Commits · 5600b6b6d67795916af2b30e8703d9e58f05568f · pycodegen / pystencils

Aug 12, 2024
- Add SVE nontemporal stores and scatters, including masked variants · e77df63f
  Michael Kuron authored 7 months ago
```
The added tests revealed a number of bugs in maskStore/maskStoreS of other instructions sets, which are also fixed.
```
  e77df63f
May 16, 2024
- Support ARM64 Streaming SVE · b699c883
  Michael Kuron authored 10 months ago
```
This is part of the Scalable Matrix Extensions (SME)
```
  b699c883
Apr 22, 2024
- RISC-V cacheline zero · 8597d398
  Michael Kuron authored 1 year ago
  
  8597d398
Jan 16, 2024
- Refactor source tree layout · bcc8d818
  Frederik Hennig authored 1 year ago and Markus Holzer committed 1 year ago
  
  bcc8d818
Sep 13, 2023
- Reveal base pointer spec · 7b08c076
  Markus Holzer authored 1 year ago and Frederik Hennig committed 1 year ago
  
  7b08c076
Sep 07, 2023
- Fix integration pipeline · ae67f2e6
  Daniel Bauer authored 1 year ago and Markus Holzer committed 1 year ago
  
  ae67f2e6
Aug 23, 2023
- Vectorize all scalar symbols in vector expressions · b65d1cc1
  Daniel Bauer authored 1 year ago and Markus Holzer committed 1 year ago
  
  b65d1cc1
Oct 24, 2022
- Improve Vectorisation · 88b9fc0b
  Markus Holzer authored 2 years ago and Christoph Alt committed 2 years ago
  
  88b9fc0b
Jul 02, 2022
- Fix nontemporal stores on non-x86 for fields with variable size · a4d16275
  Michael Kuron authored 2 years ago
  
  a4d16275
May 23, 2022
- Revision2 · 1ff42b84
  Markus Holzer authored 2 years ago and Jan Hönig committed 2 years ago
  
  1ff42b84
Feb 10, 2022
- Fixed second stage pipeline · 4b2bd4d3
  Markus Holzer authored 3 years ago
  
  4b2bd4d3
- Fix Neon vector instruction set · 136f7c95
  Markus Holzer authored 3 years ago
  
  136f7c95
- Repaired vector test cases · 3c8cf943
  Jan Hönig authored 3 years ago and Markus Holzer committed 3 years ago
  
  3c8cf943
- Fixing vectorisation · 939241f2
  Markus Holzer authored 3 years ago and Markus Holzer committed 3 years ago
  
  939241f2
- First example working · 2998e4a4
  Markus Holzer authored 3 years ago
  
  2998e4a4
Nov 02, 2021
- Small clean up · bb3febb1
  Markus Holzer authored 3 years ago and Helen Schottenhamml committed 3 years ago
  
  bb3febb1
- Vectorisation bug · 8e9bded0
  Markus Holzer authored 3 years ago and Helen Schottenhamml committed 3 years ago
  
  8e9bded0
Oct 31, 2021
- Fix deepcopy issue with Sympy 1.9 · c9e59035
  Michael Kuron authored 3 years ago
  
  c9e59035
Sep 10, 2021
- `create_kernel` API Update · 2a6d08ec
  Jan Hönig authored 3 years ago and Markus Holzer committed 3 years ago
  
  2a6d08ec
May 21, 2021
- Sizeless vectorization · cdf73d8f
  Michael Kuron authored 3 years ago and Jan Hönig committed 3 years ago
  
  cdf73d8f
Apr 20, 2021
- Vectorization tests: run with all available instruction sets, add test for maskStore · 686a3ad8
  Michael Kuron authored 3 years ago
  
  686a3ad8
- SVE vectorization · 7238f9bf
  Michael Kuron authored 3 years ago and Markus Holzer committed 3 years ago
  
  7238f9bf
Apr 03, 2021
- logic fix · 1d39f5f5
  Michael Kuron authored 3 years ago
  
  1d39f5f5
Apr 02, 2021
- remove stream from instruction sets that don't have it · 8f92d147
  Michael Kuron authored 3 years ago
  
  8f92d147
Apr 01, 2021
- Allow fields to be aligned to cacheline size · bcd2d628
  Michael Kuron authored 3 years ago
  
  bcd2d628
Mar 31, 2021
- cache line zero: require alignment to cache line size when OpenMP enabled · a100b5b0
  Michael Kuron authored 3 years ago
```
otherwise we can delete stuff from the next outer loop, which in 2D would be the one OpenMP-parallelized
```
  a100b5b0
- nontemporal cleanup · 31be359b
  Michael Kuron authored 3 years ago
  
  31be359b
- nontemporal stores: flush cacheline if available · 7e6d2a21
  Michael Kuron authored 3 years ago
  
  7e6d2a21
- use cache line zero on architectures without nontemporal store · 9c784763
  Michael Kuron authored 3 years ago
  
  9c784763
Mar 29, 2021
- add fence after non-temporal stores · b1750b81
  Michael Kuron authored 3 years ago
```
Fixes #25
```
  b1750b81
- Vectorization improvements · 8440bc6e
  Michael Kuron authored 3 years ago and Markus Holzer committed 3 years ago
  
  8440bc6e
Feb 22, 2021
- Improve ARM64 support · 75d4bdf0
  Michael Kuron authored 4 years ago and Markus Holzer committed 4 years ago
  
  75d4bdf0
Feb 18, 2021
- vectorization: improve treatment of unary minus · 39c214af
  Michael Kuron authored 4 years ago and Markus Holzer committed 4 years ago
  
  39c214af
Jan 11, 2021
- Extended Test suit · 1c2653a3
  Markus Holzer authored 4 years ago and Michael Kuron committed 4 years ago
  
  1c2653a3
Dec 08, 2020
- FIX Quicktests · 391da2e7
  Markus Holzer authored 4 years ago and Stephan Seitz committed 4 years ago
  
  391da2e7
Dec 06, 2020
- FIX pystencils install bug · 50d53928
  Markus Holzer authored 4 years ago and Stephan Seitz committed 4 years ago
  
  50d53928
Aug 10, 2020
- Added test cases to vectorisation · cd3a2f3e
  Markus Holzer authored 4 years ago
  
  cd3a2f3e
Oct 01, 2019
- Autoformat in test_vectorization.py · 9d258f76
  Stephan Seitz authored 5 years ago
  
  9d258f76
Jul 11, 2019
- Import sorting using isort · 6373c03a
  Martin Bauer authored 5 years ago
  
  6373c03a
Jun 12, 2019

Bugfix - Vectorization of in-place LB update wrong · 754c7767

Martin Bauer authored 5 years ago

Block.subs method tried to be too smart:
a = field[..]
b = a + b

was "simplified" incorrectly to
b = field[...] + b

754c7767

Admin message