Commit Graph

5666 Commits

Author SHA1 Message Date
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
Patrick DeVivo
19ac2540b2 A+C: add Patrick DeVivo 2020-03-13 13:57:01 +10:30
Patrick DeVivo
a553592644 README: add TODOs badge 2020-03-13 13:57:01 +10:30
Vladimir Chalupecky
4363550baf mat: add TriDense.SliceTri 2020-03-12 17:47:46 +01:00
Vladimir Chalupecky
5f268d9394 mat: rename CloneVec to CloneFromVec 2020-03-08 11:00:58 +01:00
Vladimir Chalupecky
8ba336e3c9 mat: use internal setVec function in CloneVec 2020-03-06 22:03:58 +01:00
Vladimir Chalupecky
b12183f05a mat: add MutableVector interface 2020-03-06 22:03:58 +01:00
Dan Kortschak
fe0017ca91 gopher: add svg format version 2020-03-06 06:24:18 +10:30
Sebastien Binet
1dd0408014 num/hyperdual: fix Number example 2020-03-03 11:49:22 +01:00
Sebastien Binet
87edfd4a95 num/dual: fix Number example 2020-03-03 11:49:22 +01:00
Dan Kortschak
b8468606bd mathext/internal/amos: use consistent random source 2020-03-01 18:22:15 +10:30
Dan Kortschak
960a37950c mat: document that Trace will use a Tracer's implementation v0.7.0 2020-02-27 06:58:45 +10:30
Dan Kortschak
32189f28b7 lapack: use binary exponent float constants 2020-02-26 11:41:24 +10:30
Dan Kortschak
ee02969383 appveyor: update test matrix for go1.14 2020-02-26 11:41:24 +10:30
Dan Kortschak
5a328f3baa travis: update test matrix for go1.14 2020-02-26 11:41:24 +10:30
Dan Kortschak
49112b3415 modules: bump Go version 2020-02-26 11:41:24 +10:30
Dan Kortschak
833936c7d4 README: note supported architectures 2020-02-25 15:15:10 +10:30
Vladimir Chalupecky
a37400476f mat: add MulVecTo to BandDense and SymBandDense 2020-02-24 13:06:34 +01:00
Dan Kortschak
a272e7fbf8 math32: add arm64 assembly implementation for Sqrt 2020-02-24 15:38:42 +10:30
Vladimir Chalupecky
de969dc2da testlapack: remove unused equalApprox helper 2020-02-23 09:15:54 +01:00
Vladimir Chalupecky
459cb8bec3 testlapack: remove unused helpers 2020-02-23 09:15:54 +01:00
Vladimir Chalupecky
8928ea27e9 testlapack: move and document residualOrthogonal 2020-02-23 09:15:54 +01:00
Vladimir Chalupecky
25a8a664a4 testlapack: replace isOrthogonal with residualOrthogonal 2020-02-23 09:15:54 +01:00
Vladimir Chalupecky
2786b46b92 testlapack: rework DgerqfTest 2020-02-23 09:15:54 +01:00
Vladimir Chalupecky
580023157c testlapack: rework Dgerq2Test 2020-02-23 09:15:54 +01:00
Vladimir Chalupecky
5666e1d767 testlapack: rework Dgeqr2Test 2020-02-23 09:15:54 +01:00
Vladimir Chalupecky
6f5fb37b77 testlapack: rework Dgeqp3Test 2020-02-23 09:15:54 +01:00
Dan Kortschak
332e2c4547 travis: require arm64 tests to pass
Do not perform coverage analysis on arm64; travis times out.
2020-02-22 19:47:24 +10:30
Vladimir Chalupecky
3973e30147 testlapack: print byte constants as chars 2020-02-22 09:15:10 +01:00
Dan Kortschak
939a2b38a3 optimize/functions: fix ExtendedRosenbrock for BFGS test
With the fused operation, grad diverges, resulting in a non-progression
of the location.
2020-02-22 11:38:52 +10:30
Dan Kortschak
be8b0445de optimize/functions: fix BrownBadlyScaled for BFGS and LBFGS tests
With the fused operation, f3 is calculated to -9e-17 rather than zero,
allowing another iteration, which fails to progress due to underflow.
2020-02-22 09:15:27 +10:30
Vladimir Chalupecky
16e319c960 testlapack: use residualOrthogonal helper in Dorg2lTest and DgesvdTest 2020-02-21 22:42:22 +01:00
Vladimir Chalupecky
b5d31a9a76 lapack: set strictly lower triangle in Dlaset 2020-02-21 22:39:18 +01:00
Dan Kortschak
815f35ac4b optimize: explicitly state interface satisfaction of types 2020-02-22 06:55:55 +10:30
Dan Kortschak
39972c90c7 optimize: relax gradient tolerance for BFGS test 2020-02-21 20:24:15 +10:30
Dan Kortschak
89be31a4e6 optimize: fix doc comment typo 2020-02-21 19:37:42 +10:30
Dan Kortschak
c3867503e7 optimize: relax gradient tolerance for Newton test 2020-02-21 01:24:01 +10:30
Dan Kortschak
fd1732d5e6 dsp/window: make Gaussian a type and add a Transform method 2020-02-20 21:36:17 +10:30
Dan Kortschak
c9a7355ed9 dsp/window: add lookup table window functions 2020-02-20 21:36:17 +10:30
Dan Kortschak
efc4dabf2a dsp/window: use half offset to exclude flanking zeros 2020-02-20 19:30:39 +10:30
Vladimir Chalupecky
43ba13d1a9 optimize: relax gradient tolerance in two tests 2020-02-20 09:33:46 +01:00
Dan Kortschak
d763153d4c fourier/...: move packages to within a dsp hierarchy 2020-02-19 20:03:23 +10:30
Blinnikov Andrei
3573cf5299 A+C: add Andrei Blinnikov 2020-02-19 09:37:28 +10:30
Andrei Blinnikov
54b6ee192e dsp/window: new package for functions to control spectral leakage of FFT 2020-02-19 09:36:40 +10:30
Dan Kortschak
ca302525a3 optimize: add doc comments for Location fields 2020-02-18 21:09:08 +10:30
Vladimir Chalupecky
315a975e1d testlapack: use bad matrices in tests for Dtrexc and Dlaexc 2020-02-17 11:47:50 +01:00
Vladimir Chalupecky
887ddf9cc3 testlapack: rework test for Dlaexc 2020-02-17 11:47:50 +01:00
Vladimir Chalupecky
14cb76f5d8 testlapack: small cleanup in DtrexcTest 2020-02-17 11:47:50 +01:00
Dan Kortschak
8ad895b51b optimize: avoid unnecessary allocations for Gradient and Hessian 2020-02-17 09:22:36 +10:30
Dan Kortschak
7d9e94571f optimize: fix typo 2020-02-16 10:24:52 +10:30