Vladimir Chalupecky
194082cf5b
blas/testblas: clean up Drotmg test and relax tolerance
2023-07-29 11:54:43 +02:00
Vladimir Chalupecky
fd2bc7697a
blas/testblas: add flag to DrotgTest to skip extreme values
2023-07-27 17:56:55 +02:00
Chris Nobody
45b800a8c6
all: add pkg.go.dev reference README badges
...
As discussed during the review of the immediately preceding changesets.
2023-03-16 06:40:50 +10:30
Chris Nobody
f040f6bd07
all: replace godoc.org README badges with godocs.io ones
...
The links have been redirecting to `pkg.go.dev` for a while now. In the
case of the main README, that made the badge a virtual duplicate of the
"go.dev reference" one.
2023-03-16 06:40:50 +10:30
Dan Kortschak
5f0141ca4c
all: run gofmt and generate all packages
...
Changes made in dsp/fourier/internal/fftpack break the formatting used
there, so these are reverted. There will be complaints in CI.
[git-generate]
gofmt -w .
go generate gonum.org/v1/gonum/blas
go generate gonum.org/v1/gonum/blas/gonum
go generate gonum.org/v1/gonum/unit
go generate gonum.org/v1/gonum/unit/constant
go generate gonum.org/v1/gonum/graph/formats/dot
go generate gonum.org/v1/gonum/graph/formats/rdf
go generate gonum.org/v1/gonum/stat/card
git checkout -- dsp/fourier/internal/fftpack
2022-08-06 07:05:17 +09:30
Dan Kortschak
a2c6f817bf
ci: replace golangci-lint with staticcheck
2022-05-18 21:35:57 +09:30
Dan Kortschak
d8ad7756b6
all: fix spelling and typos
2022-03-14 21:32:06 +10:30
Jinesi Yelizati
befe5e99e1
blas: update README
2021-10-01 05:45:00 +09:30
Vladimir Chalupecky
e25273b06b
blas/gonum: rewrite Drotg and extend its test
2021-05-19 11:41:04 +02:00
Dan Kortschak
62923c3fe1
blas/gonum,internal/asm/f32: harmonise gemv implementations in blas/gonum
2021-04-05 17:33:26 +09:30
Dan Kortschak
72a4f789d0
blas/{gonum,testblas}: increase coverage cases for alpha and beta zeros
2021-03-31 16:26:44 +10:30
Dan Kortschak
da72779e7a
floats/scalar: new package containing non-vector functions from floats
2020-08-07 07:59:02 +09:30
Dan Kortschak
78ef2f096f
blas/{gonum,testblas}: zero out elements of c when beta is zero
2020-03-30 20:22:58 +10:30
Egon Elbre
028ce68c53
blas/gonum: improve dgemmParallel performance
...
name old time/op new time/op delta
DgemmSmSmSm-32 820ns ± 1% 823ns ± 0% ~ (p=0.127 n=5+5)
DgemmMedMedMed-32 137µs ± 1% 139µs ± 0% +1.12% (p=0.008 n=5+5)
DgemmMedLgMed-32 463µs ± 0% 450µs ± 0% -2.88% (p=0.008 n=5+5)
DgemmLgLgLg-32 25.0ms ± 1% 24.9ms ± 0% ~ (p=1.000 n=5+5)
DgemmLgSmLg-32 685µs ± 1% 694µs ± 1% +1.40% (p=0.008 n=5+5)
DgemmLgLgSm-32 808µs ± 1% 761µs ± 0% -5.77% (p=0.008 n=5+5)
DgemmHgHgSm-32 71.7ms ± 0% 68.5ms ± 0% -4.40% (p=0.008 n=5+5)
DgemmMedMedMedTNT-32 345µs ±10% 228µs ± 1% -33.97% (p=0.008 n=5+5)
DgemmMedMedMedNTT-32 142µs ± 0% 149µs ± 1% +5.05% (p=0.008 n=5+5)
DgemmMedMedMedTT-32 584µs ±33% 417µs ± 4% -28.48% (p=0.008 n=5+5)
2020-03-14 20:16:27 +10:30
Dan Kortschak
86e8186781
blas/gonum: make drotmg test pass
...
This change prevents the compiler from emitting FMA instructions during
Drotmg and Drotm which would otherwise cause the arm64 build to fail.
2020-01-27 07:43:03 +10:30
Dan Kortschak
70a775755b
blas/{blas32,blas64,cblas64,cblas128}: check vector length agreement
2019-11-07 20:54:07 +10:30
Dan Kortschak
d70617abe8
blas/blas32: add length field to Vector
2019-11-07 20:54:07 +10:30
Dan Kortschak
40a79689c0
blas/testblas: clean up lint
2019-11-07 19:56:23 +10:30
Dan Kortschak
89d1544a09
blas/{cblas64,cblas128}: add length field to Vector
2019-11-06 23:06:23 +10:30
Dan Kortschak
06383a2742
blas/{blas64,cblas128}: test wrapping of blas interfaces
2019-11-06 23:06:23 +10:30
Dan Kortschak
e2ba7f0950
internal/asm,blas,floats: move level 2 norm to asm
...
This allows sharing of the blas implementation with floats and opens the
possibility of an assembly implementation of this function.
2019-10-25 16:27:36 +10:30
Dan Kortschak
17ea55aedb
blas,lapack: clean up docs and comments
...
Apply (with manual curation after the fact):
* s/^T/U+1d40/g
* s/^H/U+1d34/g
* s/, {2,3}if / $1/g
Some additional manual editing of odd formatting.
2019-09-06 20:02:29 +09:30
Dan Kortschak
035030958a
blas: fix copy paste error
2019-08-19 20:36:41 +09:30
Vladimir Chalupecky
9996f1428e
blas/gonum: call internal/asm functions directly
2019-03-27 06:30:44 +01:00
Vladimir Chalupecky
03b8f50d86
blas/gonum: rename files for consistency
2019-03-27 06:30:44 +01:00
Vladimir Chalupecky
f5dd0ae53c
blas/gonum: generate Complex64 implementation
2019-03-27 06:30:44 +01:00
Vladimir Chalupecky
e59520b803
blas/gonum: slice arguments to exact length
...
Also, avoid unnecessary use of ...To routines.
2019-02-12 23:50:47 +01:00
Vladimir Chalupecky
9ec2aaecbb
blas/blas64: reorder fields of Vector
...
... so that the increment is after the data slice. This matches the
order in which the parameters are passed to the BLAS routines.
2019-01-28 10:32:12 +01:00
Dan Kortschak
d54847ab4d
testblas: use testing Log instead of log.Println
2019-01-19 12:11:24 +10:30
Vladimir Chalupecky
54df3f38fd
blas/testblas: rewrite DtrsmTest to test various matrix strides
2019-01-19 00:17:25 +01:00
Vladimir Chalupecky
c78761093b
blas/gonum: fix matrix strides in Dtrsm and Dsyr2k
2019-01-17 17:29:26 +01:00
Vladimir Chalupecky
3681b8c377
blas/gonum: clean up Dtrsm and use asm functions where possible
2019-01-17 17:29:26 +01:00
Vladimir Chalupecky
61b0fc21a9
blas/gonum: add Ztrsm with test
2019-01-17 17:29:26 +01:00
Vladimir Chalupecky
6b65044604
blas/gonum: clean up Dtrmm and use asm functions where possible
2019-01-17 17:29:26 +01:00
Vladimir Chalupecky
24791c35e9
blas/gonum: add Ztrmm with test
2019-01-17 17:29:26 +01:00
Vladimir Chalupecky
1b07048b32
blas/gonum: add Zhemm with test
2019-01-15 21:56:57 +01:00
Vladimir Chalupecky
ff4d47a8e5
blas/gonum: add Zsymm with test
2019-01-15 21:56:57 +01:00
Vladimir Chalupecky
9a8f312b2c
blas/gonum: clean up Dsymm
...
Remove harmless but unnecessary shadowing of ctmp slice.
2019-01-15 21:56:57 +01:00
Vladimir Chalupecky
cbd96b6d6f
blas/gonum: use switch instead of 'else if'
2019-01-15 21:56:57 +01:00
Vladimir Chalupecky
14c5f9c2d5
blas/gonum: add Zher2k with test
2019-01-12 14:37:56 +01:00
Vladimir Chalupecky
fc3de8ad21
blas/gonum: add Zsyr2k with test
2019-01-12 14:37:56 +01:00
Vladimir Chalupecky
20cf5b6541
blas/gonum: handle special values of beta consistently in Zherk
2019-01-10 12:18:00 +01:00
Vladimir Chalupecky
9653a9356e
blas/gonum: add Zherk with test
2019-01-10 12:18:00 +01:00
Vladimir Chalupecky
20d2a2bc77
blas/gonum: add Zsyrk with test
2019-01-10 12:18:00 +01:00
Vladimir Chalupecky
73c94a2aff
blas/gonum: handle special values of beta consistently in Dsyrk
2019-01-10 12:18:00 +01:00
Vladimir Chalupecky
f41a0d0905
blas/gonum: fix bug in Dsyrk
...
The intention of the affected code was an optimization by avoiding
unnecessary multiplication of C when beta == 1. The bug was that the
condition compared beta to 0 instead of 1 (when uplo == blas.Lower)
which behaved correctly for all values except 0. This went unnoticed
since no test case used beta of 0 and the coverage was complete. This
commit fixes the bug and adds test cases for both Lower and Upper values
of uplo.
2019-01-10 12:18:00 +01:00
Vladimir Chalupecky
bf4bccac52
blas/gonum: add Zgemm with test
2019-01-10 12:18:00 +01:00
Brendan Tracey
572d9101fe
blas64: add length field N to Vector
...
blas64: add length field N to Vector
Alongside, fix the implementation of mat.VecDense and mat.Diagonal, as well as other changes needed to fix this change.
Fixes #736 .
2018-12-10 08:36:04 +00:00
Brendan Tracey
6b862e40c9
blas: Remove variable copying in dgemm ( #734 )
...
* blas: Remove variable copying in dgemm
Fixes #732 .
2018-12-09 16:34:14 +00:00
Vladimir Chalupecky
e2f95e5c31
blas/gonum: add comments to parameter checks
2018-11-15 06:14:16 +01:00