=7 -> (SUB a (SUBshiftLL <x.Type> x x [log2(c+1)]))
|
||||||
|
(MADD a x (MOVDconst [c])) && c%3 == 0 && isPowerOfTwo(c/3) -> (SUBshiftLL a (SUBshiftLL <x.Type> x x [2]) [log2(c/3)])
|
||||||
|
(MADD a x (MOVDconst [c])) && c%5 == 0 && isPowerOfTwo(c/5) -> (ADDshiftLL a (ADDshiftLL <x.Type> x x [2]) [log2(c/5)])
|
||||||
|
(MADD a x (MOVDconst [c])) && c%7 == 0 && isPowerOfTwo(c/7) -> (SUBshiftLL a (SUBshiftLL <x.Type> x x [3]) [log2(c/7)])
|
||||||
|
(MADD a x (MOVDconst [c])) && c%9 == 0 && isPowerOfTwo(c/9) -> (ADDshiftLL a (ADDshiftLL <x.Type> x x [3]) [log2(c/9)])
|
||||||
|
|
||||||
|
(MADD a (MOVDconst [-1]) x) -> (SUB a x)
|
||||||
|
(MADD a (MOVDconst [0]) _) -> a
|
||||||
|
(MADD a (MOVDconst [1]) x) -> (ADD a x)
|
||||||
|
(MADD a (MOVDconst [c]) x) && isPowerOfTwo(c) -> (ADDshiftLL a x [log2(c)])
|
||||||
|
(MADD a (MOVDconst [c]) x) && isPowerOfTwo(c-1) && c>=3 -> (ADD a (ADDshiftLL <x.Type> x x [log2(c-1)]))
|
||||||
|
(MADD a (MOVDconst [c]) x) && isPowerOfTwo(c+1) && c>=7 -> (SUB a (SUBshiftLL <x.Type> x x [log2(c+1)]))
|
||||||
|
(MADD a (MOVDconst [c]) x) && c%3 == 0 && isPowerOfTwo(c/3) -> (SUBshiftLL a (SUBshiftLL <x.Type> x x [2]) [log2(c/3)])
|
||||||
|
(MADD a (MOVDconst [c]) x) && c%5 == 0 && isPowerOfTwo(c/5) -> (ADDshiftLL a (ADDshiftLL <x.Type> x x [2]) [log2(c/5)])
|
||||||
|
(MADD a (MOVDconst [c]) x) && c%7 == 0 && isPowerOfTwo(c/7) -> (SUBshiftLL a (SUBshiftLL <x.Type> x x [3]) [log2(c/7)])
|
||||||
|
(MADD a (MOVDconst [c]) x) && c%9 == 0 && isPowerOfTwo(c/9) -> (ADDshiftLL a (ADDshiftLL <x.Type> x x [3]) [log2(c/9)])
|
||||||
|
|
||||||
|
(MADDW a x (MOVDconst [c])) && int32(c)==-1 -> (SUB a x)
|
||||||
|
(MADDW a _ (MOVDconst [c])) && int32(c)==0 -> a
|
||||||
|
(MADDW a x (MOVDconst [c])) && int32(c)==1 -> (ADD a x)
|
||||||
|
(MADDW a x (MOVDconst [c])) && isPowerOfTwo(c) -> (ADDshiftLL a x [log2(c)])
|
||||||
|
(MADDW a x (MOVDconst [c])) && isPowerOfTwo(c-1) && int32(c)>=3 -> (ADD a (ADDshiftLL <x.Type> x x [log2(c-1)]))
|
||||||
|
(MADDW a x (MOVDconst [c])) && isPowerOfTwo(c+1) && int32(c)>=7 -> (SUB a (SUBshiftLL <x.Type> x x [log2(c+1)]))
|
||||||
|
(MADDW a x (MOVDconst [c])) && c%3 == 0 && isPowerOfTwo(c/3) && is32Bit(c) -> (SUBshiftLL a (SUBshiftLL <x.Type> x x [2]) [log2(c/3)])
|
||||||
|
(MADDW a x (MOVDconst [c])) && c%5 == 0 && isPowerOfTwo(c/5) && is32Bit(c) -> (ADDshiftLL a (ADDshiftLL <x.Type> x x [2]) [log2(c/5)])
|
||||||
|
(MADDW a x (MOVDconst [c])) && c%7 == 0 && isPowerOfTwo(c/7) && is32Bit(c) -> (SUBshiftLL a (SUBshiftLL <x.Type> x x [3]) [log2(c/7)])
|
||||||
|
(MADDW a x (MOVDconst [c])) && c%9 == 0 && isPowerOfTwo(c/9) && is32Bit(c) -> (ADDshiftLL a (ADDshiftLL <x.Type> x x [3]) [log2(c/9)])
|
||||||
|
|
||||||
|
(MADDW a (MOVDconst [c]) x) && int32(c)==-1 -> (SUB a x)
|
||||||
|
(MADDW a (MOVDconst [c]) _) && int32(c)==0 -> a
|
||||||
|
(MADDW a (MOVDconst [c]) x) && int32(c)==1 -> (ADD a x)
|
||||||
|
(MADDW a (MOVDconst [c]) x) && isPowerOfTwo(c) -> (ADDshiftLL a x [log2(c)])
|
||||||
|
(MADDW a (MOVDconst [c]) x) && isPowerOfTwo(c-1) && int32(c)>=3 -> (ADD a (ADDshiftLL <x.Type> x x [log2(c-1)]))
|
||||||
|
(MADDW a (MOVDconst [c]) x) && isPowerOfTwo(c+1) && int32(c)>=7 -> (SUB a (SUBshiftLL <x.Type> x x [log2(c+1)]))
|
||||||
|
(MADDW a (MOVDconst [c]) x) && c%3 == 0 && isPowerOfTwo(c/3) && is32Bit(c) -> (SUBshiftLL a (SUBshiftLL <x.Type> x x [2]) [log2(c/3)])
|
||||||
|
(MADDW a (MOVDconst [c]) x) && c%5 == 0 && isPowerOfTwo(c/5) && is32Bit(c) -> (ADDshiftLL a (ADDshiftLL <x.Type> x x [2]) [log2(c/5)])
|
||||||
|
(MADDW a (MOVDconst [c]) x) && c%7 == 0 && isPowerOfTwo(c/7) && is32Bit(c) -> (SUBshiftLL a (SUBshiftLL <x.Type> x x [3]) [log2(c/7)])
|
||||||
|
(MADDW a (MOVDconst [c]) x) && c%9 == 0 && isPowerOfTwo(c/9) && is32Bit(c) -> (ADDshiftLL a (ADDshiftLL <x.Type> x x [3]) [log2(c/9)])
|
||||||
|
|
||||||
|
(MSUB a x (MOVDconst [-1])) -> (ADD a x)
|
||||||
|
(MSUB a _ (MOVDconst [0])) -> a
|
||||||
|
(MSUB a x (MOVDconst [1])) -> (SUB a x)
|
||||||
|
(MSUB a x (MOVDconst [c])) && isPowerOfTwo(c) -> (SUBshiftLL a x [log2(c)])
|
||||||
|
(MSUB a x (MOVDconst [c])) && isPowerOfTwo(c-1) && c>=3 -> (SUB a (ADDshiftLL <x.Type> x x [log2(c-1)]))
|
||||||
|
(MSUB a x (MOVDconst [c])) && isPowerOfTwo(c+1) && c>=7 -> (ADD a (SUBshiftLL <x.Type> x x [log2(c+1)]))
|
||||||
|
(MSUB a x (MOVDconst [c])) && c%3 == 0 && isPowerOfTwo(c/3) -> (ADDshiftLL a (SUBshiftLL <x.Type> x x [2]) [log2(c/3)])
|
||||||
|
(MSUB a x (MOVDconst [c])) && c%5 == 0 && isPowerOfTwo(c/5) -> (SUBshiftLL a (ADDshiftLL <x.Type> x x [2]) [log2(c/5)])
|
||||||
|
(MSUB a x (MOVDconst [c])) && c%7 == 0 && isPowerOfTwo(c/7) -> (ADDshiftLL a (SUBshiftLL <x.Type> x x [3]) [log2(c/7)])
|
||||||
|
(MSUB a x (MOVDconst [c])) && c%9 == 0 && isPowerOfTwo(c/9) -> (SUBshiftLL a (ADDshiftLL <x.Type> x x [3]) [log2(c/9)])
|
||||||
|
|
||||||
|
(MSUB a (MOVDconst [-1]) x) -> (ADD a x)
|
||||||
|
(MSUB a (MOVDconst [0]) _) -> a
|
||||||
|
(MSUB a (MOVDconst [1]) x) -> (SUB a x)
|
||||||
|
(MSUB a (MOVDconst [c]) x) && isPowerOfTwo(c) -> (SUBshiftLL a x [log2(c)])
|
||||||
|
(MSUB a (MOVDconst [c]) x) && isPowerOfTwo(c-1) && c>=3 -> (SUB a (ADDshiftLL <x.Type> x x [log2(c-1)]))
|
||||||
|
(MSUB a (MOVDconst [c]) x) && isPowerOfTwo(c+1) && c>=7 -> (ADD a (SUBshiftLL <x.Type> x x [log2(c+1)]))
|
||||||
|
(MSUB a (MOVDconst [c]) x) && c%3 == 0 && isPowerOfTwo(c/3) -> (ADDshiftLL a (SUBshiftLL <x.Type> x x [2]) [log2(c/3)])
|
||||||
|
(MSUB a (MOVDconst [c]) x) && c%5 == 0 && isPowerOfTwo(c/5) -> (SUBshiftLL a (ADDshiftLL <x.Type> x x [2]) [log2(c/5)])
|
||||||
|
(MSUB a (MOVDconst [c]) x) && c%7 == 0 && isPowerOfTwo(c/7) -> (ADDshiftLL a (SUBshiftLL <x.Type> x x [3]) [log2(c/7)])
|
||||||
|
(MSUB a (MOVDconst [c]) x) && c%9 == 0 && isPowerOfTwo(c/9) -> (SUBshiftLL a (ADDshiftLL <x.Type> x x [3]) [log2(c/9)])
|
||||||
|
|
||||||
|
(MSUBW a x (MOVDconst [c])) && int32(c)==-1 -> (ADD a x)
|
||||||
|
(MSUBW a _ (MOVDconst [c])) && int32(c)==0 -> a
|
||||||
|
(MSUBW a x (MOVDconst [c])) && int32(c)==1 -> (SUB a x)
|
||||||
|
(MSUBW a x (MOVDconst [c])) && isPowerOfTwo(c) -> (SUBshiftLL a x [log2(c)])
|
||||||
|
(MSUBW a x (MOVDconst [c])) && isPowerOfTwo(c-1) && int32(c)>=3 -> (SUB a (ADDshiftLL <x.Type> x x [log2(c-1)]))
|
||||||
|
(MSUBW a x (MOVDconst [c])) && isPowerOfTwo(c+1) && int32(c)>=7 -> (ADD a (SUBshiftLL <x.Type> x x [log2(c+1)]))
|
||||||
|
(MSUBW a x (MOVDconst [c])) && c%3 == 0 && isPowerOfTwo(c/3) && is32Bit(c) -> (ADDshiftLL a (SUBshiftLL <x.Type> x x [2]) [log2(c/3)])
|
||||||
|
(MSUBW a x (MOVDconst [c])) && c%5 == 0 && isPowerOfTwo(c/5) && is32Bit(c) -> (SUBshiftLL a (ADDshiftLL <x.Type> x x [2]) [log2(c/5)])
|
||||||
|
(MSUBW a x (MOVDconst [c])) && c%7 == 0 && isPowerOfTwo(c/7) && is32Bit(c) -> (ADDshiftLL a (SUBshiftLL <x.Type> x x [3]) [log2(c/7)])
|
||||||
|
(MSUBW a x (MOVDconst [c])) && c%9 == 0 && isPowerOfTwo(c/9) && is32Bit(c) -> (SUBshiftLL a (ADDshiftLL <x.Type> x x [3]) [log2(c/9)])
|
||||||
|
|
||||||
|
(MSUBW a (MOVDconst [c]) x) && int32(c)==-1 -> (ADD a x)
|
||||||
|
(MSUBW a (MOVDconst [c]) _) && int32(c)==0 -> a
|
||||||
|
(MSUBW a (MOVDconst [c]) x) && int32(c)==1 -> (SUB a x)
|
||||||
|
(MSUBW a (MOVDconst [c]) x) && isPowerOfTwo(c) -> (SUBshiftLL a x [log2(c)])
|
||||||
|
(MSUBW a (MOVDconst [c]) x) && isPowerOfTwo(c-1) && int32(c)>=3 -> (SUB a (ADDshiftLL <x.Type> x x [log2(c-1)]))
|
||||||
|
(MSUBW a (MOVDconst [c]) x) && isPowerOfTwo(c+1) && int32(c)>=7 -> (ADD a (SUBshiftLL <x.Type> x x [log2(c+1)]))
|
||||||
|
(MSUBW a (MOVDconst [c]) x) && c%3 == 0 && isPowerOfTwo(c/3) && is32Bit(c) -> (ADDshiftLL a (SUBshiftLL <x.Type> x x [2]) [log2(c/3)])
|
||||||
|
(MSUBW a (MOVDconst [c]) x) && c%5 == 0 && isPowerOfTwo(c/5) && is32Bit(c) -> (SUBshiftLL a (ADDshiftLL <x.Type> x x [2]) [log2(c/5)])
|
||||||
|
(MSUBW a (MOVDconst [c]) x) && c%7 == 0 && isPowerOfTwo(c/7) && is32Bit(c) -> (ADDshiftLL a (SUBshiftLL <x.Type> x x [3]) [log2(c/7)])
|
||||||
|
(MSUBW a (MOVDconst [c]) x) && c%9 == 0 && isPowerOfTwo(c/9) && is32Bit(c) -> (SUBshiftLL a (ADDshiftLL <x.Type> x x [3]) [log2(c/9)])
|
||||||
|
|
||||||
// div by constant
|
// div by constant
|
||||||
(UDIV x (MOVDconst [1])) -> x
|
(UDIV x (MOVDconst [1])) -> x
|
||||||
(UDIV x (MOVDconst [c])) && isPowerOfTwo(c) -> (SRLconst [log2(c)] x)
|
(UDIV x (MOVDconst [c])) && isPowerOfTwo(c) -> (SRLconst [log2(c)] x)
|
||||||
|
|
@ -1261,6 +1349,14 @@
|
||||||
(MULW (MOVDconst [c]) (MOVDconst [d])) -> (MOVDconst [int64(int32(c)*int32(d))])
|
(MULW (MOVDconst [c]) (MOVDconst [d])) -> (MOVDconst [int64(int32(c)*int32(d))])
|
||||||
(MNEG (MOVDconst [c]) (MOVDconst [d])) -> (MOVDconst [-c*d])
|
(MNEG (MOVDconst [c]) (MOVDconst [d])) -> (MOVDconst [-c*d])
|
||||||
(MNEGW (MOVDconst [c]) (MOVDconst [d])) -> (MOVDconst [-int64(int32(c)*int32(d))])
|
(MNEGW (MOVDconst [c]) (MOVDconst [d])) -> (MOVDconst [-int64(int32(c)*int32(d))])
|
||||||
|
(MADD (MOVDconst [c]) x y) -> (ADDconst [c] (MUL <x.Type> x y))
|
||||||
|
(MADDW (MOVDconst [c]) x y) -> (ADDconst [c] (MULW <x.Type> x y))
|
||||||
|
(MSUB (MOVDconst [c]) x y) -> (ADDconst [c] (MNEG <x.Type> x y))
|
||||||
|
(MSUBW (MOVDconst [c]) x y) -> (ADDconst [c] (MNEGW <x.Type> x y))
|
||||||
|
(MADD a (MOVDconst [c]) (MOVDconst [d])) -> (ADDconst [c*d] a)
|
||||||
|
(MADDW a (MOVDconst [c]) (MOVDconst [d])) -> (ADDconst [int64(int32(c)*int32(d))] a)
|
||||||
|
(MSUB a (MOVDconst [c]) (MOVDconst [d])) -> (SUBconst [c*d] a)
|
||||||
|
(MSUBW a (MOVDconst [c]) (MOVDconst [d])) -> (SUBconst [int64(int32(c)*int32(d))] a)
|
||||||
(DIV (MOVDconst [c]) (MOVDconst [d])) -> (MOVDconst [c/d])
|
(DIV (MOVDconst [c]) (MOVDconst [d])) -> (MOVDconst [c/d])
|
||||||
(UDIV (MOVDconst [c]) (MOVDconst [d])) -> (MOVDconst [int64(uint64(c)/uint64(d))])
|
(UDIV (MOVDconst [c]) (MOVDconst [d])) -> (MOVDconst [int64(uint64(c)/uint64(d))])
|
||||||
(DIVW (MOVDconst [c]) (MOVDconst [d])) -> (MOVDconst [int64(int32(c)/int32(d))])
|
(DIVW (MOVDconst [c]) (MOVDconst [d])) -> (MOVDconst [int64(int32(c)/int32(d))])
|
||||||
|
|
|
||||||
File diff suppressed because it is too large
Load Diff
|
|
@ -206,8 +206,15 @@ func AddMul(x int) int {
|
||||||
return 2*x + 1
|
return 2*x + 1
|
||||||
}
|
}
|
||||||
|
|
||||||
func MULA(a, b, c uint32) uint32 {
|
func MULA(a, b, c uint32) (uint32, uint32, uint32) {
|
||||||
// arm:`MULA`
|
// arm:`MULA`,-`MUL\s`
|
||||||
// arm64:`MADDW`
|
// arm64:`MADDW`,-`MULW`
|
||||||
return a*b + c
|
r0 := a*b + c
|
||||||
|
// arm:`MULA`-`MUL\s`
|
||||||
|
// arm64:`MADDW`,-`MULW`
|
||||||
|
r1 := c*79 + a
|
||||||
|
// arm:`ADD`,-`MULA`-`MUL\s`
|
||||||
|
// arm64:`ADD`,-`MADD`,-`MULW`
|
||||||
|
r2 := b*64 + c
|
||||||
|
return r0, r1, r2
|
||||||
}
|
}
|
||||||
|
|
|
||||||
Loading…
Reference in New Issue