[llvm] r249990 - [X86] Completed SHL cost model tests
Simon Pilgrim via llvm-commits
llvm-commits at lists.llvm.org
Sun Oct 11 11:33:48 PDT 2015
Author: rksimon
Date: Sun Oct 11 13:33:48 2015
New Revision: 249990
URL: http://llvm.org/viewvc/llvm-project?rev=249990&view=rev
Log:
[X86] Completed SHL cost model tests
As discussed in D8690.
Modified:
llvm/trunk/test/Analysis/CostModel/X86/vshift-shl-cost.ll
Modified: llvm/trunk/test/Analysis/CostModel/X86/vshift-shl-cost.ll
URL: http://llvm.org/viewvc/llvm-project/llvm/trunk/test/Analysis/CostModel/X86/vshift-shl-cost.ll?rev=249990&r1=249989&r2=249990&view=diff
==============================================================================
--- llvm/trunk/test/Analysis/CostModel/X86/vshift-shl-cost.ll (original)
+++ llvm/trunk/test/Analysis/CostModel/X86/vshift-shl-cost.ll Sun Oct 11 13:33:48 2015
@@ -5,9 +5,407 @@
; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -mcpu=bdver2 -cost-model -analyze | FileCheck %s -check-prefix=CHECK -check-prefix=XOP -check-prefix=XOPAVX
; RUN: opt < %s -mtriple=x86_64-unknown-linux-gnu -mcpu=bdver4 -cost-model -analyze | FileCheck %s -check-prefix=CHECK -check-prefix=XOP -check-prefix=XOPAVX2
-
; Verify the cost of vector shift left instructions.
+;
+;
+; Variable Shifts
+;
+
+define <2 x i64> @var_shift_v2i64(<2 x i64> %a, <2 x i64> %b) {
+; CHECK: 'Cost Model Analysis' for function 'var_shift_v2i64':
+; SSE2: Found an estimated cost of 4 for instruction: %shift
+; SSE41: Found an estimated cost of 4 for instruction: %shift
+; AVX: Found an estimated cost of 4 for instruction: %shift
+; AVX2: Found an estimated cost of 1 for instruction: %shift
+; XOPAVX: Found an estimated cost of 1 for instruction: %shift
+; XOPAVX2: Found an estimated cost of 1 for instruction: %shift
+ %shift = shl <2 x i64> %a, %b
+ ret <2 x i64> %shift
+}
+
+define <4 x i64> @var_shift_v4i64(<4 x i64> %a, <4 x i64> %b) {
+; CHECK: 'Cost Model Analysis' for function 'var_shift_v4i64':
+; SSE2: Found an estimated cost of 8 for instruction: %shift
+; SSE41: Found an estimated cost of 8 for instruction: %shift
+; AVX: Found an estimated cost of 8 for instruction: %shift
+; AVX2: Found an estimated cost of 1 for instruction: %shift
+; XOPAVX: Found an estimated cost of 2 for instruction: %shift
+; XOPAVX2: Found an estimated cost of 1 for instruction: %shift
+ %shift = shl <4 x i64> %a, %b
+ ret <4 x i64> %shift
+}
+
+define <4 x i32> @var_shift_v4i32(<4 x i32> %a, <4 x i32> %b) {
+; CHECK: 'Cost Model Analysis' for function 'var_shift_v4i32':
+; SSE2: Found an estimated cost of 10 for instruction: %shift
+; SSE41: Found an estimated cost of 10 for instruction: %shift
+; AVX: Found an estimated cost of 10 for instruction: %shift
+; AVX2: Found an estimated cost of 1 for instruction: %shift
+; XOPAVX: Found an estimated cost of 1 for instruction: %shift
+; XOPAVX2: Found an estimated cost of 1 for instruction: %shift
+ %shift = shl <4 x i32> %a, %b
+ ret <4 x i32> %shift
+}
+
+define <8 x i32> @var_shift_v8i32(<8 x i32> %a, <8 x i32> %b) {
+; CHECK: 'Cost Model Analysis' for function 'var_shift_v8i32':
+; SSE2: Found an estimated cost of 20 for instruction: %shift
+; SSE41: Found an estimated cost of 20 for instruction: %shift
+; AVX: Found an estimated cost of 2 for instruction: %shift
+; AVX2: Found an estimated cost of 1 for instruction: %shift
+; XOPAVX: Found an estimated cost of 2 for instruction: %shift
+; XOPAVX2: Found an estimated cost of 1 for instruction: %shift
+ %shift = shl <8 x i32> %a, %b
+ ret <8 x i32> %shift
+}
+
+define <8 x i16> @var_shift_v8i16(<8 x i16> %a, <8 x i16> %b) {
+; CHECK: 'Cost Model Analysis' for function 'var_shift_v8i16':
+; SSE2: Found an estimated cost of 32 for instruction: %shift
+; SSE41: Found an estimated cost of 32 for instruction: %shift
+; AVX: Found an estimated cost of 32 for instruction: %shift
+; AVX2: Found an estimated cost of 32 for instruction: %shift
+; XOP: Found an estimated cost of 1 for instruction: %shift
+ %shift = shl <8 x i16> %a, %b
+ ret <8 x i16> %shift
+}
+
+define <16 x i16> @var_shift_v16i16(<16 x i16> %a, <16 x i16> %b) {
+; CHECK: 'Cost Model Analysis' for function 'var_shift_v16i16':
+; SSE2: Found an estimated cost of 64 for instruction: %shift
+; SSE41: Found an estimated cost of 64 for instruction: %shift
+; AVX: Found an estimated cost of 2 for instruction: %shift
+; AVX2: Found an estimated cost of 10 for instruction: %shift
+; XOP: Found an estimated cost of 2 for instruction: %shift
+ %shift = shl <16 x i16> %a, %b
+ ret <16 x i16> %shift
+}
+
+define <16 x i8> @var_shift_v16i8(<16 x i8> %a, <16 x i8> %b) {
+; CHECK: 'Cost Model Analysis' for function 'var_shift_v16i8':
+; SSE2: Found an estimated cost of 26 for instruction: %shift
+; SSE41: Found an estimated cost of 26 for instruction: %shift
+; AVX: Found an estimated cost of 26 for instruction: %shift
+; AVX2: Found an estimated cost of 26 for instruction: %shift
+; XOP: Found an estimated cost of 1 for instruction: %shift
+ %shift = shl <16 x i8> %a, %b
+ ret <16 x i8> %shift
+}
+
+define <32 x i8> @var_shift_v32i8(<32 x i8> %a, <32 x i8> %b) {
+; CHECK: 'Cost Model Analysis' for function 'var_shift_v32i8':
+; SSE2: Found an estimated cost of 52 for instruction: %shift
+; SSE41: Found an estimated cost of 52 for instruction: %shift
+; AVX: Found an estimated cost of 2 for instruction: %shift
+; AVX2: Found an estimated cost of 11 for instruction: %shift
+; XOP: Found an estimated cost of 2 for instruction: %shift
+ %shift = shl <32 x i8> %a, %b
+ ret <32 x i8> %shift
+}
+
+;
+; Uniform Variable Shifts
+;
+
+define <2 x i64> @splatvar_shift_v2i64(<2 x i64> %a, <2 x i64> %b) {
+; CHECK: 'Cost Model Analysis' for function 'splatvar_shift_v2i64':
+; SSE2: Found an estimated cost of 4 for instruction: %shift
+; SSE41: Found an estimated cost of 4 for instruction: %shift
+; AVX: Found an estimated cost of 4 for instruction: %shift
+; AVX2: Found an estimated cost of 1 for instruction: %shift
+; XOPAVX: Found an estimated cost of 1 for instruction: %shift
+; XOPAVX2: Found an estimated cost of 1 for instruction: %shift
+ %splat = shufflevector <2 x i64> %b, <2 x i64> undef, <2 x i32> zeroinitializer
+ %shift = shl <2 x i64> %a, %splat
+ ret <2 x i64> %shift
+}
+
+define <4 x i64> @splatvar_shift_v4i64(<4 x i64> %a, <4 x i64> %b) {
+; CHECK: 'Cost Model Analysis' for function 'splatvar_shift_v4i64':
+; SSE2: Found an estimated cost of 8 for instruction: %shift
+; SSE41: Found an estimated cost of 8 for instruction: %shift
+; AVX: Found an estimated cost of 8 for instruction: %shift
+; AVX2: Found an estimated cost of 1 for instruction: %shift
+; XOPAVX: Found an estimated cost of 2 for instruction: %shift
+; XOPAVX2: Found an estimated cost of 1 for instruction: %shift
+ %splat = shufflevector <4 x i64> %b, <4 x i64> undef, <4 x i32> zeroinitializer
+ %shift = shl <4 x i64> %a, %splat
+ ret <4 x i64> %shift
+}
+
+define <4 x i32> @splatvar_shift_v4i32(<4 x i32> %a, <4 x i32> %b) {
+; CHECK: 'Cost Model Analysis' for function 'splatvar_shift_v4i32':
+; SSE2: Found an estimated cost of 10 for instruction: %shift
+; SSE41: Found an estimated cost of 10 for instruction: %shift
+; AVX: Found an estimated cost of 10 for instruction: %shift
+; AVX2: Found an estimated cost of 1 for instruction: %shift
+; XOPAVX: Found an estimated cost of 1 for instruction: %shift
+; XOPAVX2: Found an estimated cost of 1 for instruction: %shift
+ %splat = shufflevector <4 x i32> %b, <4 x i32> undef, <4 x i32> zeroinitializer
+ %shift = shl <4 x i32> %a, %splat
+ ret <4 x i32> %shift
+}
+
+define <8 x i32> @splatvar_shift_v8i32(<8 x i32> %a, <8 x i32> %b) {
+; CHECK: 'Cost Model Analysis' for function 'splatvar_shift_v8i32':
+; SSE2: Found an estimated cost of 20 for instruction: %shift
+; SSE41: Found an estimated cost of 20 for instruction: %shift
+; AVX: Found an estimated cost of 2 for instruction: %shift
+; AVX2: Found an estimated cost of 1 for instruction: %shift
+; XOPAVX: Found an estimated cost of 2 for instruction: %shift
+; XOPAVX2: Found an estimated cost of 1 for instruction: %shift
+ %splat = shufflevector <8 x i32> %b, <8 x i32> undef, <8 x i32> zeroinitializer
+ %shift = shl <8 x i32> %a, %splat
+ ret <8 x i32> %shift
+}
+
+define <8 x i16> @splatvar_shift_v8i16(<8 x i16> %a, <8 x i16> %b) {
+; CHECK: 'Cost Model Analysis' for function 'splatvar_shift_v8i16':
+; SSE2: Found an estimated cost of 32 for instruction: %shift
+; SSE41: Found an estimated cost of 32 for instruction: %shift
+; AVX: Found an estimated cost of 32 for instruction: %shift
+; AVX2: Found an estimated cost of 32 for instruction: %shift
+; XOP: Found an estimated cost of 1 for instruction: %shift
+ %splat = shufflevector <8 x i16> %b, <8 x i16> undef, <8 x i32> zeroinitializer
+ %shift = shl <8 x i16> %a, %splat
+ ret <8 x i16> %shift
+}
+
+define <16 x i16> @splatvar_shift_v16i16(<16 x i16> %a, <16 x i16> %b) {
+; CHECK: 'Cost Model Analysis' for function 'splatvar_shift_v16i16':
+; SSE2: Found an estimated cost of 64 for instruction: %shift
+; SSE41: Found an estimated cost of 64 for instruction: %shift
+; AVX: Found an estimated cost of 2 for instruction: %shift
+; AVX2: Found an estimated cost of 10 for instruction: %shift
+; XOP: Found an estimated cost of 2 for instruction: %shift
+ %splat = shufflevector <16 x i16> %b, <16 x i16> undef, <16 x i32> zeroinitializer
+ %shift = shl <16 x i16> %a, %splat
+ ret <16 x i16> %shift
+}
+
+define <16 x i8> @splatvar_shift_v16i8(<16 x i8> %a, <16 x i8> %b) {
+; CHECK: 'Cost Model Analysis' for function 'splatvar_shift_v16i8':
+; SSE2: Found an estimated cost of 26 for instruction: %shift
+; SSE41: Found an estimated cost of 26 for instruction: %shift
+; AVX: Found an estimated cost of 26 for instruction: %shift
+; AVX2: Found an estimated cost of 26 for instruction: %shift
+; XOP: Found an estimated cost of 1 for instruction: %shift
+ %splat = shufflevector <16 x i8> %b, <16 x i8> undef, <16 x i32> zeroinitializer
+ %shift = shl <16 x i8> %a, %splat
+ ret <16 x i8> %shift
+}
+
+define <32 x i8> @splatvar_shift_v32i8(<32 x i8> %a, <32 x i8> %b) {
+; CHECK: 'Cost Model Analysis' for function 'splatvar_shift_v32i8':
+; SSE2: Found an estimated cost of 52 for instruction: %shift
+; SSE41: Found an estimated cost of 52 for instruction: %shift
+; AVX: Found an estimated cost of 2 for instruction: %shift
+; AVX2: Found an estimated cost of 11 for instruction: %shift
+; XOP: Found an estimated cost of 2 for instruction: %shift
+ %splat = shufflevector <32 x i8> %b, <32 x i8> undef, <32 x i32> zeroinitializer
+ %shift = shl <32 x i8> %a, %splat
+ ret <32 x i8> %shift
+}
+
+;
+; Constant Shifts
+;
+
+define <2 x i64> @constant_shift_v2i64(<2 x i64> %a) {
+; CHECK: 'Cost Model Analysis' for function 'constant_shift_v2i64':
+; SSE2: Found an estimated cost of 4 for instruction: %shift
+; SSE41: Found an estimated cost of 4 for instruction: %shift
+; AVX: Found an estimated cost of 4 for instruction: %shift
+; AVX2: Found an estimated cost of 1 for instruction: %shift
+; XOPAVX: Found an estimated cost of 1 for instruction: %shift
+; XOPAVX2: Found an estimated cost of 1 for instruction: %shift
+ %shift = shl <2 x i64> %a, <i64 1, i64 7>
+ ret <2 x i64> %shift
+}
+
+define <4 x i64> @constant_shift_v4i64(<4 x i64> %a) {
+; CHECK: 'Cost Model Analysis' for function 'constant_shift_v4i64':
+; SSE2: Found an estimated cost of 8 for instruction: %shift
+; SSE41: Found an estimated cost of 8 for instruction: %shift
+; AVX: Found an estimated cost of 8 for instruction: %shift
+; AVX2: Found an estimated cost of 1 for instruction: %shift
+; XOPAVX: Found an estimated cost of 2 for instruction: %shift
+; XOPAVX2: Found an estimated cost of 1 for instruction: %shift
+ %shift = shl <4 x i64> %a, <i64 1, i64 7, i64 15, i64 31>
+ ret <4 x i64> %shift
+}
+
+define <4 x i32> @constant_shift_v4i32(<4 x i32> %a) {
+; CHECK: 'Cost Model Analysis' for function 'constant_shift_v4i32':
+; SSE2: Found an estimated cost of 6 for instruction: %shift
+; SSE41: Found an estimated cost of 1 for instruction: %shift
+; AVX: Found an estimated cost of 1 for instruction: %shift
+; AVX2: Found an estimated cost of 1 for instruction: %shift
+; XOPAVX: Found an estimated cost of 1 for instruction: %shift
+; XOPAVX2: Found an estimated cost of 1 for instruction: %shift
+ %shift = shl <4 x i32> %a, <i32 4, i32 5, i32 6, i32 7>
+ ret <4 x i32> %shift
+}
+
+define <8 x i32> @constant_shift_v8i32(<8 x i32> %a) {
+; CHECK: 'Cost Model Analysis' for function 'constant_shift_v8i32':
+; SSE2: Found an estimated cost of 12 for instruction: %shift
+; SSE41: Found an estimated cost of 2 for instruction: %shift
+; AVX: Found an estimated cost of 4 for instruction: %shift
+; AVX2: Found an estimated cost of 1 for instruction: %shift
+; XOPAVX: Found an estimated cost of 2 for instruction: %shift
+; XOPAVX2: Found an estimated cost of 1 for instruction: %shift
+ %shift = shl <8 x i32> %a, <i32 4, i32 5, i32 6, i32 7, i32 0, i32 1, i32 2, i32 3>
+ ret <8 x i32> %shift
+}
+
+define <8 x i16> @constant_shift_v8i16(<8 x i16> %a) {
+; CHECK: 'Cost Model Analysis' for function 'constant_shift_v8i16':
+; SSE2: Found an estimated cost of 1 for instruction: %shift
+; SSE41: Found an estimated cost of 1 for instruction: %shift
+; AVX: Found an estimated cost of 1 for instruction: %shift
+; AVX2: Found an estimated cost of 1 for instruction: %shift
+; XOP: Found an estimated cost of 1 for instruction: %shift
+ %shift = shl <8 x i16> %a, <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>
+ ret <8 x i16> %shift
+}
+
+define <16 x i16> @constant_shift_v16i16(<16 x i16> %a) {
+; CHECK: 'Cost Model Analysis' for function 'constant_shift_v16i16':
+; SSE2: Found an estimated cost of 2 for instruction: %shift
+; SSE41: Found an estimated cost of 2 for instruction: %shift
+; AVX: Found an estimated cost of 4 for instruction: %shift
+; AVX2: Found an estimated cost of 1 for instruction: %shift
+; XOPAVX: Found an estimated cost of 2 for instruction: %shift
+; XOPAVX2: Found an estimated cost of 1 for instruction: %shift
+ %shift = shl <16 x i16> %a, <i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7, i16 0, i16 1, i16 2, i16 3, i16 4, i16 5, i16 6, i16 7>
+ ret <16 x i16> %shift
+}
+
+define <16 x i8> @constant_shift_v16i8(<16 x i8> %a) {
+; CHECK: 'Cost Model Analysis' for function 'constant_shift_v16i8':
+; SSE2: Found an estimated cost of 26 for instruction: %shift
+; SSE41: Found an estimated cost of 26 for instruction: %shift
+; AVX: Found an estimated cost of 26 for instruction: %shift
+; AVX2: Found an estimated cost of 26 for instruction: %shift
+; XOP: Found an estimated cost of 1 for instruction: %shift
+ %shift = shl <16 x i8> %a, <i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7, i8 7, i8 6, i8 5, i8 4, i8 3, i8 2, i8 1, i8 0>
+ ret <16 x i8> %shift
+}
+
+define <32 x i8> @constant_shift_v32i8(<32 x i8> %a) {
+; CHECK: 'Cost Model Analysis' for function 'constant_shift_v32i8':
+; SSE2: Found an estimated cost of 52 for instruction: %shift
+; SSE41: Found an estimated cost of 52 for instruction: %shift
+; AVX: Found an estimated cost of 2 for instruction: %shift
+; AVX2: Found an estimated cost of 11 for instruction: %shift
+; XOP: Found an estimated cost of 2 for instruction: %shift
+ %shift = shl <32 x i8> %a, <i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7, i8 7, i8 6, i8 5, i8 4, i8 3, i8 2, i8 1, i8 0, i8 0, i8 1, i8 2, i8 3, i8 4, i8 5, i8 6, i8 7, i8 7, i8 6, i8 5, i8 4, i8 3, i8 2, i8 1, i8 0>
+ ret <32 x i8> %shift
+}
+
+;
+; Uniform Constant Shifts
+;
+
+define <2 x i64> @splatconstant_shift_v2i64(<2 x i64> %a) {
+; CHECK: 'Cost Model Analysis' for function 'splatconstant_shift_v2i64':
+; SSE2: Found an estimated cost of 1 for instruction: %shift
+; SSE41: Found an estimated cost of 1 for instruction: %shift
+; AVX: Found an estimated cost of 1 for instruction: %shift
+; AVX2: Found an estimated cost of 1 for instruction: %shift
+; XOPAVX: Found an estimated cost of 1 for instruction: %shift
+; XOPAVX2: Found an estimated cost of 1 for instruction: %shift
+ %shift = shl <2 x i64> %a, <i64 7, i64 7>
+ ret <2 x i64> %shift
+}
+
+define <4 x i64> @splatconstant_shift_v4i64(<4 x i64> %a) {
+; CHECK: 'Cost Model Analysis' for function 'splatconstant_shift_v4i64':
+; SSE2: Found an estimated cost of 2 for instruction: %shift
+; SSE41: Found an estimated cost of 2 for instruction: %shift
+; AVX: Found an estimated cost of 8 for instruction: %shift
+; AVX2: Found an estimated cost of 1 for instruction: %shift
+; XOPAVX: Found an estimated cost of 2 for instruction: %shift
+; XOPAVX2: Found an estimated cost of 1 for instruction: %shift
+ %shift = shl <4 x i64> %a, <i64 7, i64 7, i64 7, i64 7>
+ ret <4 x i64> %shift
+}
+
+define <4 x i32> @splatconstant_shift_v4i32(<4 x i32> %a) {
+; CHECK: 'Cost Model Analysis' for function 'splatconstant_shift_v4i32':
+; SSE2: Found an estimated cost of 1 for instruction: %shift
+; SSE41: Found an estimated cost of 1 for instruction: %shift
+; AVX: Found an estimated cost of 1 for instruction: %shift
+; AVX2: Found an estimated cost of 1 for instruction: %shift
+; XOPAVX: Found an estimated cost of 1 for instruction: %shift
+; XOPAVX2: Found an estimated cost of 1 for instruction: %shift
+ %shift = shl <4 x i32> %a, <i32 5, i32 5, i32 5, i32 5>
+ ret <4 x i32> %shift
+}
+
+define <8 x i32> @splatconstant_shift_v8i32(<8 x i32> %a) {
+; CHECK: 'Cost Model Analysis' for function 'splatconstant_shift_v8i32':
+; SSE2: Found an estimated cost of 2 for instruction: %shift
+; SSE41: Found an estimated cost of 2 for instruction: %shift
+; AVX: Found an estimated cost of 2 for instruction: %shift
+; AVX2: Found an estimated cost of 1 for instruction: %shift
+; XOPAVX: Found an estimated cost of 2 for instruction: %shift
+; XOPAVX2: Found an estimated cost of 1 for instruction: %shift
+ %shift = shl <8 x i32> %a, <i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5, i32 5>
+ ret <8 x i32> %shift
+}
+
+define <8 x i16> @splatconstant_shift_v8i16(<8 x i16> %a) {
+; CHECK: 'Cost Model Analysis' for function 'splatconstant_shift_v8i16':
+; SSE2: Found an estimated cost of 1 for instruction: %shift
+; SSE41: Found an estimated cost of 1 for instruction: %shift
+; AVX: Found an estimated cost of 1 for instruction: %shift
+; AVX2: Found an estimated cost of 1 for instruction: %shift
+; XOP: Found an estimated cost of 1 for instruction: %shift
+ %shift = shl <8 x i16> %a, <i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3>
+ ret <8 x i16> %shift
+}
+
+define <16 x i16> @splatconstant_shift_v16i16(<16 x i16> %a) {
+; CHECK: 'Cost Model Analysis' for function 'splatconstant_shift_v16i16':
+; SSE2: Found an estimated cost of 2 for instruction: %shift
+; SSE41: Found an estimated cost of 2 for instruction: %shift
+; AVX: Found an estimated cost of 2 for instruction: %shift
+; AVX2: Found an estimated cost of 1 for instruction: %shift
+; XOPAVX: Found an estimated cost of 2 for instruction: %shift
+; XOPAVX2: Found an estimated cost of 1 for instruction: %shift
+ %shift = shl <16 x i16> %a, <i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3, i16 3>
+ ret <16 x i16> %shift
+}
+
+define <16 x i8> @splatconstant_shift_v16i8(<16 x i8> %a) {
+; CHECK: 'Cost Model Analysis' for function 'splatconstant_shift_v16i8':
+; SSE2: Found an estimated cost of 1 for instruction: %shift
+; SSE41: Found an estimated cost of 1 for instruction: %shift
+; AVX: Found an estimated cost of 1 for instruction: %shift
+; AVX2: Found an estimated cost of 1 for instruction: %shift
+; XOP: Found an estimated cost of 1 for instruction: %shift
+ %shift = shl <16 x i8> %a, <i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3>
+ ret <16 x i8> %shift
+}
+
+define <32 x i8> @splatconstant_shift_v32i8(<32 x i8> %a) {
+; CHECK: 'Cost Model Analysis' for function 'splatconstant_shift_v32i8':
+; SSE2: Found an estimated cost of 2 for instruction: %shift
+; SSE41: Found an estimated cost of 2 for instruction: %shift
+; AVX: Found an estimated cost of 2 for instruction: %shift
+; AVX2: Found an estimated cost of 11 for instruction: %shift
+; XOP: Found an estimated cost of 2 for instruction: %shift
+ %shift = shl <32 x i8> %a, <i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3, i8 3>
+ ret <32 x i8> %shift
+}
+
+;
+; Special Cases
+;
+
; We always emit a single pmullw in the case of v8i16 vector shifts by
; non-uniform constant.
More information about the llvm-commits
mailing list