mulimoen
8873f458b4
useless conversion
2021-06-29 18:07:26 +02:00
mulimoen
b5360af6de
Add Evaluator for defining IC/BC in json config
2021-06-29 17:55:19 +02:00
mulimoen
300f6a34e5
Fix interactive eulerplot
2021-06-27 12:34:59 +02:00
mulimoen
091076a9ad
Neon supports fma
2021-04-18 23:20:08 +02:00
mulimoen
2ee55de8c8
Split into components for subcomputation
2021-03-30 22:19:04 +02:00
mulimoen
8051954c29
Update convergence test after trait rework
2021-03-30 19:09:11 +02:00
mulimoen
5f7d38dd55
Introduce Diff type for euler
2021-03-30 18:46:28 +02:00
mulimoen
27abaebc2c
Move vortex stuff to separate file
2021-03-30 18:29:23 +02:00
mulimoen
57c897ec94
Upgrade arrayvec to use const-generics
2021-03-27 11:46:02 +01:00
mulimoen
22a3464326
Add euler iterators
2021-03-26 19:40:04 +01:00
mulimoen
9e2ce3ae24
Add Metrics iterator
2021-03-26 16:27:45 +01:00
mulimoen
75338698a4
Conditionally use fma
2021-03-26 15:57:15 +01:00
mulimoen
7ab6f311c1
Replace hypot with cheaper impl
2021-03-26 15:52:55 +01:00
mulimoen
52c21dbbe9
Add boolean for switching serial/parallel execution
2021-03-26 00:00:42 +01:00
mulimoen
a33e1d37ba
Document what compiler is doing for diffxi
2021-03-25 22:39:00 +01:00
mulimoen
76f5291131
Elide bounds check in diffxi
...
array_windows.skip did not elide bounds checks as it should. If
the slice is instead offset by the skipped amount, we have the
same behavour, but aids the compiler enough.
The two changed lines allows SIMD optimisations, giving an
impressive reduction in instructions by two thirds in the
benchmark.
2021-03-25 17:23:01 +01:00
mulimoen
a81ba0365d
Use core over std in no_std dep
2021-03-25 17:18:25 +01:00
mulimoen
9b5192b357
Add documentation for utilities
2021-03-23 19:37:27 +01:00
mulimoen
ce2e8f17ec
Move FastFloat to separate crate (archival)
2021-03-23 19:36:44 +01:00
mulimoen
4ae5c02bb1
Replace FastFloat with mul_add
2021-03-23 19:21:38 +01:00
mulimoen
df05c06270
Add no_std and forbid(unsafe) in utils
2021-03-23 19:12:09 +01:00
mulimoen
00fcdf1031
Replace StructOpt with argh for comp.time+space
2021-03-22 20:56:16 +01:00
mulimoen
ff9a477b67
Remove non-global Threadpool
2021-03-22 19:28:01 +01:00
mulimoen
7aadda3de9
Move integrate to separate crate
2021-03-22 17:49:35 +01:00
mulimoen
be1330ec02
Add constrmatrix as separate crate
2021-03-22 16:24:32 +01:00
mulimoen
502679c9a1
Move Float to separate crate
2021-03-22 16:17:27 +01:00
mulimoen
be984fbdac
Bump sprs to 0.10
2021-03-18 23:27:03 +01:00
mulimoen
550b43b4cd
Bump ndarray
2021-03-16 19:03:35 +01:00
mulimoen
f098981d3e
Update email
2021-03-16 19:00:24 +01:00
mulimoen
8383517ba3
ensure slice can be cast to Matrix
2021-03-15 20:18:19 +01:00
mulimoen
17ab18e953
zero-pad diffxi kernel
2021-03-15 20:07:41 +01:00
mulimoen
e43e71a4d8
Make flip_XX impl on Matrix
2021-03-15 19:31:41 +01:00
mulimoen
6fc045ae17
Replace transmute with cast
2021-02-12 19:02:13 +01:00
mulimoen
5d59a51bbe
Add option to output information in JSON
2021-02-12 18:28:30 +01:00
mulimoen
743cca00e2
Revert hdf5 patch (TODO: add back info here)
2021-02-12 18:09:15 +01:00
mulimoen
78f0c9ed2c
Update json5 crate to include location errors
2021-02-12 18:08:10 +01:00
mulimoen
d11abff1c3
Add checkbox for gridlines
2021-02-12 17:16:36 +01:00
mulimoen
dc94aa8e6b
Merge branch 'feature/const_matrix'
...
Using constr generics improves performance of diffxi by about
15% without fast-float and to 30% with the fast-float flag
2021-02-10 21:25:08 +01:00
mulimoen
a02c7daafc
remove iterator inhibiting optimisation
2021-02-10 21:17:05 +01:00
mulimoen
02175d1734
use some unsafe...
2021-02-10 19:29:26 +01:00
mulimoen
8a6dc60edf
remove some unsafe from simd
2021-02-10 19:02:48 +01:00
mulimoen
87c055f81e
Add back simd column algo
2021-02-09 21:44:35 +01:00
mulimoen
cf4d8f1e9b
add Zero for constmatrix
2021-02-03 08:41:11 +01:00
mulimoen
7ec426b5a8
add more fast intrinsics
2021-02-03 08:31:33 +01:00
mulimoen
64a4e92dd2
specialise on contigous ny
2021-02-02 00:33:37 +01:00
mulimoen
c709cf465e
remove ndarray transmute
2021-02-02 00:12:03 +01:00
mulimoen
74d99a4a18
try ndarray transmute
2021-02-02 00:12:03 +01:00
mulimoen
b15ea57e6d
inline for perf
2021-02-02 00:12:03 +01:00
mulimoen
299b4f8083
ensure FastFloat flag works
2021-02-02 00:12:03 +01:00
mulimoen
6f7268bf33
use matrices everywhere
2021-02-02 00:12:02 +01:00
mulimoen