Thanks for using Compiler Explorer
Sponsors
Jakt
C++
Ada
Analysis
Android Java
Android Kotlin
Assembly
C
C3
Carbon
C++ (Circle)
CIRCT
Clean
CMake
CMakeScript
COBOL
C++ for OpenCL
MLIR
Cppx
Cppx-Blue
Cppx-Gold
Cpp2-cppfront
Crystal
C#
CUDA C++
D
Dart
Elixir
Erlang
Fortran
F#
GLSL
Go
Haskell
HLSL
Hook
Hylo
IL
ispc
Java
Julia
Kotlin
LLVM IR
LLVM MIR
Modula-2
Nim
Objective-C
Objective-C++
OCaml
Odin
OpenCL C
Pascal
Pony
Python
Racket
Ruby
Rust
Snowball
Scala
Slang
Solidity
Spice
SPIR-V
Swift
LLVM TableGen
Toit
TypeScript Native
V
Vala
Visual Basic
Vyper
WASM
Zig
Javascript
GIMPLE
Ygen
llvm source #1
Output
Compile to binary object
Link to binary
Execute the code
Intel asm syntax
Demangle identifiers
Verbose demangling
Filters
Unused labels
Library functions
Directives
Comments
Horizontal whitespace
Debug intrinsics
Compiler
clang (assertions trunk)
clang (trunk)
clang 10.0.0
clang 10.0.1
clang 11.0.0
clang 11.0.1
clang 12.0.0
clang 12.0.1
clang 13.0.0
clang 14.0.0
clang 15.0.0
clang 16.0.0
clang 17.0.1
clang 18.1.0
clang 19.1.0
clang 4.0.1
clang 5.0.0
clang 6.0.0
clang 7.0.0
clang 8.0.0
clang 9.0.0
llc (assertions trunk)
llc (trunk)
llc 10.0.0
llc 10.0.1
llc 11.0.0
llc 11.0.1
llc 12.0.0
llc 12.0.1
llc 13.0.0
llc 14.0.0
llc 15.0.0
llc 16.0.0
llc 17.0.1
llc 18.1.0
llc 19.1.0
llc 3.2
llc 3.3
llc 3.9.1
llc 4.0.0
llc 4.0.1
llc 5.0.0
llc 6.0.0
llc 7.0.0
llc 8.0.0
llc 9.0.0
opt (assertions trunk)
opt (trunk)
opt 10.0.0
opt 10.0.1
opt 11.0.0
opt 11.0.1
opt 12.0.0
opt 12.0.1
opt 13.0.0
opt 14.0.0
opt 15.0.0
opt 16.0.0
opt 17.0.1
opt 18.1.0
opt 19.1.0
opt 3.2
opt 3.3
opt 3.9.1
opt 4.0.0
opt 4.0.1
opt 5.0.0
opt 6.0.0
opt 7.0.0
opt 8.0.0
opt 9.0.0
Options
Source code
define void @load_i32_stride6_vf16(<96 x i32>* %in.vec, <16 x i32>* %out.vec0, <16 x i32>* %out.vec1, <16 x i32>* %out.vec2, <16 x i32>* %out.vec3, <16 x i32>* %out.vec4, <16 x i32>* %out.vec5) nounwind { %wide.vec = load <96 x i32>, <96 x i32>* %in.vec, align 32 tail call void asm sideeffect "# LLVM-MCA-BEGIN", "~{dirflag},~{fpsr},~{flags}"() %strided.vec0 = shufflevector <96 x i32> %wide.vec, <96 x i32> poison, <16 x i32> <i32 0, i32 6, i32 12, i32 18, i32 24, i32 30, i32 36, i32 42, i32 48, i32 54, i32 60, i32 66, i32 72, i32 78, i32 84, i32 90> %strided.vec1 = shufflevector <96 x i32> %wide.vec, <96 x i32> poison, <16 x i32> <i32 1, i32 7, i32 13, i32 19, i32 25, i32 31, i32 37, i32 43, i32 49, i32 55, i32 61, i32 67, i32 73, i32 79, i32 85, i32 91> %strided.vec2 = shufflevector <96 x i32> %wide.vec, <96 x i32> poison, <16 x i32> <i32 2, i32 8, i32 14, i32 20, i32 26, i32 32, i32 38, i32 44, i32 50, i32 56, i32 62, i32 68, i32 74, i32 80, i32 86, i32 92> %strided.vec3 = shufflevector <96 x i32> %wide.vec, <96 x i32> poison, <16 x i32> <i32 3, i32 9, i32 15, i32 21, i32 27, i32 33, i32 39, i32 45, i32 51, i32 57, i32 63, i32 69, i32 75, i32 81, i32 87, i32 93> %strided.vec4 = shufflevector <96 x i32> %wide.vec, <96 x i32> poison, <16 x i32> <i32 4, i32 10, i32 16, i32 22, i32 28, i32 34, i32 40, i32 46, i32 52, i32 58, i32 64, i32 70, i32 76, i32 82, i32 88, i32 94> %strided.vec5 = shufflevector <96 x i32> %wide.vec, <96 x i32> poison, <16 x i32> <i32 5, i32 11, i32 17, i32 23, i32 29, i32 35, i32 41, i32 47, i32 53, i32 59, i32 65, i32 71, i32 77, i32 83, i32 89, i32 95> tail call void asm sideeffect "# LLVM-MCA-END", "~{dirflag},~{fpsr},~{flags}"() store <16 x i32> %strided.vec0, <16 x i32>* %out.vec0, align 32 store <16 x i32> %strided.vec1, <16 x i32>* %out.vec1, align 32 store <16 x i32> %strided.vec2, <16 x i32>* %out.vec2, align 32 store <16 x i32> %strided.vec3, <16 x i32>* %out.vec3, align 32 store <16 x i32> %strided.vec4, <16 x i32>* %out.vec4, align 32 store <16 x i32> %strided.vec5, <16 x i32>* %out.vec5, align 32 ret void }
analysis source #2
Output
Compile to binary object
Link to binary
Execute the code
Intel asm syntax
Demangle identifiers
Verbose demangling
Filters
Unused labels
Library functions
Directives
Comments
Horizontal whitespace
Debug intrinsics
Compiler
OSACA (0.6.1)
llvm-mca (assertions trunk)
llvm-mca (trunk)
Options
Source code
vmovaps xmm14, xmmword ptr [rip + .LCPI0_0] # xmm14 = <0,6,4,u> vblendps ymm10, ymm0, ymm1, 48 # ymm10 = ymm0[0,1,2,3],ymm1[4,5],ymm0[6,7] vpermps ymm0, ymm14, ymm10 vperm2f128 ymm11, ymm2, ymm5, 32 # ymm11 = ymm2[0,1],ymm5[0,1] vpermilps ymm12, ymm11, 164 # ymm12 = ymm11[0,1,2,2,4,5,6,6] vpermilps ymm13, ymm5, 170 # ymm13 = ymm5[2,2,2,2,6,6,6,6] vblendps ymm12, ymm12, ymm13, 34 # ymm12 = ymm12[0],ymm13[1],ymm12[2,3,4],ymm13[5],ymm12[6,7] vblendps ymm0, ymm0, ymm12, 56 # ymm0 = ymm0[0,1,2],ymm12[3,4,5],ymm0[6,7] vblendps ymm12, ymm6, ymm7, 48 # ymm12 = ymm6[0,1,2,3],ymm7[4,5],ymm6[6,7] vbroadcastsd ymm9, qword ptr [rip + .LCPI0_1] # ymm9 = [8589934596,8589934596,8589934596,8589934596] vpermps ymm13, ymm9, ymm12 vblendps ymm0, ymm0, ymm13, 192 # ymm0 = ymm0[0,1,2,3,4,5],ymm13[6,7] vperm2f128 ymm3, ymm3, ymm8, 32 # ymm3 = ymm3[0,1],ymm8[0,1] vpermilps ymm1, ymm3, 164 # ymm1 = ymm3[0,1,2,2,4,5,6,6] vpermilps ymm13, ymm8, 170 # ymm13 = ymm8[2,2,2,2,6,6,6,6] vblendps ymm13, ymm1, ymm13, 34 # ymm13 = ymm1[0],ymm13[1],ymm1[2,3,4],ymm13[5],ymm1[6,7] vblendps ymm0, ymm4, ymm15, 48 # ymm0 = ymm4[0,1,2,3],ymm15[4,5],ymm4[6,7] vpermps ymm14, ymm14, ymm0 vblendps ymm1, ymm14, ymm13, 56 # ymm1 = ymm14[0,1,2],ymm13[3,4,5],ymm14[6,7] vblendps ymm2, ymm13, ymm14, 48 # ymm2 = ymm13[0,1,2,3],ymm14[4,5],ymm13[6,7] vpermps ymm9, ymm9, ymm2 vblendps ymm1, ymm1, ymm9, 192 # ymm1 = ymm1[0,1,2,3,4,5],ymm9[6,7] vpermilps ymm1, ymm11, 229 # ymm1 = ymm11[1,1,2,3,5,5,6,7] vpermilps ymm9, ymm5, 238 # ymm9 = ymm5[2,3,2,3,6,7,6,7] vblendps ymm1, ymm1, ymm9, 34 # ymm1 = ymm1[0],ymm9[1],ymm1[2,3,4],ymm9[5],ymm1[6,7] vmovaps xmm9, xmmword ptr [rip + .LCPI0_2] # xmm9 = <1,7,5,u> vpermps ymm10, ymm9, ymm10 vblendps ymm1, ymm10, ymm1, 56 # ymm1 = ymm10[0,1,2],ymm1[3,4,5],ymm10[6,7] vbroadcastsd ymm10, qword ptr [rip + .LCPI0_3] # ymm10 = [12884901893,12884901893,12884901893,12884901893] vpermps ymm11, ymm10, ymm12 vblendps ymm1, ymm1, ymm11, 192 # ymm1 = ymm1[0,1,2,3,4,5],ymm11[6,7] vpermps ymm0, ymm9, ymm0 vpermilps ymm1, ymm3, 229 # ymm1 = ymm3[1,1,2,3,5,5,6,7] vpermilps ymm3, ymm8, 238 # ymm3 = ymm8[2,3,2,3,6,7,6,7] vblendps ymm1, ymm1, ymm3, 34 # ymm1 = ymm1[0],ymm3[1],ymm1[2,3,4],ymm3[5],ymm1[6,7] vblendps ymm0, ymm0, ymm1, 56 # ymm0 = ymm0[0,1,2],ymm1[3,4,5],ymm0[6,7] vpermps ymm1, ymm10, ymm2 vblendps ymm0, ymm0, ymm1, 192 # ymm0 = ymm0[0,1,2,3,4,5],ymm1[6,7] vpermilps ymm0, ymm10, 238 # ymm0 = ymm10[2,3,2,3,6,7,6,7] vpermilps ymm1, ymm12, 224 # ymm1 = ymm12[0,0,2,3,4,4,6,7] vblendps ymm0, ymm1, ymm0, 17 # ymm0 = ymm0[0],ymm1[1,2,3],ymm0[4],ymm1[5,6,7] vblendps ymm1, ymm9, ymm5, 204 # ymm1 = ymm9[0,1],ymm5[2,3],ymm9[4,5],ymm5[6,7] vpermilps ymm1, ymm1, 226 # ymm1 = ymm1[2,0,2,3,6,4,6,7] vpermpd ymm1, ymm1, 200 # ymm1 = ymm1[0,2,0,3] vpermpd ymm0, ymm0, 236 # ymm0 = ymm0[0,3,2,3] vblendps ymm0, ymm0, ymm1, 24 # ymm0 = ymm0[0,1,2],ymm1[3,4],ymm0[5,6,7] vpermilps ymm1, ymm6, 0 # ymm1 = ymm6[0,0,0,0,4,4,4,4] vpermilps ymm2, ymm7, 224 # ymm2 = ymm7[0,0,2,3,4,4,6,7] vblendps ymm1, ymm2, ymm1, 136 # ymm1 = ymm2[0,1,2],ymm1[3],ymm2[4,5,6],ymm1[7] vpermpd ymm1, ymm1, 196 # ymm1 = ymm1[0,1,0,3] vblendps ymm0, ymm0, ymm1, 224 # ymm0 = ymm0[0,1,2,3,4],ymm1[5,6,7] vpermilps ymm0, ymm4, 238 # ymm0 = ymm4[2,3,2,3,6,7,6,7] vpermilps ymm1, ymm15, 224 # ymm1 = ymm15[0,0,2,3,4,4,6,7] vblendps ymm0, ymm1, ymm0, 17 # ymm0 = ymm0[0],ymm1[1,2,3],ymm0[4],ymm1[5,6,7] vblendps ymm1, ymm3, ymm8, 204 # ymm1 = ymm3[0,1],ymm8[2,3],ymm3[4,5],ymm8[6,7] vpermilps ymm1, ymm1, 226 # ymm1 = ymm1[2,0,2,3,6,4,6,7] vpermpd ymm1, ymm1, 200 # ymm1 = ymm1[0,2,0,3] vpermpd ymm0, ymm0, 236 # ymm0 = ymm0[0,3,2,3] vblendps ymm0, ymm0, ymm1, 24 # ymm0 = ymm0[0,1,2],ymm1[3,4],ymm0[5,6,7] vpermilps ymm1, ymm13, 0 # ymm1 = ymm13[0,0,0,0,4,4,4,4] vpermilps ymm2, ymm14, 224 # ymm2 = ymm14[0,0,2,3,4,4,6,7] vblendps ymm1, ymm2, ymm1, 136 # ymm1 = ymm2[0,1,2],ymm1[3],ymm2[4,5,6],ymm1[7] vpermpd ymm1, ymm1, 196 # ymm1 = ymm1[0,1,0,3] vblendps ymm0, ymm0, ymm1, 224 # ymm0 = ymm0[0,1,2,3,4],ymm1[5,6,7] vpermilps ymm0, ymm10, 255 # ymm0 = ymm10[3,3,3,3,7,7,7,7] vpermilps ymm1, ymm12, 244 # ymm1 = ymm12[0,1,3,3,4,5,7,7] vblendps ymm0, ymm1, ymm0, 17 # ymm0 = ymm0[0],ymm1[1,2,3],ymm0[4],ymm1[5,6,7] vpermilps ymm1, ymm5, 255 # ymm1 = ymm5[3,3,3,3,7,7,7,7] vblendps ymm1, ymm1, ymm9, 34 # ymm1 = ymm1[0],ymm9[1],ymm1[2,3,4],ymm9[5],ymm1[6,7] vpermpd ymm1, ymm1, 200 # ymm1 = ymm1[0,2,0,3] vpermpd ymm0, ymm0, 236 # ymm0 = ymm0[0,3,2,3] vblendps ymm0, ymm0, ymm1, 24 # ymm0 = ymm0[0,1,2],ymm1[3,4],ymm0[5,6,7] vpermilps ymm1, ymm6, 68 # ymm1 = ymm6[0,1,0,1,4,5,4,5] vmovaps ymm11, ymm6 vpermilps ymm2, ymm7, 244 # ymm2 = ymm7[0,1,3,3,4,5,7,7] vmovaps ymm10, ymm7 vblendps ymm1, ymm2, ymm1, 136 # ymm1 = ymm2[0,1,2],ymm1[3],ymm2[4,5,6],ymm1[7] vpermpd ymm1, ymm1, 196 # ymm1 = ymm1[0,1,0,3] vblendps ymm0, ymm0, ymm1, 224 # ymm0 = ymm0[0,1,2,3,4],ymm1[5,6,7] vpermilps ymm0, ymm4, 255 # ymm0 = ymm4[3,3,3,3,7,7,7,7] vpermilps ymm1, ymm15, 244 # ymm1 = ymm15[0,1,3,3,4,5,7,7] vblendps ymm0, ymm1, ymm0, 17 # ymm0 = ymm0[0],ymm1[1,2,3],ymm0[4],ymm1[5,6,7] vpermilps ymm1, ymm8, 255 # ymm1 = ymm8[3,3,3,3,7,7,7,7] vblendps ymm1, ymm1, ymm3, 34 # ymm1 = ymm1[0],ymm3[1],ymm1[2,3,4],ymm3[5],ymm1[6,7] vpermpd ymm1, ymm1, 200 # ymm1 = ymm1[0,2,0,3] vpermpd ymm0, ymm0, 236 # ymm0 = ymm0[0,3,2,3] vblendps ymm0, ymm0, ymm1, 24 # ymm0 = ymm0[0,1,2],ymm1[3,4],ymm0[5,6,7] vpermilps ymm1, ymm13, 68 # ymm1 = ymm13[0,1,0,1,4,5,4,5] vpermilps ymm2, ymm14, 244 # ymm2 = ymm14[0,1,3,3,4,5,7,7] vblendps ymm1, ymm2, ymm1, 136 # ymm1 = ymm2[0,1,2],ymm1[3],ymm2[4,5,6],ymm1[7] vpermpd ymm1, ymm1, 196 # ymm1 = ymm1[0,1,0,3] vblendps ymm0, ymm0, ymm1, 224 # ymm0 = ymm0[0,1,2,3,4],ymm1[5,6,7] vblendps ymm1, ymm15, ymm4, 48 # ymm1 = ymm15[0,1,2,3],ymm4[4,5],ymm15[6,7] vblendps ymm2, ymm3, ymm8, 240 # ymm2 = ymm3[0,1,2,3],ymm8[4,5,6,7] vblendps ymm0, ymm2, ymm3, 204 # ymm0 = ymm2[0,1],ymm3[2,3],ymm2[4,5],ymm3[6,7] vpermilps ymm0, ymm0, 132 # ymm0 = ymm0[0,1,0,2,4,5,4,6] vmovaps xmm4, xmmword ptr [rip + .LCPI0_4] # xmm4 = <4,2,u,u> vpermps ymm6, ymm4, ymm1 vblendps ymm0, ymm0, ymm6, 3 # ymm0 = ymm6[0,1],ymm0[2,3,4,5,6,7] vblendps ymm6, ymm13, ymm14, 12 # ymm6 = ymm13[0,1],ymm14[2,3],ymm13[4,5,6,7] vbroadcastf128 ymm8, xmmword ptr [rip + .LCPI0_5] # ymm8 = [0,2,0,6,0,2,0,6] # ymm8 = mem[0,1,0,1] vpermps ymm13, ymm8, ymm6 vblendps ymm0, ymm0, ymm13, 224 # ymm0 = ymm0[0,1,2,3,4],ymm13[5,6,7] vblendps ymm5, ymm9, ymm5, 240 # ymm5 = ymm9[0,1,2,3],ymm5[4,5,6,7] vblendps ymm7, ymm12, ymm12, 48 # 32-byte Folded Reload # ymm7 = ymm12[0,1,2,3],mem[4,5],ymm12[6,7] vblendps ymm14, ymm5, ymm13, 204 # ymm14 = ymm5[0,1],ymm13[2,3],ymm5[4,5],ymm13[6,7] vpermilps ymm14, ymm14, 132 # ymm14 = ymm14[0,1,0,2,4,5,4,6] vpermps ymm4, ymm4, ymm7 vblendps ymm4, ymm14, ymm4, 3 # ymm4 = ymm4[0,1],ymm14[2,3,4,5,6,7] vblendps ymm14, ymm11, ymm10, 12 # ymm14 = ymm11[0,1],ymm10[2,3],ymm11[4,5,6,7] vpermps ymm8, ymm8, ymm14 vblendps ymm4, ymm4, ymm8, 224 # ymm4 = ymm4[0,1,2,3,4],ymm8[5,6,7] vpermilps ymm2, ymm2, 85 # ymm2 = ymm2[1,1,1,1,5,5,5,5] vblendps ymm2, ymm2, ymm3, 136 # ymm2 = ymm2[0,1,2],ymm3[3],ymm2[4,5,6],ymm3[7] vmovaps xmm3, xmmword ptr [rip + .LCPI0_6] # xmm3 = <5,3,u,u> vpermps ymm1, ymm3, ymm1 vblendps ymm1, ymm2, ymm1, 3 # ymm1 = ymm1[0,1],ymm2[2,3,4,5,6,7] vbroadcastf128 ymm2, xmmword ptr [rip + .LCPI0_7] # ymm2 = [0,3,1,7,0,3,1,7] # ymm2 = mem[0,1,0,1] vpermps ymm6, ymm2, ymm6 vblendps ymm1, ymm1, ymm6, 224 # ymm1 = ymm1[0,1,2,3,4],ymm6[5,6,7] vpermilps ymm5, ymm5, 85 # ymm5 = ymm5[1,1,1,1,5,5,5,5] vblendps ymm5, ymm5, ymm13, 136 # ymm5 = ymm5[0,1,2],ymm13[3],ymm5[4,5,6],ymm13[7] vpermps ymm3, ymm3, ymm7 vblendps ymm3, ymm5, ymm3, 3 # ymm3 = ymm3[0,1],ymm5[2,3,4,5,6,7] vpermps ymm2, ymm2, ymm14 vblendps ymm2, ymm3, ymm2, 224 # ymm2 = ymm3[0,1,2,3,4],ymm2[5,6,7]
Become a Patron
Sponsor on GitHub
Donate via PayPal
Source on GitHub
Mailing list
Installed libraries
Wiki
Report an issue
How it works
Contact the author
CE on Mastodon
CE on Bluesky
About the author
Statistics
Changelog
Version tree