Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Batched DGEMM (CPU, GPU Hybrid) #3

Open
v1j4y opened this issue Mar 29, 2021 · 0 comments
Open

Batched DGEMM (CPU, GPU Hybrid) #3

v1j4y opened this issue Mar 29, 2021 · 0 comments

Comments

@v1j4y
Copy link
Contributor

v1j4y commented Mar 29, 2021

A batched DGEMM based implementation with tiling of the matrices and using a batched DGEMM interface.
v1j4y added a commit to v1j4y/irpjast that referenced this issue Mar 29, 2021
@v1j4y
[WIP] Added a provider for mod 16 based nelec and nnuc. TREX-CoE#3
2077028
v1j4y added a commit to v1j4y/irpjast that referenced this issue Mar 29, 2021
@v1j4y
[WIP] Added tiled version of three necesary tensors. TREX-CoE#3
fce1929
v1j4y added a commit to v1j4y/irpjast that referenced this issue Mar 29, 2021
@v1j4y
Added tiled DGEMM. TREX-CoE#3
a3379b4
v1j4y added a commit to v1j4y/irpjast that referenced this issue Mar 29, 2021
@v1j4y
Fixed some bugs in index ordering. TREX-CoE#3
5f570ff
v1j4y added a commit to v1j4y/irpjast that referenced this issue Mar 30, 2021
@v1j4y
Fixed bug in tiled_dgemm. TREX-CoE#3
67bbef2
v1j4y pushed a commit to v1j4y/irpjast that referenced this issue Mar 31, 2021
Working on final factors. TREX-CoE#3
d8d13bd
v1j4y pushed a commit to v1j4y/irpjast that referenced this issue Apr 8, 2021
Changed ordering of indices for tiled matrices. TREX-CoE#3
e03ee95
v1j4y pushed a commit to v1j4y/irpjast that referenced this issue Apr 8, 2021
Removed providers for old tmp_c and dtmp_c. TREX-CoE#3
3c2e6ac
v1j4y pushed a commit to v1j4y/irpjast that referenced this issue Apr 8, 2021
Added tiled intermediates in codelets. TREX-CoE#3
4b0011c
v1j4y pushed a commit to v1j4y/irpjast that referenced this issue Apr 8, 2021
Added makefile modifications. TREX-CoE#3
1985c43
v1j4y added a commit to v1j4y/irpjast that referenced this issue Apr 8, 2021
@v1j4y
Fixed bugs in indices. TREX-CoE#3
5e60703
v1j4y added a commit to v1j4y/irpjast that referenced this issue Apr 8, 2021
@v1j4y
Fixed bug in dtmp_c_tiled. TREX-CoE#3
7047392
v1j4y added a commit to v1j4y/irpjast that referenced this issue Apr 8, 2021
@v1j4y
Added new Jastrow providers with tiling and no tiling. TREX-CoE#3
78f6d0f
v1j4y added a commit to v1j4y/irpjast that referenced this issue Apr 8, 2021
@v1j4y
Replaced 16 with tile_size. TREX-CoE#3
c94778c
v1j4y added a commit to v1j4y/irpjast that referenced this issue Apr 8, 2021
@v1j4y
Now hand written tiled dgemm is also fast. TREX-CoE#3
9c384dd
v1j4y added a commit to v1j4y/irpjast that referenced this issue Apr 9, 2021
@v1j4y
Added Anthonys idea. TREX-CoE#3
2e7f44f
v1j4y pushed a commit to v1j4y/irpjast that referenced this issue Apr 21, 2021
Compiling with StarPU functions. StarPU functions and files added. TR…
bdc3b9d 
…EX-CoE#3
v1j4y pushed a commit to v1j4y/irpjast that referenced this issue Apr 21, 2021
Now calling StarPU dgemm works. TREX-CoE#3
48b2f55
v1j4y pushed a commit to v1j4y/irpjast that referenced this issue Apr 21, 2021
Added starpu dgemm also for dtmp_c. TREX-CoE#3
cbf86b9
v1j4y pushed a commit to v1j4y/irpjast that referenced this issue Apr 22, 2021
Fixed reversed loop in xgemm matrix initialization. TREX-CoE#3
32ae5f1
v1j4y pushed a commit to v1j4y/irpjast that referenced this issue May 24, 2021
Bug in starpu dgemm. TREX-CoE#3
b3bde6f
v1j4y pushed a commit to v1j4y/irpjast that referenced this issue May 24, 2021
Fixed Makefile. TREX-CoE#3
daea16c
v1j4y pushed a commit to v1j4y/irpjast that referenced this issue May 24, 2021
Added tiling and interface modifications. TREX-CoE#3
98117a2
v1j4y pushed a commit to v1j4y/irpjast that referenced this issue May 24, 2021
Added remaining files. TREX-CoE#3
b30bd3a
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees
No one assigned
Labels
None yet
Projects
None yet
Milestone
No milestone
Development

No branches or pull requests
1 participant
@v1j4y