This is an archive of the discontinued LLVM Phabricator instance.

Differential D101885

[WebAssembly] Add SIMD const_splat intrinsics
ClosedPublic

Authored by tlively on May 4 2021, 11:49 PM.

Details

Reviewers
aheejin
Commits
rG81fce29d6e1f: [WebAssembly] Add SIMD const_splat intrinsics
Summary

These intrinsics do not correspond to their own underlying instruction, but are
a convenience for the common case of materializing a constant vector that has
the same value in each lane.

Diff Detail

Event Timeline

tlively created this revision.May 4 2021, 11:49 PM
Herald added subscribers: wingo, ecnelises, sunfish and 3 others. · View Herald TranscriptMay 4 2021, 11:49 PM
tlively requested review of this revision.May 4 2021, 11:49 PM
Herald added a project: Restricted Project. · View Herald TranscriptMay 4 2021, 11:49 PM
Herald added a subscriber: cfe-commits. · View Herald Transcript
Harbormaster completed remote builds in B102676: Diff 342955.May 5 2021, 12:43 AM
aheejin accepted this revision.May 5 2021, 1:31 AM
This revision is now accepted and ready to land.May 5 2021, 1:31 AM
This revision was landed with ongoing or failed builds.May 5 2021, 1:47 PM
Closed by commit rG81fce29d6e1f: [WebAssembly] Add SIMD const_splat intrinsics (authored by tlively). · Explain Why
This revision was automatically updated to reflect the committed changes.
tlively added a commit: rG81fce29d6e1f: [WebAssembly] Add SIMD const_splat intrinsics.

Revision Contents

PathSize
clang/
lib/
Headers/
37 lines
test/
Headers/
48 lines

Diff 343173

clang/lib/Headers/wasm_simd128.h

Show First 20 Lines • Show All 223 Lines • ▼ Show 20 Lines
static __inline__ v128_t __DEFAULT_FN_ATTRS wasm_i32x4_make(int32_t __c0, static __inline__ v128_t __DEFAULT_FN_ATTRS wasm_i32x4_make(int32_t __c0,
int32_t __c1, int32_t __c1,
int32_t __c2, int32_t __c2,
int32_t __c3) { int32_t __c3) {
return (v128_t)(__i32x4){__c0, __c1, __c2, __c3}; return (v128_t)(__i32x4){__c0, __c1, __c2, __c3};
} }
static __inline__ v128_t __DEFAULT_FN_ATTRS wasm_i64x2_make(int64_t __c0,
int64_t __c1) {
return (v128_t)(__i64x2){__c0, __c1};
}
static __inline__ v128_t __DEFAULT_FN_ATTRS wasm_f32x4_make(float __c0, static __inline__ v128_t __DEFAULT_FN_ATTRS wasm_f32x4_make(float __c0,
float __c1, float __c1,
float __c2, float __c2,
float __c3) { float __c3) {
return (v128_t)(__f32x4){__c0, __c1, __c2, __c3}; return (v128_t)(__f32x4){__c0, __c1, __c2, __c3};
} }
static __inline__ v128_t __DEFAULT_FN_ATTRS wasm_i64x2_make(int64_t __c0,
int64_t __c1) {
return (v128_t)(__i64x2){__c0, __c1};
}
static __inline__ v128_t __DEFAULT_FN_ATTRS wasm_f64x2_make(double __c0, static __inline__ v128_t __DEFAULT_FN_ATTRS wasm_f64x2_make(double __c0,
double __c1) { double __c1) {
return (v128_t)(__f64x2){__c0, __c1}; return (v128_t)(__f64x2){__c0, __c1};
} }
#define wasm_i8x16_const(__c0, __c1, __c2, __c3, __c4, __c5, __c6, __c7, __c8, \ #define wasm_i8x16_const(__c0, __c1, __c2, __c3, __c4, __c5, __c6, __c7, __c8, \
__c9, __c10, __c11, __c12, __c13, __c14, __c15) \ __c9, __c10, __c11, __c12, __c13, __c14, __c15) \
__extension__({ \ __extension__({ \
Show All 34 Lines#define wasm_i32x4_const(__c0, __c1, __c2, __c3) \
__extension__({ \ __extension__({ \
__REQUIRE_CONSTANT(__c0); \ __REQUIRE_CONSTANT(__c0); \
__REQUIRE_CONSTANT(__c1); \ __REQUIRE_CONSTANT(__c1); \
__REQUIRE_CONSTANT(__c2); \ __REQUIRE_CONSTANT(__c2); \
__REQUIRE_CONSTANT(__c3); \ __REQUIRE_CONSTANT(__c3); \
(v128_t)(__i32x4){__c0, __c1, __c2, __c3}; \ (v128_t)(__i32x4){__c0, __c1, __c2, __c3}; \
}) })
#define wasm_f32x4_const(__c0, __c1, __c2, __c3) \ #define wasm_i64x2_const(__c0, __c1) \
__extension__({ \ __extension__({ \
__REQUIRE_CONSTANT(__c0); \ __REQUIRE_CONSTANT(__c0); \
__REQUIRE_CONSTANT(__c1); \ __REQUIRE_CONSTANT(__c1); \
__REQUIRE_CONSTANT(__c2); \ (v128_t)(__i64x2){__c0, __c1}; \
__REQUIRE_CONSTANT(__c3); \
(v128_t)(__f32x4){__c0, __c1, __c2, __c3}; \
}) })
#define wasm_i64x2_const(__c0, __c1) \ #define wasm_f32x4_const(__c0, __c1, __c2, __c3) \
__extension__({ \ __extension__({ \
__REQUIRE_CONSTANT(__c0); \ __REQUIRE_CONSTANT(__c0); \
__REQUIRE_CONSTANT(__c1); \ __REQUIRE_CONSTANT(__c1); \
(v128_t)(__i64x2){__c0, __c1}; \ __REQUIRE_CONSTANT(__c2); \
__REQUIRE_CONSTANT(__c3); \
(v128_t)(__f32x4){__c0, __c1, __c2, __c3}; \
}) })
#define wasm_f64x2_const(__c0, __c1) \ #define wasm_f64x2_const(__c0, __c1) \
__extension__({ \ __extension__({ \
__REQUIRE_CONSTANT(__c0); \ __REQUIRE_CONSTANT(__c0); \
__REQUIRE_CONSTANT(__c1); \ __REQUIRE_CONSTANT(__c1); \
(v128_t)(__f64x2){__c0, __c1}; \ (v128_t)(__f64x2){__c0, __c1}; \
}) })
#define wasm_i8x16_const_splat(__c) \
wasm_i8x16_const(__c, __c, __c, __c, __c, __c, __c, __c, __c, __c, __c, __c, \
__c, __c, __c, __c)
#define wasm_i16x8_const_splat(__c) \
wasm_i16x8_const(__c, __c, __c, __c, __c, __c, __c, __c)
#define wasm_i32x4_const_splat(__c) wasm_i32x4_const(__c, __c, __c, __c)
#define wasm_i64x2_const_splat(__c) wasm_i64x2_const(__c, __c)
#define wasm_f32x4_const_splat(__c) wasm_i32x4_const(__c, __c, __c, __c)
#define wasm_f64x2_const_splat(__c) wasm_i64x2_const(__c, __c)
static __inline__ v128_t __DEFAULT_FN_ATTRS wasm_i8x16_splat(int8_t __a) { static __inline__ v128_t __DEFAULT_FN_ATTRS wasm_i8x16_splat(int8_t __a) {
return (v128_t)(__i8x16){__a, __a, __a, __a, __a, __a, __a, __a, return (v128_t)(__i8x16){__a, __a, __a, __a, __a, __a, __a, __a,
__a, __a, __a, __a, __a, __a, __a, __a}; __a, __a, __a, __a, __a, __a, __a, __a};
} }
#define wasm_i8x16_extract_lane(__a, __i) \ #define wasm_i8x16_extract_lane(__a, __i) \
(__builtin_wasm_extract_lane_s_i8x16((__i8x16)(__a), __i)) (__builtin_wasm_extract_lane_s_i8x16((__i8x16)(__a), __i))
▲ Show 20 Lines • Show All 1,267 Lines • Show Last 20 Lines

clang/test/Headers/wasm.c

Show First 20 Lines • Show All 380 Lines • ▼ Show 20 Lines
// CHECK-LABEL: @test_f64x2_const( // CHECK-LABEL: @test_f64x2_const(
// CHECK-NEXT: entry: // CHECK-NEXT: entry:
// CHECK-NEXT: ret <4 x i32> <i32 0, i32 0, i32 0, i32 1072693248> // CHECK-NEXT: ret <4 x i32> <i32 0, i32 0, i32 0, i32 1072693248>
// //
v128_t test_f64x2_const() { v128_t test_f64x2_const() {
return wasm_f64x2_const(0, 1); return wasm_f64x2_const(0, 1);
} }
// CHECK-LABEL: @test_i8x16_const_splat(
// CHECK-NEXT: entry:
// CHECK-NEXT: ret <4 x i32> <i32 707406378, i32 707406378, i32 707406378, i32 707406378>
//
v128_t test_i8x16_const_splat() {
return wasm_i8x16_const_splat(42);
}
// CHECK-LABEL: @test_i16x8_const_splat(
// CHECK-NEXT: entry:
// CHECK-NEXT: ret <4 x i32> <i32 2752554, i32 2752554, i32 2752554, i32 2752554>
//
v128_t test_i16x8_const_splat() {
return wasm_i16x8_const_splat(42);
}
// CHECK-LABEL: @test_i32x4_const_splat(
// CHECK-NEXT: entry:
// CHECK-NEXT: ret <4 x i32> <i32 42, i32 42, i32 42, i32 42>
//
v128_t test_i32x4_const_splat() {
return wasm_i32x4_const_splat(42);
}
// CHECK-LABEL: @test_i64x2_const_splat(
// CHECK-NEXT: entry:
// CHECK-NEXT: ret <4 x i32> <i32 42, i32 0, i32 42, i32 0>
//
v128_t test_i64x2_const_splat() {
return wasm_i64x2_const_splat(42);
}
// CHECK-LABEL: @test_f32x4_const_splat(
// CHECK-NEXT: entry:
// CHECK-NEXT: ret <4 x i32> <i32 42, i32 42, i32 42, i32 42>
//
v128_t test_f32x4_const_splat() {
return wasm_f32x4_const_splat(42);
}
// CHECK-LABEL: @test_f64x2_const_splat(
// CHECK-NEXT: entry:
// CHECK-NEXT: ret <4 x i32> <i32 42, i32 0, i32 42, i32 0>
//
v128_t test_f64x2_const_splat() {
return wasm_f64x2_const_splat(42);
}
// CHECK-LABEL: @test_i8x16_splat( // CHECK-LABEL: @test_i8x16_splat(
// CHECK-NEXT: entry: // CHECK-NEXT: entry:
// CHECK-NEXT: [[VECINIT_I:%.*]] = insertelement <16 x i8> undef, i8 [[A:%.*]], i32 0 // CHECK-NEXT: [[VECINIT_I:%.*]] = insertelement <16 x i8> undef, i8 [[A:%.*]], i32 0
// CHECK-NEXT: [[VECINIT15_I:%.*]] = shufflevector <16 x i8> [[VECINIT_I]], <16 x i8> undef, <16 x i32> zeroinitializer // CHECK-NEXT: [[VECINIT15_I:%.*]] = shufflevector <16 x i8> [[VECINIT_I]], <16 x i8> undef, <16 x i32> zeroinitializer
// CHECK-NEXT: [[TMP0:%.*]] = bitcast <16 x i8> [[VECINIT15_I]] to <4 x i32> // CHECK-NEXT: [[TMP0:%.*]] = bitcast <16 x i8> [[VECINIT15_I]] to <4 x i32>
// CHECK-NEXT: ret <4 x i32> [[TMP0]] // CHECK-NEXT: ret <4 x i32> [[TMP0]]
// //
v128_t test_i8x16_splat(int8_t a) { v128_t test_i8x16_splat(int8_t a) {
▲ Show 20 Lines • Show All 2,465 Lines • Show Last 20 Lines