Skip to content

Commit

Permalink
Fix log2 for LLVM < 8, bump deps to avoid broken LineNumberNode uses …
Browse files Browse the repository at this point in the history 
…on master, and bump version.
  • Loading branch information
@chriselrod
chriselrod committed Feb 19, 2020
1 parent a908f0c commit 304f30f
Show file tree
Hide file tree
Showing 4 changed files with 87 additions and 44 deletions.
5 changes: 5 additions & 0 deletions .travis.yml
View file
Original file line number Diff line number Diff line change
Expand Up @@ -24,3 +24,8 @@ jobs:
Pkg.instantiate();
include("docs/make.jl");'
after_success: skip
env:
global:
- COVERALLS_PARALLEL=true
notifications:
webhooks: https://coveralls.io/webhook
8 changes: 4 additions & 4 deletions Manifest.toml
View file
Original file line number Diff line number Diff line change
Expand Up @@ -37,11 +37,11 @@ uuid = "9a3f8284-a2c9-5f02-9a11-845980a1fd5c"

[[SIMDPirates]]
deps = ["VectorizationBase"]
git-tree-sha1 = "9bbd51cc7e31471d9957a107300d2edd109c3c42"
git-tree-sha1 = "16e34b8028fa9d419a0e0b7f4d702c9b590a1977"
repo-rev = "master"
repo-url = "https://github.com/chriselrod/SIMDPirates.jl"
uuid = "21efa798-c60a-11e8-04d3-e1a92915a26a"
version = "0.3.8"
version = "0.3.13"

[[Serialization]]
uuid = "9e88b42a-f829-5b0c-bbe9-9e923198166b"
Expand All @@ -55,8 +55,8 @@ uuid = "8dfed614-e22c-5e08-85e1-65c5234f0b40"

[[VectorizationBase]]
deps = ["CpuId", "LinearAlgebra"]
git-tree-sha1 = "74b956a54e2692d4db2f5e53d2b990e440d4fdd5"
git-tree-sha1 = "e1093ff0fc183880a6f836026309ba06672c92ec"
repo-rev = "master"
repo-url = "https://github.com/chriselrod/VectorizationBase.jl"
uuid = "3d5dd08c-fd9d-11e8-17fa-ed2836048c2f"
version = "0.2.5"
version = "0.3.1"
6 changes: 3 additions & 3 deletions Project.toml
View file
Original file line number Diff line number Diff line change
@@ -1,16 +1,16 @@
name = "SLEEFPirates"
uuid = "476501e8-09a2-5ece-8869-fb82de89a1fa"
authors = ["chriselrod <[email protected]>"]
version = "0.3.5"
version = "0.3.6"

[deps]
Libdl = "8f399da3-3557-5675-b5ff-fb832c97cbdb"
SIMDPirates = "21efa798-c60a-11e8-04d3-e1a92915a26a"
VectorizationBase = "3d5dd08c-fd9d-11e8-17fa-ed2836048c2f"

[compat]
SIMDPirates = "0.3, 0.4"
VectorizationBase = "0.2, 0.3"
SIMDPirates = "0.3.13, 0.4, 0.5, 0.6"
VectorizationBase = "0.3.1, 0.3, 0.4, 0.5, 0.6"
julia = "1"

[extras]
Expand Down
112 changes: 75 additions & 37 deletions src/sleef.jl
View file
Original file line number Diff line number Diff line change
Expand Up @@ -31,45 +31,83 @@ declare <8 x double> @llvm.x86.avx512.mask.getmant.pd.512(<8 x double>, i32, <8
"""), Vec{8,Float64}, Tuple{Vec{8,Float64}}, v)
end


@inline function log2(v::Vec{8,Float64})
Base.llvmcall(("""
declare <8 x double> @llvm.fma.v8f64(<8 x double>, <8 x double>, <8 x double>)
declare <8 x double> @llvm.x86.avx512.mask.fixupimm.pd.512(<8 x double>, <8 x double>, <8 x i64>, i32, i8, i32)
declare <8 x double> @llvm.x86.avx512.mask.getexp.pd.512(<8 x double>, <8 x double>, i8, i32)
declare <8 x double> @llvm.x86.avx512.mask.getmant.pd.512(<8 x double>, i32, <8 x double>, i8, i32)
""","""
%2 = fmul <8 x double> %0, <double 0x3FF5555555555555, double 0x3FF5555555555555, double 0x3FF5555555555555, double 0x3FF5555555555555, double 0x3FF5555555555555, double 0x3FF5555555555555, double 0x3FF5555555555555, double 0x3FF5555555555555>
%3 = tail call <8 x double> @llvm.x86.avx512.mask.getexp.pd.512(<8 x double> %2, <8 x double> zeroinitializer, i8 -1, i32 4) #13
%4 = fcmp oeq <8 x double> %3, <double 0x7FF0000000000000, double 0x7FF0000000000000, double 0x7FF0000000000000, double 0x7FF0000000000000, double 0x7FF0000000000000, double 0x7FF0000000000000, double 0x7FF0000000000000, double 0x7FF0000000000000>
%5 = select <8 x i1> %4, <8 x double> <double 1.024000e+03, double 1.024000e+03, double 1.024000e+03, double 1.024000e+03, double 1.024000e+03, double 1.024000e+03, double 1.024000e+03, double 1.024000e+03>, <8 x double> %3
%6 = tail call <8 x double> @llvm.x86.avx512.mask.getmant.pd.512(<8 x double> %0, i32 11, <8 x double> zeroinitializer, i8 -1, i32 4) #13
%7 = fadd <8 x double> %6, <double -1.000000e+00, double -1.000000e+00, double -1.000000e+00, double -1.000000e+00, double -1.000000e+00, double -1.000000e+00, double -1.000000e+00, double -1.000000e+00>
%8 = fadd <8 x double> %6, <double 1.000000e+00, double 1.000000e+00, double 1.000000e+00, double 1.000000e+00, double 1.000000e+00, double 1.000000e+00, double 1.000000e+00, double 1.000000e+00>
%9 = fdiv <8 x double> %7, %8
%10 = fmul <8 x double> %9, %9
%11 = tail call <8 x double> @llvm.fma.v8f64(<8 x double> %10, <8 x double> <double 0x3FCC501739F17BA9, double 0x3FCC501739F17BA9, double 0x3FCC501739F17BA9, double 0x3FCC501739F17BA9, double 0x3FCC501739F17BA9, double 0x3FCC501739F17BA9, double 0x3FCC501739F17BA9, double 0x3FCC501739F17BA9>, <8 x double> <double 0x3FCC2B7A962850E9, double 0x3FCC2B7A962850E9, double 0x3FCC2B7A962850E9, double 0x3FCC2B7A962850E9, double 0x3FCC2B7A962850E9, double 0x3FCC2B7A962850E9, double 0x3FCC2B7A962850E9, double 0x3FCC2B7A962850E9>) #13
%12 = tail call <8 x double> @llvm.fma.v8f64(<8 x double> %11, <8 x double> %10, <8 x double> <double 0x3FD0CAAEEB877481, double 0x3FD0CAAEEB877481, double 0x3FD0CAAEEB877481, double 0x3FD0CAAEEB877481, double 0x3FD0CAAEEB877481, double 0x3FD0CAAEEB877481, double 0x3FD0CAAEEB877481, double 0x3FD0CAAEEB877481>) #13
%13 = tail call <8 x double> @llvm.fma.v8f64(<8 x double> %12, <8 x double> %10, <8 x double> <double 0x3FD484AC6A7CB2DD, double 0x3FD484AC6A7CB2DD, double 0x3FD484AC6A7CB2DD, double 0x3FD484AC6A7CB2DD, double 0x3FD484AC6A7CB2DD, double 0x3FD484AC6A7CB2DD, double 0x3FD484AC6A7CB2DD, double 0x3FD484AC6A7CB2DD>) #13
%14 = tail call <8 x double> @llvm.fma.v8f64(<8 x double> %13, <8 x double> %10, <8 x double> <double 0x3FDA617636C2C254, double 0x3FDA617636C2C254, double 0x3FDA617636C2C254, double 0x3FDA617636C2C254, double 0x3FDA617636C2C254, double 0x3FDA617636C2C254, double 0x3FDA617636C2C254, double 0x3FDA617636C2C254>) #13
%15 = tail call <8 x double> @llvm.fma.v8f64(<8 x double> %14, <8 x double> %10, <8 x double> <double 0x3FE2776C50E7EDE9, double 0x3FE2776C50E7EDE9, double 0x3FE2776C50E7EDE9, double 0x3FE2776C50E7EDE9, double 0x3FE2776C50E7EDE9, double 0x3FE2776C50E7EDE9, double 0x3FE2776C50E7EDE9, double 0x3FE2776C50E7EDE9>) #13
%16 = tail call <8 x double> @llvm.fma.v8f64(<8 x double> %15, <8 x double> %10, <8 x double> <double 0x3FEEC709DC3A07B2, double 0x3FEEC709DC3A07B2, double 0x3FEEC709DC3A07B2, double 0x3FEEC709DC3A07B2, double 0x3FEEC709DC3A07B2, double 0x3FEEC709DC3A07B2, double 0x3FEEC709DC3A07B2, double 0x3FEEC709DC3A07B2>) #13
%17 = fmul <8 x double> %9, <double 0x40071547652B82FE, double 0x40071547652B82FE, double 0x40071547652B82FE, double 0x40071547652B82FE, double 0x40071547652B82FE, double 0x40071547652B82FE, double 0x40071547652B82FE, double 0x40071547652B82FE>
%18 = fneg <8 x double> %17
%19 = tail call <8 x double> @llvm.fma.v8f64(<8 x double> %9, <8 x double> <double 0x40071547652B82FE, double 0x40071547652B82FE, double 0x40071547652B82FE, double 0x40071547652B82FE, double 0x40071547652B82FE, double 0x40071547652B82FE, double 0x40071547652B82FE, double 0x40071547652B82FE>, <8 x double> %18) #13
%20 = fadd <8 x double> %5, %17
%21 = fsub <8 x double> %5, %20
%22 = fadd <8 x double> %17, %21
%23 = fadd <8 x double> %19, %22
%24 = fmul <8 x double> %9, %10
%25 = fadd <8 x double> %20, %23
%26 = tail call <8 x double> @llvm.fma.v8f64(<8 x double> %16, <8 x double> %24, <8 x double> %25) #13
%27 = tail call <8 x double> @llvm.x86.avx512.mask.fixupimm.pd.512(<8 x double> %26, <8 x double> %0, <8 x i64> <i64 167482009228346368, i64 167482009228346368, i64 167482009228346368, i64 167482009228346368, i64 167482009228346368, i64 167482009228346368, i64 167482009228346368, i64 167482009228346368>, i32 0, i8 -1, i32 4)
ret <8 x double> %27
"""), Vec{8,Float64}, Tuple{Vec{8,Float64}}, v)
@static if Base.libllvm_version > v"8"
# Support different LLVM versions. Only difference is fneg in llvm 8+
@inline function log2(v::Vec{8,Float64})
Base.llvmcall(("""
declare <8 x double> @llvm.fma.v8f64(<8 x double>, <8 x double>, <8 x double>)
declare <8 x double> @llvm.x86.avx512.mask.fixupimm.pd.512(<8 x double>, <8 x double>, <8 x i64>, i32, i8, i32)
declare <8 x double> @llvm.x86.avx512.mask.getexp.pd.512(<8 x double>, <8 x double>, i8, i32)
declare <8 x double> @llvm.x86.avx512.mask.getmant.pd.512(<8 x double>, i32, <8 x double>, i8, i32)
""","""
%2 = fmul <8 x double> %0, <double 0x3FF5555555555555, double 0x3FF5555555555555, double 0x3FF5555555555555, double 0x3FF5555555555555, double 0x3FF5555555555555, double 0x3FF5555555555555, double 0x3FF5555555555555, double 0x3FF5555555555555>
%3 = tail call <8 x double> @llvm.x86.avx512.mask.getexp.pd.512(<8 x double> %2, <8 x double> zeroinitializer, i8 -1, i32 4) #13
%4 = fcmp oeq <8 x double> %3, <double 0x7FF0000000000000, double 0x7FF0000000000000, double 0x7FF0000000000000, double 0x7FF0000000000000, double 0x7FF0000000000000, double 0x7FF0000000000000, double 0x7FF0000000000000, double 0x7FF0000000000000>
%5 = select <8 x i1> %4, <8 x double> <double 1.024000e+03, double 1.024000e+03, double 1.024000e+03, double 1.024000e+03, double 1.024000e+03, double 1.024000e+03, double 1.024000e+03, double 1.024000e+03>, <8 x double> %3
%6 = tail call <8 x double> @llvm.x86.avx512.mask.getmant.pd.512(<8 x double> %0, i32 11, <8 x double> zeroinitializer, i8 -1, i32 4) #13
%7 = fadd <8 x double> %6, <double -1.000000e+00, double -1.000000e+00, double -1.000000e+00, double -1.000000e+00, double -1.000000e+00, double -1.000000e+00, double -1.000000e+00, double -1.000000e+00>
%8 = fadd <8 x double> %6, <double 1.000000e+00, double 1.000000e+00, double 1.000000e+00, double 1.000000e+00, double 1.000000e+00, double 1.000000e+00, double 1.000000e+00, double 1.000000e+00>
%9 = fdiv <8 x double> %7, %8
%10 = fmul <8 x double> %9, %9
%11 = tail call <8 x double> @llvm.fma.v8f64(<8 x double> %10, <8 x double> <double 0x3FCC501739F17BA9, double 0x3FCC501739F17BA9, double 0x3FCC501739F17BA9, double 0x3FCC501739F17BA9, double 0x3FCC501739F17BA9, double 0x3FCC501739F17BA9, double 0x3FCC501739F17BA9, double 0x3FCC501739F17BA9>, <8 x double> <double 0x3FCC2B7A962850E9, double 0x3FCC2B7A962850E9, double 0x3FCC2B7A962850E9, double 0x3FCC2B7A962850E9, double 0x3FCC2B7A962850E9, double 0x3FCC2B7A962850E9, double 0x3FCC2B7A962850E9, double 0x3FCC2B7A962850E9>) #13
%12 = tail call <8 x double> @llvm.fma.v8f64(<8 x double> %11, <8 x double> %10, <8 x double> <double 0x3FD0CAAEEB877481, double 0x3FD0CAAEEB877481, double 0x3FD0CAAEEB877481, double 0x3FD0CAAEEB877481, double 0x3FD0CAAEEB877481, double 0x3FD0CAAEEB877481, double 0x3FD0CAAEEB877481, double 0x3FD0CAAEEB877481>) #13
%13 = tail call <8 x double> @llvm.fma.v8f64(<8 x double> %12, <8 x double> %10, <8 x double> <double 0x3FD484AC6A7CB2DD, double 0x3FD484AC6A7CB2DD, double 0x3FD484AC6A7CB2DD, double 0x3FD484AC6A7CB2DD, double 0x3FD484AC6A7CB2DD, double 0x3FD484AC6A7CB2DD, double 0x3FD484AC6A7CB2DD, double 0x3FD484AC6A7CB2DD>) #13
%14 = tail call <8 x double> @llvm.fma.v8f64(<8 x double> %13, <8 x double> %10, <8 x double> <double 0x3FDA617636C2C254, double 0x3FDA617636C2C254, double 0x3FDA617636C2C254, double 0x3FDA617636C2C254, double 0x3FDA617636C2C254, double 0x3FDA617636C2C254, double 0x3FDA617636C2C254, double 0x3FDA617636C2C254>) #13
%15 = tail call <8 x double> @llvm.fma.v8f64(<8 x double> %14, <8 x double> %10, <8 x double> <double 0x3FE2776C50E7EDE9, double 0x3FE2776C50E7EDE9, double 0x3FE2776C50E7EDE9, double 0x3FE2776C50E7EDE9, double 0x3FE2776C50E7EDE9, double 0x3FE2776C50E7EDE9, double 0x3FE2776C50E7EDE9, double 0x3FE2776C50E7EDE9>) #13
%16 = tail call <8 x double> @llvm.fma.v8f64(<8 x double> %15, <8 x double> %10, <8 x double> <double 0x3FEEC709DC3A07B2, double 0x3FEEC709DC3A07B2, double 0x3FEEC709DC3A07B2, double 0x3FEEC709DC3A07B2, double 0x3FEEC709DC3A07B2, double 0x3FEEC709DC3A07B2, double 0x3FEEC709DC3A07B2, double 0x3FEEC709DC3A07B2>) #13
%17 = fmul <8 x double> %9, <double 0x40071547652B82FE, double 0x40071547652B82FE, double 0x40071547652B82FE, double 0x40071547652B82FE, double 0x40071547652B82FE, double 0x40071547652B82FE, double 0x40071547652B82FE, double 0x40071547652B82FE>
%18 = fneg <8 x double> %17
%19 = tail call <8 x double> @llvm.fma.v8f64(<8 x double> %9, <8 x double> <double 0x40071547652B82FE, double 0x40071547652B82FE, double 0x40071547652B82FE, double 0x40071547652B82FE, double 0x40071547652B82FE, double 0x40071547652B82FE, double 0x40071547652B82FE, double 0x40071547652B82FE>, <8 x double> %18) #13
%20 = fadd <8 x double> %5, %17
%21 = fsub <8 x double> %5, %20
%22 = fadd <8 x double> %17, %21
%23 = fadd <8 x double> %19, %22
%24 = fmul <8 x double> %9, %10
%25 = fadd <8 x double> %20, %23
%26 = tail call <8 x double> @llvm.fma.v8f64(<8 x double> %16, <8 x double> %24, <8 x double> %25) #13
%27 = tail call <8 x double> @llvm.x86.avx512.mask.fixupimm.pd.512(<8 x double> %26, <8 x double> %0, <8 x i64> <i64 167482009228346368, i64 167482009228346368, i64 167482009228346368, i64 167482009228346368, i64 167482009228346368, i64 167482009228346368, i64 167482009228346368, i64 167482009228346368>, i32 0, i8 -1, i32 4)
ret <8 x double> %27
"""), Vec{8,Float64}, Tuple{Vec{8,Float64}}, v)
end
else
@inline function log2(v::Vec{8,Float64})
Base.llvmcall(("""
declare <8 x double> @llvm.fma.v8f64(<8 x double>, <8 x double>, <8 x double>)
declare <8 x double> @llvm.x86.avx512.mask.fixupimm.pd.512(<8 x double>, <8 x double>, <8 x i64>, i32, i8, i32)
declare <8 x double> @llvm.x86.avx512.mask.getexp.pd.512(<8 x double>, <8 x double>, i8, i32)
declare <8 x double> @llvm.x86.avx512.mask.getmant.pd.512(<8 x double>, i32, <8 x double>, i8, i32)
""","""
%2 = fmul <8 x double> %0, <double 0x3FF5555555555555, double 0x3FF5555555555555, double 0x3FF5555555555555, double 0x3FF5555555555555, double 0x3FF5555555555555, double 0x3FF5555555555555, double 0x3FF5555555555555, double 0x3FF5555555555555>
%3 = tail call <8 x double> @llvm.x86.avx512.mask.getexp.pd.512(<8 x double> %2, <8 x double> zeroinitializer, i8 -1, i32 4) #13
%4 = fcmp oeq <8 x double> %3, <double 0x7FF0000000000000, double 0x7FF0000000000000, double 0x7FF0000000000000, double 0x7FF0000000000000, double 0x7FF0000000000000, double 0x7FF0000000000000, double 0x7FF0000000000000, double 0x7FF0000000000000>
%5 = select <8 x i1> %4, <8 x double> <double 1.024000e+03, double 1.024000e+03, double 1.024000e+03, double 1.024000e+03, double 1.024000e+03, double 1.024000e+03, double 1.024000e+03, double 1.024000e+03>, <8 x double> %3
%6 = tail call <8 x double> @llvm.x86.avx512.mask.getmant.pd.512(<8 x double> %0, i32 11, <8 x double> zeroinitializer, i8 -1, i32 4) #13
%7 = fadd <8 x double> %6, <double -1.000000e+00, double -1.000000e+00, double -1.000000e+00, double -1.000000e+00, double -1.000000e+00, double -1.000000e+00, double -1.000000e+00, double -1.000000e+00>
%8 = fadd <8 x double> %6, <double 1.000000e+00, double 1.000000e+00, double 1.000000e+00, double 1.000000e+00, double 1.000000e+00, double 1.000000e+00, double 1.000000e+00, double 1.000000e+00>
%9 = fdiv <8 x double> %7, %8
%10 = fmul <8 x double> %9, %9
%11 = tail call <8 x double> @llvm.fma.v8f64(<8 x double> %10, <8 x double> <double 0x3FCC501739F17BA9, double 0x3FCC501739F17BA9, double 0x3FCC501739F17BA9, double 0x3FCC501739F17BA9, double 0x3FCC501739F17BA9, double 0x3FCC501739F17BA9, double 0x3FCC501739F17BA9, double 0x3FCC501739F17BA9>, <8 x double> <double 0x3FCC2B7A962850E9, double 0x3FCC2B7A962850E9, double 0x3FCC2B7A962850E9, double 0x3FCC2B7A962850E9, double 0x3FCC2B7A962850E9, double 0x3FCC2B7A962850E9, double 0x3FCC2B7A962850E9, double 0x3FCC2B7A962850E9>) #13
%12 = tail call <8 x double> @llvm.fma.v8f64(<8 x double> %11, <8 x double> %10, <8 x double> <double 0x3FD0CAAEEB877481, double 0x3FD0CAAEEB877481, double 0x3FD0CAAEEB877481, double 0x3FD0CAAEEB877481, double 0x3FD0CAAEEB877481, double 0x3FD0CAAEEB877481, double 0x3FD0CAAEEB877481, double 0x3FD0CAAEEB877481>) #13
%13 = tail call <8 x double> @llvm.fma.v8f64(<8 x double> %12, <8 x double> %10, <8 x double> <double 0x3FD484AC6A7CB2DD, double 0x3FD484AC6A7CB2DD, double 0x3FD484AC6A7CB2DD, double 0x3FD484AC6A7CB2DD, double 0x3FD484AC6A7CB2DD, double 0x3FD484AC6A7CB2DD, double 0x3FD484AC6A7CB2DD, double 0x3FD484AC6A7CB2DD>) #13
%14 = tail call <8 x double> @llvm.fma.v8f64(<8 x double> %13, <8 x double> %10, <8 x double> <double 0x3FDA617636C2C254, double 0x3FDA617636C2C254, double 0x3FDA617636C2C254, double 0x3FDA617636C2C254, double 0x3FDA617636C2C254, double 0x3FDA617636C2C254, double 0x3FDA617636C2C254, double 0x3FDA617636C2C254>) #13
%15 = tail call <8 x double> @llvm.fma.v8f64(<8 x double> %14, <8 x double> %10, <8 x double> <double 0x3FE2776C50E7EDE9, double 0x3FE2776C50E7EDE9, double 0x3FE2776C50E7EDE9, double 0x3FE2776C50E7EDE9, double 0x3FE2776C50E7EDE9, double 0x3FE2776C50E7EDE9, double 0x3FE2776C50E7EDE9, double 0x3FE2776C50E7EDE9>) #13
%16 = tail call <8 x double> @llvm.fma.v8f64(<8 x double> %15, <8 x double> %10, <8 x double> <double 0x3FEEC709DC3A07B2, double 0x3FEEC709DC3A07B2, double 0x3FEEC709DC3A07B2, double 0x3FEEC709DC3A07B2, double 0x3FEEC709DC3A07B2, double 0x3FEEC709DC3A07B2, double 0x3FEEC709DC3A07B2, double 0x3FEEC709DC3A07B2>) #13
%17 = fmul <8 x double> %9, <double 0x40071547652B82FE, double 0x40071547652B82FE, double 0x40071547652B82FE, double 0x40071547652B82FE, double 0x40071547652B82FE, double 0x40071547652B82FE, double 0x40071547652B82FE, double 0x40071547652B82FE>
%18 = fsub fast <8 x double> <double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00, double -0.000000e+00>, %17
%19 = tail call <8 x double> @llvm.fma.v8f64(<8 x double> %9, <8 x double> <double 0x40071547652B82FE, double 0x40071547652B82FE, double 0x40071547652B82FE, double 0x40071547652B82FE, double 0x40071547652B82FE, double 0x40071547652B82FE, double 0x40071547652B82FE, double 0x40071547652B82FE>, <8 x double> %18) #13
%20 = fadd <8 x double> %5, %17
%21 = fsub <8 x double> %5, %20
%22 = fadd <8 x double> %17, %21
%23 = fadd <8 x double> %19, %22
%24 = fmul <8 x double> %9, %10
%25 = fadd <8 x double> %20, %23
%26 = tail call <8 x double> @llvm.fma.v8f64(<8 x double> %16, <8 x double> %24, <8 x double> %25) #13
%27 = tail call <8 x double> @llvm.x86.avx512.mask.fixupimm.pd.512(<8 x double> %26, <8 x double> %0, <8 x i64> <i64 167482009228346368, i64 167482009228346368, i64 167482009228346368, i64 167482009228346368, i64 167482009228346368, i64 167482009228346368, i64 167482009228346368, i64 167482009228346368>, i32 0, i8 -1, i32 4)
ret <8 x double> %27
"""), Vec{8,Float64}, Tuple{Vec{8,Float64}}, v)
end
end


@inline log(v::SVec{8,Float64}) = SVec(log(extract_data(v)))
@inline Base.log(v::SVec{8,Float64}) = SVec(log(extract_data(v)))
@inline log2(v::SVec{8,Float64}) = SVec(log2(extract_data(v)))
Expand Down

2 comments on commit 304f30f

@chriselrod
Copy link
Member Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@JuliaRegistrator register()

@JuliaRegistrator
Copy link

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Registration pull request created: JuliaRegistries/General/9736

After the above pull request is merged, it is recommended that a tag is created on this repository for the registered package version.

This will be done automatically if Julia TagBot is installed, or can be done manually through the github interface, or via:

git tag -a v0.3.6 -m "<description of version>" 304f30fa32cdea353576b5c68aeccd0c9779dff1
git push origin v0.3.6

Please sign in to comment.