diff options
| author | fanzha02 <fannie.zhang@arm.com> | 2020-12-07 19:15:15 +0800 |
|---|---|---|
| committer | fannie zhang <Fannie.Zhang@arm.com> | 2021-03-02 06:38:07 +0000 |
| commit | 2b50ab2aee75d3c361fcd1eb39e830e2e73056b6 (patch) | |
| tree | 86c8adabbab0b962827eebb995e054fc5ac6098d /test/codegen/math.go | |
| parent | ebb92dfed96fadb3c563ff11cead85bbb7536793 (diff) | |
| download | go-2b50ab2aee75d3c361fcd1eb39e830e2e73056b6.tar.xz | |
cmd/compile: optimize single-precision floating point square root
Add generic rule to rewrite the single-precision square root expression
with one single-precision instruction. The optimization will reduce two
times of precision converting between double-precision and single-precision.
On arm64 flatform.
previous:
FCVTSD F0, F0
FSQRTD F0, F0
FCVTDS F0, F0
optimized:
FSQRTS S0, S0
And this patch adds the test case to check the correctness.
This patch refers to CL 241877, contributed by Alice Xu
(dianhong.xu@arm.com)
Change-Id: I6de5d02281c693017ac4bd4c10963dd55989bd7e
Reviewed-on: https://go-review.googlesource.com/c/go/+/276873
Trust: fannie zhang <Fannie.Zhang@arm.com>
Run-TryBot: fannie zhang <Fannie.Zhang@arm.com>
TryBot-Result: Go Bot <gobot@golang.org>
Reviewed-by: Keith Randall <khr@golang.org>
Diffstat (limited to 'test/codegen/math.go')
| -rw-r--r-- | test/codegen/math.go | 11 |
1 files changed, 11 insertions, 0 deletions
diff --git a/test/codegen/math.go b/test/codegen/math.go index ac8071400e..243ddb0494 100644 --- a/test/codegen/math.go +++ b/test/codegen/math.go @@ -55,6 +55,17 @@ func sqrt(x float64) float64 { return math.Sqrt(x) } +func sqrt32(x float32) float32 { + // amd64:"SQRTSS" + // 386/sse2:"SQRTSS" 386/softfloat:-"SQRTS" + // arm64:"FSQRTS" + // arm/7:"SQRTF" + // mips/hardfloat:"SQRTF" mips/softfloat:-"SQRTF" + // mips64/hardfloat:"SQRTF" mips64/softfloat:-"SQRTF" + // wasm:"F32Sqrt" + return float32(math.Sqrt(float64(x))) +} + // Check that it's using integer registers func abs(x, y float64) { // amd64:"BTRQ\t[$]63" |