From c515852732a490bab64f35d001ddc444b0f0f553 Mon Sep 17 00:00:00 2001 From: Heisenberg Date: Tue, 8 Sep 2020 14:31:39 +0800 Subject: [PATCH] runtime: add 2-byte and 8-byte sub-benchmarks for memmove load/store Change-Id: I6389d7efe90836b6ece44d2e75053d1ad9f35d08 Reviewed-on: https://go-review.googlesource.com/c/go/+/253417 Trust: Emmanuel Odeke Reviewed-by: Keith Randall --- src/runtime/memmove_test.go | 39 +++++++++++++++++++++++-------------- 1 file changed, 24 insertions(+), 15 deletions(-) diff --git a/src/runtime/memmove_test.go b/src/runtime/memmove_test.go index 396c1304c5..b549433f71 100644 --- a/src/runtime/memmove_test.go +++ b/src/runtime/memmove_test.go @@ -538,21 +538,30 @@ func BenchmarkCopyFat1024(b *testing.B) { } } +// BenchmarkIssue18740 ensures that memmove uses 4 and 8 byte load/store to move 4 and 8 bytes. +// It used to do 2 2-byte load/stores, which leads to a pipeline stall +// when we try to read the result with one 4-byte load. func BenchmarkIssue18740(b *testing.B) { - // This tests that memmove uses one 4-byte load/store to move 4 bytes. - // It used to do 2 2-byte load/stores, which leads to a pipeline stall - // when we try to read the result with one 4-byte load. - var buf [4]byte - for j := 0; j < b.N; j++ { - s := uint32(0) - for i := 0; i < 4096; i += 4 { - copy(buf[:], g[i:]) - s += binary.LittleEndian.Uint32(buf[:]) - } - sink = uint64(s) + benchmarks := []struct { + name string + nbyte int + f func([]byte) uint64 + }{ + {"2byte", 2, func(buf []byte) uint64 { return uint64(binary.LittleEndian.Uint16(buf)) }}, + {"4byte", 4, func(buf []byte) uint64 { return uint64(binary.LittleEndian.Uint32(buf)) }}, + {"8byte", 8, func(buf []byte) uint64 { return binary.LittleEndian.Uint64(buf) }}, + } + + var g [4096]byte + for _, bm := range benchmarks { + buf := make([]byte, bm.nbyte) + b.Run(bm.name, func(b *testing.B) { + for j := 0; j < b.N; j++ { + for i := 0; i < 4096; i += bm.nbyte { + copy(buf[:], g[i:]) + sink += bm.f(buf[:]) + } + } + }) } } - -// TODO: 2 byte and 8 byte benchmarks also. - -var g [4096]byte -- 2.50.0