Commits · a2bd28d30f437535f5d74362b4bf4d03d9ed1f91 · elpa / elpa

10 Jan, 2018 1 commit
- Fix merge error in AVX512 kernel · 354ae8ec
  Andreas Marek authored Jan 10, 2018
  
  354ae8ec
02 Jan, 2018 1 commit
- passing useGPU to tridiagonal solver · 6febaf23
  Pavel Kus authored Nov 27, 2017
  
  6febaf23
14 Dec, 2017 3 commits
- Initialize j in compute_hh_trafo to avoid (wrong) compiler warning · 50a8623c
  Andreas Marek authored Dec 14, 2017
  
  50a8623c
- Explicitly do the conversion of data types where needed · 7a7d45e1
  Andreas Marek authored Dec 14, 2017
  
  7a7d45e1
- Replace tabs with spaces · 8c10c2dc
  Andreas Marek authored Dec 14, 2017
  
  8c10c2dc
27 Nov, 2017 1 commit
- Some changes for K-computer · 4c7521e0
  Andreas Marek authored Nov 27, 2017
  
  4c7521e0
23 Nov, 2017 3 commits
- Double precision real block6 kernel for Sparc64 · c38695bd
  Andreas Marek authored Nov 23, 2017
  
  c38695bd
- Double precision real block4 kernel for Sparc64 · 46ed16fd
  Andreas Marek authored Nov 23, 2017
  
  46ed16fd
- Make double precision real Power8 kernels available · c111ef44
  Andreas Marek authored Nov 23, 2017
  
  c111ef44
20 Nov, 2017 9 commits
- Fix real block6 kernel on power8 · 44098022
  Andreas Marek authored Nov 20, 2017
  
  44098022
- Fix real block6 kernel on power8 · 26401f3b
  Andreas Marek authored Nov 20, 2017
  
  26401f3b
- Fix error in Power8 block2 and 4 kernel · e13b6e7f
  Andreas Marek authored Nov 20, 2017
  
  e13b6e7f
- Fix sign error in Power8 block2 kernel · db81a315
  Andreas Marek authored Nov 20, 2017
  
  db81a315
- Fix real block4 kernel on power8 · c3c2ab26
  Andreas Marek authored Nov 20, 2017
  
  c3c2ab26
- Correct errors in VSX block2 kernel · 567fc568
  Andreas Marek authored Nov 20, 2017
  
  567fc568
- Real block4 single/double-precision kernel for Power8 · afa7e5e3
  Andreas Marek authored Nov 20, 2017
  
  afa7e5e3
- Fix error in compute_hh_trafo · 3bb78876
  Andreas Marek authored Nov 20, 2017
  
  3bb78876
- Different name for SPARC64 routines · a4fc25e4
  Andreas Marek authored Nov 20, 2017
  
  a4fc25e4
19 Nov, 2017 1 commit
- Fix make distcheck · 7389664a
  Andreas Marek authored Nov 19, 2017
  
  7389664a
18 Nov, 2017 2 commits
- Real block2 double-precision kernel for Power8-Computer · fa3e9892
  Andreas Marek authored Nov 18, 2017
  
  fa3e9892
- Real block2 double-precision kernel for K-Computer · ab06e091
  Andreas Marek authored Nov 18, 2017
  
  ab06e091
15 Nov, 2017 1 commit
- Kernels for Skylake · 9516f944
  Andreas Marek authored Nov 15, 2017
  
  9516f944
02 Oct, 2017 3 commits
- finally getting rid of CONST_* macros for numbers · e4e42c43
  Pavel Kus authored Oct 02, 2017
  
  e4e42c43
- further real/complex unifications · 58c393cd
  Pavel Kus authored Oct 02, 2017
```
in elpa2_bandred_template.F90
```
  58c393cd
- real/complex unification · e62e2dd8
  Pavel Kus authored Oct 02, 2017
  
  e62e2dd8
13 Sep, 2017 2 commits
- real/complex unification of declarations · ec8bc696
  Pavel Kus authored Sep 12, 2017
```
in elpa2_bandred_template
```
  ec8bc696
- polished elpa2_compute_real_template · 8567bbe5
  Pavel Kus authored Sep 12, 2017
  
  8567bbe5
11 Sep, 2017 2 commits
- Remove print of blocking factor · a490a3f4
  Andreas Marek authored Sep 11, 2017
  
  a490a3f4
- Remove some unused variables · cce423fc
  Andreas Marek authored Sep 09, 2017
  
  cce423fc
03 Sep, 2017 11 commits
- Work on issue #47 · 2f1cd089
  Andreas Marek authored Aug 31, 2017
  
  2f1cd089
- Remove complex GPU call from OpenMP region · fa6778b7
  Andreas Marek authored Aug 31, 2017
```
This closes issue #51.
```
  fa6778b7
- Remove complex GPU call from OpenMP region · 8638a970
  Andreas Marek authored Aug 31, 2017
```
This closes issue #51.
```
  8638a970
- bit of cleanup · 51ad5b39
  Andreas Marek authored Aug 30, 2017
  
  51ad5b39
- bit of cleanup · 59df91ce
  Andreas Marek authored Aug 30, 2017
  
  59df91ce
- further polishing elpa2_trans_ev_tridi_to_band_template · 5f26aa45
  Pavel Kus authored Aug 29, 2017
  
  5f26aa45
- removed piece of code same for real and complex · c20ab6b3
  Pavel Kus authored Aug 29, 2017
  
  c20ab6b3
- real/complex unification in MPI calls · 1de3ba91
  Pavel Kus authored Aug 29, 2017
```
in elpa2_trans_ev_tridi_to_band_template
```
  1de3ba91
- real/complex unifications in initializations · 510d7933
  Pavel Kus authored Aug 29, 2017
```
in elpa2_trans_ev_tridi_to_band_template
```
  510d7933
- real/complex unification in declarations · e1fae9bc
  Pavel Kus authored Aug 29, 2017
```
in elpa2_trans_ev_tridi_to_band_template
```
  e1fae9bc
- Blocking factor in trans_ev_band_to_full tunable · 7dd73a2c
  Andreas Marek authored Aug 30, 2017
```
The blocking factor can be set now with the new interface
set method. The default is 3

Autotuning still has to be implemented.

This closses issue #44
```
  7dd73a2c