]>
Cypherpunks repositories - gostls13.git/commit
cmd/compile: use arm64 neon in LoweredMemmove/LoweredMemmoveLoop
Raspberry Pi 5 (Cortex-A76)
│ base.log │ opt.log │
│ sec/op │ sec/op vs base │
MemmoveKnownSize112 3.549n ± 0% 3.652n ± 0% +2.92% (p=0.000 n=10)
MemmoveKnownSize128 3.979n ± 0% 3.617n ± 0% -9.09% (p=0.000 n=10)
MemmoveKnownSize192 7.566n ± 0% 5.074n ± 0% -32.94% (p=0.000 n=10)
MemmoveKnownSize248 8.549n ± 0% 7.184n ± 1% -15.97% (p=0.000 n=10)
MemmoveKnownSize256 10.010n ± 0% 6.827n ± 0% -31.80% (p=0.000 n=10)
MemmoveKnownSize512 19.81n ± 0% 13.59n ± 0% -31.40% (p=0.000 n=10)
MemmoveKnownSize1024 39.66n ± 0% 27.00n ± 0% -31.93% (p=0.000 n=10)
geomean 9.538n 7.392n -22.50%
Change-Id: I7b17408cd0a500ceaa80bc93ffe2f19ddeea9c0d
Reviewed-on: https://go-review.googlesource.com/c/go/+/692315
Reviewed-by: Keith Randall <khr@google.com>
Reviewed-by: David Chase <drchase@google.com>
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>