https://github.com/N-BodyShop/changa

sort by:
Revision Author Date Message Commit Date
702eec5 Update testcollapse script to python3 12 June 2024, 00:19:37 UTC
ef28225 TreePiece::massMetalsEnergyCheck(): correctly account SUPERBUBBLE energy. (#173) 12 June 2024, 00:16:54 UTC
94506ea Flexible GPU and CPU Usage for Gravity (#170) * TreePieces and ListCompute can execute code for CPU tree walk when CUDA flag enabled * Move repeated code into separate functions * Minimal working version, gravity is done on CPU when CUDA flag enabled * Main signals whether gravity happens on GPU based on rung population * Missing ifdef around DataManager::unmarkTreePiecesForCleanup * Fix issue with hang during calculateEwald on CPU step * Fix indentation in ListCompute::stateReady * Remove TODO comments near useckloop checks * Remove initAccel() call before first big step * Remove comments about resetting bUseGpu * Dont set lpdata in TreePiece::nextBucket if using CUDA * Misspelled ewaldCPU * Replace bUseGpu with bUseCpu * Restructure ifdef's to get rid of schedCpuWalk * Rearrange arguments in TreePiece::startGravity * Remove unnecessary ifdef's * Forgot to switch startGravity parameters in .ci file * Remove extraneous enableCpu call * Fix startGravity callback for non-concurrent SPH 12 June 2024, 00:14:19 UTC
c32b642 Update test code in supernovaia.cpp and scripts in testonestar. 23 May 2024, 14:26:03 UTC
515770c Github workflows: change source of charm distribution. 23 May 2024, 04:18:59 UTC
b1d1f04 Add comment explaining unusual placement of setDevice 09 May 2024, 16:33:17 UTC
c4ce78e Assign GPUs to PEs based on node ID 09 May 2024, 16:33:17 UTC
6b41f91 Merge pull request #166 from spencerw/nowr Remove workRequest layer from GPU code 22 April 2024, 04:37:34 UTC
0ee3d4d Remove unused function freeDeviceMemory 21 April 2024, 21:14:57 UTC
47b32e1 Fix docstring for TransferParticleVarsBack 21 April 2024, 21:14:42 UTC
9c38c5b Fix out of order comments in adjust() 21 April 2024, 21:00:16 UTC
3bcf1be Whitespace cleanup 21 April 2024, 20:54:48 UTC
9cc0d46 Add docstrings for TreePieceDataTransferBasic and TreePieceDataTransferBasicCleanup 21 April 2024, 20:52:19 UTC
fff7110 CudaDevPtr allocation handled by local stack 21 April 2024, 20:46:39 UTC
6768f6d Remove unused partIndex variable in DataManager::serializeLocal 02 April 2024, 00:40:29 UTC
83bb251 Remove unnecessary cudaStreamSynchronize calls 01 April 2024, 02:09:50 UTC
50c3ac5 TreeWalk state not properly initalized when GPU used for interaction list calculation 29 March 2024, 23:17:43 UTC
611ed4b Merge remote-tracking branch 'upstream/master' into nowr Incorporate gpu_multicopy changes 02 March 2024, 06:33:41 UTC
2ff770b Merge pull request #73 from N-BodyShop/trq/gpu-multicopy Trq/gpu multicopy 01 March 2024, 05:12:25 UTC
0ca8444 Remove more unused code 27 February 2024, 18:18:38 UTC
3a7c12a Add more comments 27 February 2024, 18:01:24 UTC
386cdd8 DataManager::serializeLocal(): HAPI_TRACE ifdefs around trace calls. Also deleted a superfluous CmiWallTimer() call. 27 February 2024, 05:32:03 UTC
189de0f Added a python script for fixing the headers of starlog files output prior to the fix for issue #294. 09 February 2024, 01:45:53 UTC
6bb0247 Updated readstarlog.py to work in py3, fixed a reader bug 09 February 2024, 01:43:04 UTC
f6873fe Cleanup of GPU multicopy code including: DataManager::serializeRemoteChunk(): change name of parameter to avoid confusion with a local variable DataManager: Delete unused localParticles attribute Clean up/add documentation Some whitespace cleanup 08 February 2024, 20:18:05 UTC
16bf606 Misplaced function header from prior upstream merge 07 February 2024, 07:44:09 UTC
b8203bc d_missedParts, d_missedNodes handled entirely within HostCUDA.cu 07 February 2024, 04:14:29 UTC
b60bd10 Remove unused arguments for TransferParticleVarsBack 07 February 2024, 04:05:58 UTC
a8bea29 Remove unused code associated with freeing memory after GPU phase 07 February 2024, 04:01:06 UTC
92044d8 Device bucket and list pointers handled by HostCUDA functions 07 February 2024, 02:46:33 UTC
627f884 Remove unnecessary cudaStreamSynchronize calls 06 February 2024, 22:15:47 UTC
a00bd1a Cleanup formatting and indentation 06 February 2024, 22:12:20 UTC
b112203 Error checking for cudaFree calls outside of HostCUDA 06 February 2024, 21:55:48 UTC
1398d22 Remove workRequest instrumentation 06 February 2024, 21:43:34 UTC
a7594a5 Error checking for CUDA functions 06 February 2024, 21:31:28 UTC
fa65cc5 Rearrange HostCUDA functions for consistency 06 February 2024, 21:15:13 UTC
b0b2e1f Update docstrings for HostCUDA functions 06 February 2024, 21:10:54 UTC
4909005 Missing callback for TreePiecePartListDataTransferRemoteResume 06 February 2024, 20:03:36 UTC
6209ce0 Tracing event for TreePiecePartListDataTransferLocalSmallPhase 06 February 2024, 20:02:20 UTC
081ae76 Consistent placing of HAPI_TRACE functions 06 February 2024, 19:57:43 UTC
260be44 Uncomment and rewrite TreePiecePartListDataTransferLocalSmallPhase 06 February 2024, 19:53:47 UTC
bacfb5f Update verbose HostCUDA function outputs 06 February 2024, 18:54:10 UTC
8366958 Missing callback for TreePieceCellListDataTransferRemote 06 February 2024, 18:10:40 UTC
bacd69a Merge in gpu_multicopy changes 02 February 2024, 00:57:56 UTC
eabae46 Merge pull request #156 from spencerw/scw-gpu-multicopy Finish gpu multicopy 01 February 2024, 21:00:52 UTC
1b40b2e Code and whitespace cleanup 01 February 2024, 19:06:35 UTC
083f8c6 Add EwaldGPU trace event 30 January 2024, 06:11:58 UTC
db9a574 Starlog::logMetaData(): fix small format error for H2FracForm 26 January 2024, 00:50:33 UTC
62a6e2d Add user events 23 January 2024, 18:26:19 UTC
67677c4 Various cleanup 23 January 2024, 01:40:07 UTC
347d56c GPU device pointers for remote functions handled explicitly 19 January 2024, 21:06:22 UTC
eda5382 Updated links in testcosmo README (#162) * Updated links in testcosmo README 19 January 2024, 18:30:50 UTC
496440a Allocate space for EwaldMsg priority field Remove CkMcastBaseMsg inheritance for EwaldMsg 18 January 2024, 23:17:28 UTC
c423b57 Remove gputransfer and bEwaldInited flags EwaldGPU is now passed a message which tells whether it was invoked from serializeLocalTree or EwaldInit 18 January 2024, 23:17:28 UTC
3dd21b6 PressureSmoothParams::fcnSmoth(): Remove fDivv_Corrector because it creates instabilities. This will match newer versions of Gasoline. 14 January 2024, 02:50:31 UTC
19619c9 Device pointer and stream information now stored by TreePieces Remove WorkRequest code from TreePiecePartListDataTransferLocal 10 January 2024, 02:13:10 UTC
ead9ab9 Finish rewrite of TreePieceCellListDataTransferLocal 09 January 2024, 02:27:57 UTC
9d29c12 Change more GravityParticle attributes to cosmoType, and rearrange for better alignment. 04 January 2024, 00:44:07 UTC
325ff37 SFC domain decomposition: delete particle store after messages are sent. (#149) * SFC domain decomposition: delete particle store after messages are sent. Once sendParticlesDuringDD() has sent all the particles, myParticles can be deleted. Doing it here lowers memory use. * Small comment cleanup. 15 December 2023, 15:45:52 UTC
7b38be6 Decouple streams from TreePieces 12 December 2023, 21:53:00 UTC
8596999 Fix segfaults during ewald kernel 07 December 2023, 00:38:14 UTC
bd3bb6a Finish ewald implementation 06 December 2023, 02:34:51 UTC
d3ef55b cudaStreamSynchronize before data transfer callback 04 December 2023, 20:38:26 UTC
3f741ac All kernel launches now managed by a single stream 01 December 2023, 23:50:26 UTC
75dd54a Use double-pointers for device memory 27 November 2023, 23:55:02 UTC
a0a764d Finish Ewald GPU calculation without workRequest 27 November 2023, 05:09:58 UTC
a5831db Rewrite Ewald calculation without WR 23 November 2023, 23:31:15 UTC
2ef0f19 Transfer particle data back to host and deallocate device memory 20 November 2023, 19:43:29 UTC
82551eb Stream handled by DataManager 20 November 2023, 18:23:56 UTC
bce639e Invoke gpuLocalTreeWalk without WorkRequest, device memory pointer managed by DataManager 17 November 2023, 19:12:43 UTC
f44a8d1 Main::outputBinary(): always initialize iTypeWriting This avoids a (benign) access of uninitialized memory. 12 November 2023, 04:25:31 UTC
d59ff52 Initialize accelerations on GPU 07 November 2023, 02:55:00 UTC
9416581 Working version of parallel node+particle gather 03 November 2023, 20:04:29 UTC
7657c5b Fix pointer references, make sure relevant variables are visible to functions 30 October 2023, 21:51:18 UTC
03f3969 Merge remote-tracking branch 'upstream/master' into trq/gpu-multicopy 27 October 2023, 18:19:12 UTC
a05c6e6 Check to be sure we have enough TreePieces for PEs. (#153) * Check to be sure we have enough TreePieces for PEs. 02 September 2023, 21:44:36 UTC
eb80c64 More detailed error messages when opening tipsy array files. 02 September 2023, 21:43:15 UTC
65f6586 Added mf2.py in testcosmo (#145) * replaced mf2.pro with a python script and modified mf.sh 23 August 2023, 17:07:54 UTC
46462c9 Restored FLAG_FLOAT to Makefile.in to fix --enable-float. 18 August 2023, 17:00:26 UTC
c7ba57c TreePiece::updateuDot(): use uPred and uHotPred in PoverRho calculation. 13 July 2023, 18:15:29 UTC
7cb4912 Read nsidm array if restarting from output. (#141) * Read nsidm array if restarting from output. This only is done if SIDM interactions are being counted. 13 June 2023, 04:16:20 UTC
e4235e9 Merge branch 'ppl_master' into public 10 May 2023, 17:40:22 UTC
9623668 Add Elapsed time printouts to aid performance evaluation. 11 March 2023, 18:40:53 UTC
fcc4a32 Fix a bug in updateuDot where PoverRhoGas doesn't include the terms for hot gas, making the ExternalHeating used for two-phase particles erroneously low. Use the mean molecular weight for ionized gas in the calculation of the evaporation coefficient, rather than the value passed as the dMeanMolecularWeight parameter. 11 February 2023, 17:50:37 UTC
bf99d38 xdr: also create link in structures directory 10 January 2023, 20:39:13 UTC
1f14540 Look for xdr in utility. 10 January 2023, 20:39:13 UTC
23ccec7 doSimulation(): call CkWaitQD() before checkpointing. This seems to fix issues with CkIO hangs after restart. 29 December 2022, 16:22:11 UTC
fa47fa8 doSimulation(): do not flush starlog when running benchmarks. 28 December 2022, 00:12:33 UTC
526309a Fixed bug introduced in commit 25c23a5074a3efbfa4aeb15b89430352415988c2 First assign extraData before using star methods. 12 November 2022, 19:52:30 UTC
dd2b6a6 Implement Press (version 3) Ran object. Each Treepiece now has its own random number generator to avoid lock contention in threaded implementations. 12 November 2022, 19:51:07 UTC
98c212c Update version info for v3.5 release. 19 August 2022, 02:43:10 UTC
aaeddd7 HostCuda.cu: add casts to accommodate changed setDeviceToHostCallback() API. Change-Id: I3c6bede3434063e9d8b36225aa6eaf1dfbc8a583 19 August 2022, 00:33:59 UTC
a0af785 Avoid accessing cooling structures in when bGasCooling isn't set. 03 August 2022, 03:50:44 UTC
25c23a5 load_*_star(): initialize iGasOrder to -1. 03 August 2022, 03:50:44 UTC
2959e42 clRateColl_e_H2(): rewrite division into multiply. 26 July 2022, 18:25:47 UTC
e492080 LocalLymanWernerDistributor::work(): cleanly handle nodes with no gas. 26 July 2022, 18:23:47 UTC
b6aa731 Clean up many warnings from gcc v12. 16 July 2022, 05:04:15 UTC
acaea3c Merge branch 'master' of github.com:N-BodyShop/changa 08 July 2022, 04:26:08 UTC
981b2fb Added extended cooltable_xdr to data directory, and documented it. 08 July 2022, 04:23:45 UTC
ccc04b0 Add SMP build for actions. 07 July 2022, 03:16:58 UTC
back to top