[dev.simd] cmd/compile: zero only low 128-bit of X15

Zeroing the upper part of X15 may make the CPU think it is
"dirty" and slow down SSE operations. For now, just not zeroing
the upper part, and construct a zero value on the fly if we need
a 256- or 512-bit zero value. Maybe VZEROUPPER works better than
explicitly zeroing X15, but we need to evaluate.

Long term, we probably want to move more things from SSE to AVX.

This essentially undoes CL 698237 and CL 698238, except keeping
using X15 for 128-bit zeroing for SIMD.

Change-Id: I1564e6332c4c57f9721397c92c7c734c5497534c
Reviewed-on: https://go-review.googlesource.com/c/go/+/728240
LUCI-TryBot-Result: Go LUCI <golang-scoped@luci-project-accounts.iam.gserviceaccount.com>
Reviewed-by: David Chase <drchase@google.com>

9 files changed, 0 insertions, 60 deletions

diff --git a/src/runtime/asm_amd64.s b/src/runtime/asm_amd64.s
index bf208a4d29..391d9bcd22 100644
--- a/src/runtime/asm_amd64.s
+++ b/src/runtime/asm_amd64.s
@@ -1093,11 +1093,6 @@ needm:
 	// there's no need to handle that. Clear R14 so that there's
 	// a bad value in there, in case needm tries to use it.
 	XORPS	X15, X15
-#ifdef GOEXPERIMENT_simd
-	CMPB	internal∕cpu·X86+const_offsetX86HasAVX(SB), $1
-	JNE	2(PC)
-	VXORPS	X15, X15, X15
-#endif
 	XORQ    R14, R14
 	MOVQ	$runtime·needAndBindM<ABIInternal>(SB), AX
 	CALL	AX
@@ -1795,11 +1790,6 @@ TEXT ·sigpanic0(SB),NOSPLIT,$0-0
 	get_tls(R14)
 	MOVQ	g(R14), R14
 	XORPS	X15, X15
-#ifdef GOEXPERIMENT_simd
-	CMPB	internal∕cpu·X86+const_offsetX86HasAVX(SB), $1
-	JNE	2(PC)
-	VXORPS	X15, X15, X15
-#endif
 	JMP	·sigpanic<ABIInternal>(SB)
 
 // gcWriteBarrier informs the GC about heap pointer writes.
diff --git a/src/runtime/race_amd64.s b/src/runtime/race_amd64.s
index ade29bc5f1..e19118bd54 100644
--- a/src/runtime/race_amd64.s
+++ b/src/runtime/race_amd64.s
@@ -456,11 +456,6 @@ call:
 	// Back to Go world, set special registers.
 	// The g register (R14) is preserved in C.
 	XORPS	X15, X15
-#ifdef GOEXPERIMENT_simd
-	CMPB	internal∕cpu·X86+const_offsetX86HasAVX(SB), $1
-	JNE	2(PC)
-	VXORPS	X15, X15, X15
-#endif
 	RET
 
 // C->Go callback thunk that allows to call runtime·racesymbolize from C code.
diff --git a/src/runtime/sys_darwin_amd64.s b/src/runtime/sys_darwin_amd64.s
index e033e8b702..99d67a9cfd 100644
--- a/src/runtime/sys_darwin_amd64.s
+++ b/src/runtime/sys_darwin_amd64.s
@@ -177,11 +177,6 @@ TEXT runtime·sigtramp(SB),NOSPLIT|TOPFRAME|NOFRAME,$0
 	get_tls(R12)
 	MOVQ	g(R12), R14
 	PXOR	X15, X15
-#ifdef GOEXPERIMENT_simd
-	CMPB	internal∕cpu·X86+const_offsetX86HasAVX(SB), $1
-	JNE	2(PC)
-	VXORPS	X15, X15, X15
-#endif
 
 	// Reserve space for spill slots.
 	NOP	SP		// disable vet stack checking
diff --git a/src/runtime/sys_dragonfly_amd64.s b/src/runtime/sys_dragonfly_amd64.s
index e417d4b8a8..a223c2cf76 100644
--- a/src/runtime/sys_dragonfly_amd64.s
+++ b/src/runtime/sys_dragonfly_amd64.s
@@ -228,11 +228,6 @@ TEXT runtime·sigtramp(SB),NOSPLIT|TOPFRAME|NOFRAME,$0
 	get_tls(R12)
 	MOVQ	g(R12), R14
 	PXOR	X15, X15
-#ifdef GOEXPERIMENT_simd
-	CMPB	internal∕cpu·X86+const_offsetX86HasAVX(SB), $1
-	JNE	2(PC)
-	VXORPS	X15, X15, X15
-#endif
 
 	// Reserve space for spill slots.
 	NOP	SP		// disable vet stack checking
diff --git a/src/runtime/sys_freebsd_amd64.s b/src/runtime/sys_freebsd_amd64.s
index bab275cc72..977ea093d2 100644
--- a/src/runtime/sys_freebsd_amd64.s
+++ b/src/runtime/sys_freebsd_amd64.s
@@ -265,11 +265,6 @@ TEXT runtime·sigtramp(SB),NOSPLIT|TOPFRAME|NOFRAME,$0
 	get_tls(R12)
 	MOVQ	g(R12), R14
 	PXOR	X15, X15
-#ifdef GOEXPERIMENT_simd
-	CMPB	internal∕cpu·X86+const_offsetX86HasAVX(SB), $1
-	JNE	2(PC)
-	VXORPS	X15, X15, X15
-#endif
 
 	// Reserve space for spill slots.
 	NOP	SP		// disable vet stack checking
@@ -295,11 +290,6 @@ TEXT runtime·sigprofNonGoWrapper<>(SB),NOSPLIT|NOFRAME,$0
 	get_tls(R12)
 	MOVQ	g(R12), R14
 	PXOR	X15, X15
-#ifdef GOEXPERIMENT_simd
-	CMPB	internal∕cpu·X86+const_offsetX86HasAVX(SB), $1
-	JNE	2(PC)
-	VXORPS	X15, X15, X15
-#endif
 
 	// Reserve space for spill slots.
 	NOP	SP		// disable vet stack checking
diff --git a/src/runtime/sys_linux_amd64.s b/src/runtime/sys_linux_amd64.s
index 618553b196..878f834748 100644
--- a/src/runtime/sys_linux_amd64.s
+++ b/src/runtime/sys_linux_amd64.s
@@ -352,11 +352,6 @@ TEXT runtime·sigtramp(SB),NOSPLIT|TOPFRAME|NOFRAME,$0
 	get_tls(R12)
 	MOVQ	g(R12), R14
 	PXOR	X15, X15
-#ifdef GOEXPERIMENT_simd
-	CMPB	internal∕cpu·X86+const_offsetX86HasAVX(SB), $1
-	JNE	2(PC)
-	VXORPS	X15, X15, X15
-#endif
 
 	// Reserve space for spill slots.
 	NOP	SP		// disable vet stack checking
@@ -382,11 +377,6 @@ TEXT runtime·sigprofNonGoWrapper<>(SB),NOSPLIT|NOFRAME,$0
 	get_tls(R12)
 	MOVQ	g(R12), R14
 	PXOR	X15, X15
-#ifdef GOEXPERIMENT_simd
-	CMPB	internal∕cpu·X86+const_offsetX86HasAVX(SB), $1
-	JNE	2(PC)
-	VXORPS	X15, X15, X15
-#endif
 
 	// Reserve space for spill slots.
 	NOP	SP		// disable vet stack checking
diff --git a/src/runtime/sys_netbsd_amd64.s b/src/runtime/sys_netbsd_amd64.s
index 946b1fbe22..2f1ddcdc89 100644
--- a/src/runtime/sys_netbsd_amd64.s
+++ b/src/runtime/sys_netbsd_amd64.s
@@ -310,11 +310,6 @@ TEXT runtime·sigtramp(SB),NOSPLIT|TOPFRAME|NOFRAME,$0
 	get_tls(R12)
 	MOVQ	g(R12), R14
 	PXOR	X15, X15
-#ifdef GOEXPERIMENT_simd
-	CMPB	internal∕cpu·X86+const_offsetX86HasAVX(SB), $1
-	JNE	2(PC)
-	VXORPS	X15, X15, X15
-#endif
 
 	// Reserve space for spill slots.
 	NOP	SP		// disable vet stack checking
diff --git a/src/runtime/sys_openbsd_amd64.s b/src/runtime/sys_openbsd_amd64.s
index 7766fa5194..ff0bc2416a 100644
--- a/src/runtime/sys_openbsd_amd64.s
+++ b/src/runtime/sys_openbsd_amd64.s
@@ -64,11 +64,6 @@ TEXT runtime·sigtramp(SB),NOSPLIT|TOPFRAME|NOFRAME,$0
 	get_tls(R12)
 	MOVQ	g(R12), R14
 	PXOR	X15, X15
-#ifdef GOEXPERIMENT_simd
-	CMPB	internal∕cpu·X86+const_offsetX86HasAVX(SB), $1
-	JNE	2(PC)
-	VXORPS	X15, X15, X15
-#endif
 
 	// Reserve space for spill slots.
 	NOP	SP		// disable vet stack checking
diff --git a/src/runtime/sys_windows_amd64.s b/src/runtime/sys_windows_amd64.s
index 52a21ba89b..e438599910 100644
--- a/src/runtime/sys_windows_amd64.s
+++ b/src/runtime/sys_windows_amd64.s
@@ -32,11 +32,6 @@ TEXT sigtramp<>(SB),NOSPLIT,$0-0
 	// R14 is cleared in case there's a non-zero value in there
 	// if called from a non-go thread.
 	XORPS	X15, X15
-#ifdef GOEXPERIMENT_simd
-	CMPB	internal∕cpu·X86+const_offsetX86HasAVX(SB), $1
-	JNE	2(PC)
-	VXORPS	X15, X15, X15
-#endif
 	XORQ	R14, R14
 
 	get_tls(AX)