Dan Kortschak
d09a8892ec
internal/asm/f32: make Ger test pass on arm64
2020-01-13 08:52:57 +10:30
Dan Kortschak
7df05ce3f8
internal/asm/f64: zero out ZERO even for zero-length input to L2NormInc
2019-11-26 20:28:53 +10:30
Chad Kunde
96035f78ec
asm/f64: add asm implementation of l2distance function
...
L2 Distance kernel is a slight modification of the L2 Norm kernel,
taking two vectors and calculating the difference before summing squares.
Same overflow protections as the netlib form of the norm calculation.
2019-11-23 16:39:34 +08:00
Chad Kunde
f237e064fc
asm/f64: translate l2norm with increment to asm
...
Netlib algorithm reduces overflow while calculating the l2norm of a
vector.
Translated incremented vector calculation to asm while reducing
branches in NaN and Inf checks. Overflow protection is equivalent to
the Netlib standard implementation.
2019-11-23 16:39:34 +08:00
Chad Kunde
4e1ef9c972
translate netlib l2norm algorithm to asm and remove branches
...
Netlib algorithm reduces overflow while calculating the l2norm of a
vector.
Translated to asm while reducing branches in NaN and Inf checks.
Overflow protection is equivalent to the Netlib standard implementation.
2019-11-15 18:20:45 +08:00
Dan Kortschak
cf4c2f75fb
internal/asm/f64: fix vet issue
2019-11-01 19:10:33 +10:30
Dan Kortschak
7e1144f373
floats: match Norm behaviour in Distance
2019-10-25 16:27:36 +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
30fedc02ce
internal/asm: conform +build comment to convention
2019-04-02 20:57:16 +10:30
Vladimir Chalupecky
9996f1428e
blas/gonum: call internal/asm functions directly
2019-03-27 06:30:44 +01:00
Dan Kortschak
2dbf9a36d5
internal/asm/f64: fix copyright headers
2019-03-10 21:07:44 +10:30
Vladimir Chalupecky
fd50e23eae
asm/f64: fix GemvT by correctly slicing argument to ScalUnitary
2018-12-20 06:20:48 +01:00
Chad Kunde
cc0c958a30
floats,asm/f64: add asm routine for sum
2018-10-29 19:26:01 +10:30
Dan Kortschak
72e79ca985
all: add safe synonym for appengine tag
2018-07-22 17:17:15 +09:30
Dan Kortschak
b9db25a610
all: conform build tag comment to convention
2018-07-06 09:05:06 +09:30
Vladimir Chalupecky
e9e56344e3
all: fix capitalization of Gonum in copyright headers
2018-06-22 17:32:53 +02:00
Chad Kunde
190028012e
remove build flags for unsupported versions
2018-06-17 14:22:00 -07:00
Chad Kunde
f17cc5d4c6
asm/f64: Added GemT (Trans) assembly function.
2018-06-17 14:22:00 -07:00
Chad Kunde
c5ea572ae6
asm/f64: Added gemvN (NoTrans) assembly function.
2018-06-17 14:22:00 -07:00
Chad Kunde
98b54cc4f2
asm/f64: Added stubs, tests, and noasm for gemv functions.
2018-06-17 14:22:00 -07:00
Chad Kunde
20cf9567fb
internal/asm/f32,f64: remove old go version build tags
2018-04-21 19:20:22 -07:00
Chad Kunde
c3e5f1a446
asm/f32: generate ger tests from asm/f64
2018-04-21 07:28:46 +09:30
Chad Kunde
683c8ba07d
asm/f32: Ger asm implementation.
2018-04-21 07:28:46 +09:30
Chad Kunde
62709ddc2a
amd/f32: Stubs and go code for Ger function.
2018-04-21 07:28:46 +09:30
Chad Kunde
78fa1b548e
asm/f64: Added within and epsilon for comparison w/ tolerance.
2018-04-21 07:28:46 +09:30
Chad Kunde
5e413e98fe
asm/f64: Ger asm function added.
2018-04-21 07:28:46 +09:30
Chad Kunde
31786f433f
asm/f64: Adding stubs for Ger asm implementation.
2018-04-21 07:28:46 +09:30
kortschak
7d49308e07
README,internal/asm: remove go1.7 tag references
2017-12-15 09:04:36 +10:30
kortschak
fa0c5aa7bc
all: use golang.org/x/exp/rand instead of math/rand
2017-11-17 11:52:34 +10:30
kortschak
805531d142
all: change capitalization of gonum in license header
2017-11-02 06:54:08 +10:30
kortschak
21d5891471
all: remove go1.7 build constraints
2017-10-27 13:48:47 +10:30
Chad Kunde
e5ba063a8b
Merge pull request #207 from gonum/c128/dot
...
asm/c128: Adding dot product asm functions
2017-09-12 21:59:20 -07:00
Chad Kunde
9523cc3d27
asm/c64: Adding benchmark sink variable.
2017-09-08 18:47:23 -07:00
Chad Kunde
8af929b8ab
asm/c64: Asm dot functions.
2017-09-08 18:47:23 -07:00
Chad Kunde
66b4f42284
asm/c64: Adding tests for dot asm functions.
2017-09-08 18:47:23 -07:00
Chad Kunde
acf06961e6
asm/c64: Adding asm stubs for dot functions.
2017-09-08 18:47:23 -07:00
Chad Kunde
7416b77068
asm/c128: Adding bench sink variable.
2017-09-08 18:46:45 -07:00
Chad Kunde
962589dc69
asm/c128: Asm dot product functions.
2017-09-08 18:46:45 -07:00
Chad Kunde
03cef35c7f
asm/c128: Adding tests and benchmarks for dot functions.
2017-09-08 18:46:45 -07:00
Chad Kunde
6ccc15eb4b
asm/c128: Adding asm stubs for Dot functions.
2017-09-08 18:46:45 -07:00
Chad Kunde
9d7baa818a
asm/f32: Fixes from PR comments.
2017-08-23 11:53:08 -07:00
Chad Kunde
e0a0508f0d
asm/f32: Asm dot functions.
2017-08-22 14:22:38 -07:00
Chad Kunde
7f0e334144
asm/f32: Adding tests for dot asm functions.
2017-08-22 14:20:49 -07:00
Chad Kunde
51fc8a3f42
asm/f32: Adding stubs for dot asm functions.
2017-08-22 14:16:24 -07:00
Chad Kunde
5e3ead2971
asm/c128: Corrected duplicate dscal test case.
2017-08-21 04:30:07 -07:00
Chad Kunde
1d7ccdce34
asm/c128: Comment fixes from PR review.
2017-08-20 12:12:36 -07:00
Chad Kunde
f75ae60e6a
asm/c128: Changed int parameters to uintptr.
2017-08-19 21:50:10 -07:00
Chad Kunde
53ebc944e7
asm/c128: Fixes to dscal algorithm implementation.
2017-08-19 21:50:06 -07:00
Chad Kunde
8b113b6614
asm/c128: scalinc asm coded out and tested.
2017-08-14 21:34:28 +07:00