Commit Graph

Select branches

Hide Pull Requests

cuda

cuda-energy

cuda-test-doubles

newtuples

openacc

cuda-energy-working

Mono Color

• af42b353c4 Use acc::maybeConjugate for cpu code cuda-energy-working cuda-energy Alejandro Gallo 2023-01-26 01:50:53 +0100
• e4f326e394 Fix the reordering kernel in cuda Alejandro Gallo 2023-01-26 01:50:34 +0100
• 93cba3c3ab Implement zeroing of tensors through memcpy and cuMemcpy Alejandro Gallo 2023-01-26 01:49:55 +0100
• 4e2d1143e5 Add getSize static method to calculate the size of sources in SliceUnion Alejandro Gallo 2023-01-25 16:25:09 +0100
• 933d556c84 Fix the reordering kernel in Equations Ania Brown 2023-01-25 13:53:22 +0100
• c7e3fa45bd Add old version of energies and only generate code for doubles Alejandro Gallo 2023-01-25 12:50:19 +0100
• 2b8b3bd421 don't copy self sufficient slices when sources on gpu Anna Brown 2023-01-23 09:03:16 -0800
• 122329eca7 Fix zeroing cuda Alejandro Gallo 2023-01-23 17:01:29 +0100
• 58c0bf078e Zero Tijk correctly in CPU code Ania Brown 2023-01-23 16:58:08 +0100
• 3fe15e5e5c Fix bs and ths error in equations Alejandro Gallo 2023-01-23 16:57:07 +0100
• 0d223e6ed9 Fix vector types for energy in cpu Alejandro Gallo 2023-01-23 14:44:54 +0100
• c8bdc4239f Fix an odd character in the warmup Alejandro Gallo 2023-01-23 14:43:17 +0100
• be96e4bf8c 1.syntax error fix 2.allocate temporary buffers only once per sim Ania Brown 2022-12-16 02:37:57 -0800
• 9003c218a3 don't need to copy to separate mpi_data array on the host when sources are resident on gpu Anna Brown 2022-12-15 09:02:14 -0800
• 4af47a0bb7 Initialize sources on gpus when ATRIP_SOURCES_IN_GPU Ania Brown 2023-01-23 14:21:51 +0100
• 9a5a2487be Add warmup in the SliceUnion Ania Brown 2023-01-23 13:46:20 +0100
• c4ec227185 Clean getEnergyDistinct Alejandro Gallo 2023-01-13 16:59:19 +0100
• 1ceb4cf0d6 Fix maybeConjugate cuda scope Alejandro Gallo 2023-01-13 12:08:54 +0100
• 34a4e79db0 Initial compiling implementation of the energy kernel Alejandro Gallo 2023-01-13 11:33:42 +0100
• 017cf43381 Add preliminary openacc support, atrip bench not linking openacc Alejandro Gallo 2023-01-11 13:06:59 +0100
• 77e1aaabeb Add bureaucracy for openacc in autotools Alejandro Gallo 2023-01-05 00:06:37 +0100
• 249f1c0b51 Add raven modules for cuda Alejandro Gallo 2023-01-04 15:23:36 +0100
• 1d96800d45 Add support for reading tensors from file in atrip bench Alejandro Gallo 2022-12-06 21:20:03 +0100
• 9087e3af19 Update workflows Alejandro Gallo 2022-12-06 20:58:32 +0100
• 418fd9d389 Add simple cuda bench configuration Alejandro Gallo 2022-12-06 20:57:34 +0100
• 895cd02778 Add some documentation about running the benches Alejandro Gallo 2022-12-06 20:38:57 +0100
• 8efa3d911e Add --max-iterations to main bench Alejandro Gallo 2022-12-06 20:38:38 +0100
• 0fa24404e5 Improve the documentation in the readme for benches building Alejandro Gallo 2022-12-06 14:17:53 +0100
• 8f7d05efda Add Building information and building for sources on GPU newtuples Alejandro Gallo 2022-12-06 13:26:44 +0100
• ad542fe856 Add the slicing into the GPU Alejandro Gallo 2022-12-05 21:16:30 +0100
• 658397ebd7 Update in SliceUnion ATRIP_SOURCES_IN_GPU Alejandro Gallo 2022-12-05 17:55:23 +0100
• 26e2f2d109 Add ATRIP_SOURCES_IN_GPU and ATRIP_CUDA_AWARE_MPI defines in configure Alejandro Gallo 2022-12-05 17:49:54 +0100
• 871471aae3 Fix naive-tuples experimentation bench Alejandro Gallo 2022-10-18 16:23:43 +0200
• 65a64f3f8c Test on all pushes Alejandro Gallo 2022-10-08 16:05:40 +0200
• 4f9f09e965 Cleanup flags handling in configure Alejandro Gallo 2022-10-08 16:04:55 +0200
• 6dc943e10a Clean up DatabaseCommunicator Alejandro Gallo 2022-10-08 16:04:37 +0200
• ed347ab0d9 Rename NAIVE_SLOW into ATRIP_NAIVE_SLOW Alejandro Gallo 2022-10-08 16:04:03 +0200
• 8c5c47e208 Add atrip-def m4 macro Alejandro Gallo 2022-10-08 16:03:08 +0200
• 6871372cac Add support for only calculating DGEMM parts ATRIP_ONLY_DGEMM Alejandro Gallo 2022-10-08 16:02:49 +0200
• 452c0fe001 Fix test_main name in automake Alejandro Gallo 2022-10-08 15:59:48 +0200
• b636b89a64 Add configure-benches script Alejandro Gallo 2022-10-08 15:59:25 +0200
• e59d298a01 Rename test_main.cxx to main.cxx Alejandro Gallo 2022-10-08 00:17:57 +0200
• 6143d1ae73 Rename test_main to atrip Gallo Alejandro 2022-10-08 00:13:40 +0200
• 3addd86826 Create a barrier before the iterations start Gallo Alejandro 2022-10-07 22:51:43 +0200
• 3ddd507c17 Fix a bug in the building the abc tuples fakely Alejandro Gallo 2022-10-07 12:23:10 +0200
• ae6736fc21 Implement a faster version of naive computation Alejandro Gallo 2022-10-07 03:19:24 +0200
• d5cfe31b12 Add naive tuples scracth file Alejandro Gallo 2022-10-07 01:11:22 +0200
• ddb4574380 Fix DatabaseCommunicator Alejandro Gallo 2022-10-06 12:27:00 +0200
• 118df09128 Add tentative DatabaseCommunicator Alejandro Gallo 2022-10-06 01:10:06 +0200
• 1e391e3749 Update tuples-distribution script Alejandro Gallo 2022-10-06 01:07:53 +0200
• 7734efeb97 Add tuples distribution bench Alejandro Gallo 2022-10-03 17:13:21 +0200
• fa1a29c583 Create an implementation file of the Tuples Alejandro Gallo 2022-10-03 17:11:49 +0200
• 2cbff5c8c9 Add bench utils Alejandro Gallo 2022-10-03 17:11:33 +0200
• 50896e3cd0 Add vendor infrastructure Alejandro Gallo 2022-10-03 17:11:08 +0200
• 9f11071911 Comment out conjugation, reordering and zeroing in Equations cuda-test-doubles Alejandro Gallo 2022-09-30 14:46:31 +0200
• b11b53aca1 Add warning for not slicing Alejandro Gallo 2022-09-30 14:36:05 +0200
• 5f9725d515 Add forgotten memcpy for SelfSufficient buffers Alejandro Gallo 2022-09-30 14:35:28 +0200
• 1ca31f4929 Fix bench building for cublas-parallel-atrip Alejandro Gallo 2022-09-30 14:34:54 +0200
• 399447131c Exclude clang5 from the nvidia compilation Alejandro Gallo 2022-09-14 14:52:27 +0200
• da2823cf54 Exclude (gcc, cuda) from matrix Alejandro Gallo 2022-09-14 14:43:38 +0200
• 4092f968b7 Tell nvcc to use the correct MPICXX binary Alejandro Gallo 2022-09-14 13:28:15 +0200
• a263c0c933 Fix cuda bench Alejandro Gallo 2022-09-13 15:14:45 +0200
• 2895fd25a3 Disable debug in openblas Alejandro Gallo 2022-09-13 15:00:06 +0200
• 8abc516b1f Use --enable-cuda flag in workflows Alejandro Gallo 2022-09-13 14:51:49 +0200
• 5c41fb65e4 Fix ctf.mk for CUDA Gallo Alejandro 2022-09-13 14:15:39 +0200
• 4e2de62508 Fix memory consumption in test-cublas-parallel-atrip Alejandro Gallo 2022-09-12 20:19:41 +0200
• 81e89274ab Fix yaml workflows syntax Alejandro Gallo 2022-09-12 20:02:51 +0200
• 5ca94e3238 Try cuda in github ci Alejandro Gallo 2022-09-12 20:00:46 +0200
• 6effcbcdc8 Add ci for cuda branch too Alejandro Gallo 2022-09-12 19:58:18 +0200
• 20c29ed815 Add a couple of bench scripts for cuda and cublas testing Gallo Alejandro 2022-09-12 19:53:50 +0200
• a5bb8f3d9c Give a better message when out of memory in allocating SliceUnion Gallo Alejandro 2022-09-12 19:39:38 +0200
• 5678ac0d9c Fix AniaBug#2: Lifecycle of SelfSufficient slices was wrong (comments) Gallo Alejandro 2022-09-12 19:29:35 +0200
• 5483325626 Fix AniaBug #1: cublasCreate after context setting Gallo Alejandro 2022-09-12 19:17:52 +0200
• 23ad87214f Uncomment everything in Equations, (to test see comments) Gallo Alejandro 2022-09-12 19:13:45 +0200
• 3d7702d501 Minimal changes in Equations Gallo Alejandro 2022-09-12 19:10:14 +0200
• c20b9e3bcb Add error checking in Blas.cxx Gallo Alejandro 2022-09-12 19:07:48 +0200
• da704ad820 Add convenience _FORMAT macro Gallo Alejandro 2022-09-12 18:42:36 +0200
• f1b2f37fe2 Check CUresult for mpi_data to device Gallo Alejandro 2022-09-12 18:41:37 +0200
• 1cd7bac187 Add macros to check CUerror and cublasStatus_t Gallo Alejandro 2022-09-12 18:36:01 +0200
• 68892d5dd8 Make bootstrap work from anywhere in the project Gallo Alejandro 2022-09-12 18:35:30 +0200
• 4277c07cc2 Add memory consumption in bench Alejandro Gallo 2022-09-08 15:44:29 +0200
• 0558148937 Fix small syntactic bug Alejandro Gallo 2022-09-08 15:36:10 +0200
• 49ff3b377c Add a chrono for mpi memcpy in cuda Alejandro Gallo 2022-09-08 15:27:51 +0200
• 00a28c990c Indent more conventionally Alejandro Gallo 2022-09-08 15:27:38 +0200
• 2c5a4620ca Lint and tidy up Equations Alejandro Gallo 2022-09-08 13:51:49 +0200
• 368c5619cc Fix the autotools atrip_cublas Alejandro Gallo 2022-09-08 13:50:08 +0200
• 0b14ac7704 Add bootstrap script Alejandro Gallo 2022-09-06 15:26:28 +0200
• 76a785044d Check ngcards agains ranks per node Gallo Alejandro 2022-08-14 15:36:22 +0200
• 7241bbe9fb Implement reordering on the GPU Gallo Alejandro 2022-08-12 18:32:32 +0200
• c2e9e930ba Update main Atrip.cxx using several gpus Gallo Alejandro 2022-08-12 18:30:55 +0200
• b4aef4db9e Fix compilation issues and add KernelSizes Gallo Alejandro 2022-08-12 18:29:21 +0200
• 4651231d3b Update test bench for CUDA Gallo Alejandro 2022-08-12 18:28:20 +0200
• 4101c89907 Improve cuda m4 Gallo Alejandro 2022-08-11 13:55:52 +0200
• f06cd7f562 Fix the request free problem Alejandro Gallo 2022-08-08 18:28:51 +0200
• 8c04280a65 Fix blas Alejandro Gallo 2022-08-08 18:26:52 +0200
• a30c424781 Fix mkl nix Alejandro Gallo 2022-08-08 18:13:32 +0200
• 0b9ec59599 Fix --with-ctf Alejandro Gallo 2022-08-08 18:13:14 +0200
• 6b641e8581 Update openblas with debug symbols Alejandro Gallo 2022-08-08 15:32:16 +0200
• a5b2a74e18 Changes in source files, makes cuda run Gallo Alejandro 2022-08-05 13:42:04 +0200
• e03dd77904 Update nvcc compilation in src Alejandro Gallo 2022-07-29 14:08:11 +0200