summaryrefslogtreecommitdiff
path: root/src/stream_manager.cc
AgeCommit message (Collapse)Author
2025-08-19running formatter (#126)JRPan
* running formatter * consolidate CI runs * use cluster to run formatter * use cluster to run formatter * Add a CI-Success step
2025-03-06sst-integration-stream: make SST integration works with streams (#103)WilliamMTK
* sst-integration-stream: add apis to make SST integration works with stream * Add dev container specs
2025-02-18A bunch of maintenance fixes, the largest of which is getting the PTX ↵Tim Rogers
simulation to work with CUDA 12. (#95) * Fixing the formatter to always use a consistent format and running it on the codebase * Update linux-so-version.txt * Update Makefile * A couple of unnecessary files that are lingering around * Support CUDA 12 * Getting the PTX simulations to work with CUDA 12. The issue is that ptxas added more information (number of barriers and compile time). We have to parse these or lexx/yacc fail. * Update ptxinfo.l debug MACRO was ineffective * Update gpgpusim_check.cmake Update to make the CUDA version print a warning, not an error and updating the print to be more reflective of what the actual problem is.
2025-01-14fix_sst_callbacks: add weak definitions for sst callbacks (#81)WilliamMTK
* fix_sst_callbacks: add weak definitions for sst callbacks * Automated Format --------- Co-authored-by: purdue-jenkins <[email protected]>
2024-12-11Add SST integration into gpgpusim (#44)WilliamMTK
* Add accommodations to run gpgpusim with SST simulation framework through balar * Output setup_environment options when sourcing * Add SST directive check when creating sim thread * Add sst side test for jenkins * sst-integration: update Jenkinsfile with offical sst-elements repo and fix bugs in pipeline script * sst-integration: direct jenkins to rebuild gpgpusim before testing for sst * sst-integration: fix bugs in sst repos config * sst-integration: let Jenkins rebuilds simulator Since the simulator needs to be configured with both normal mode and sst mode, need to rebuild make target to clean prior runs. * sst-integration: Update Jenkinsfile to source env vars when running balar test * sst-integration: refactor code to remove __SST__ flag * sst-integration: fix a bug that init cluster twice for sst * sst-integration: fix a bug of not sending mem packets to SST * sst-integration: remove sst flags from makefiles and setup_env * sst-integration: add comments to SST changes * sst-integration: remove rebuilding simulator in jenkins when testing for SST * sst-integration: revert simulator build script * Add a function to support querying function argument info for SST * sst-integration: add version detection for vanadis binary * Automated Format * add version detection support for gcc 10+ * sst-integration: add cudaMallocHost for SST * sst-integration: fix a compilation bug * sst-integration: add sst balar unittest CI * sst-integration: specify GPU_ARCH for CI test * sst-integration: use bash for github actions * sst-integration: use https links for sst repos * sst-integration: add SST dependencies to CI config * sst-integration: remove sudo * sst-integration: default to yes for apt install * sst-integration: add manual trigger for github action * sst-integration: remove wrong on event * sst-integration: limit CPU usage for compilation * sst-integration: fix wrong path * sst-integration: use personal repo for testing * sst-integration: remove sst-core source in CI to free space * sst-integration: SST_Cycle use print stats with stream id * Automated Format * sst-integration: check for diskspace and try to clean it * sst-integration: move out of docker image * sst-integration: testing for ci path * sst-integration: fix syntax * sst-integration: pass env vars * sst-integration: set env properly * sst-integration: merge LLVM build and test into same job * sst-integration: fix step order * sst-integration: checkout correct branch for env-setup * sst-integration: remove resourcing gpu apps * sst-integration: revert back to docker github action * sst-integration: enable debug trace for sst testing * sst-integration: resourcing gpu app for env vars * sst-integration: use GPUAPPS_ROOT for path for gpu app * sst-integration: use GPUAPPS_ROOT for path for gpu app * sst-integration: enable parallel ci tests and fix not returning with cudaMallocHostSST * sst-integration: using debug flag for CI run * sst-integration: revert debug ci run * sst-integration: CI skips cuda sdk download and launch multiple jobs * sst-integration: reenable parallel tests * sst-integration: reduce concurrent test thread count * sst-integration: skip long test for github runner * sst-integration: try running CI with single core * sst-integrtion: add callback to SST to check thread sync is done in SST_Cycle() * sst-integration: ignore lookup if already found and add callbacks to SST * Automated Format * sst-integration: add support for indirect texture access * Automated Format * sste-integration: fix up for PR * Automated Format --------- Co-authored-by: purdue-jenkins <[email protected]>
2024-08-21Stream stats (#71)JRPan
* Temp commit for Justin and Cassie to sync on code changes for adding per-stream status. * Resolved compile errors. * Removed redundant parameter * Passed cuda_stream_id from accelsim to gpgpusim * Cleaned up unused changes * Changed vector to map, having operator problems. * StreamID defaults to zero * Implemented streams to inc_stats and so on * Fixed TOTAL_ACCESS counts * Implemented GLOBAL_TIMER. * Fixed m_shader->get_kernel SEGFAULT issue in shader.cc. * Use warp_init to track streamID instead of issue_warp * Removed temp debug print * Modified cache_stats to only print data from latest finished stream Added optional arg to cache_stats::print_stats, cache_stats::print_fail_stats and their upstream functions. When streamID is specified, print stats from that stream. When not specified, print all stats. NOTE: current implementation depending on streamid never equals -1 * Removed default arg values of streamID * modified constructor of mem_fetch to pass in streamID * changed get_streamid to get_streamID * Added TODO to gpgpusim_entrypoint.cc and power_stat.cc * Only collect power stats when enabled * print last finished stream in PTX mode using last_streamID * take out additional printf * Add a field to baseline cache to indicate cache level * save gpu object in cache * Print stream ID only once per kernel * rm test print * use -1 for default stream id * cleanup debug prints * remove GLOABL_TIMER * Automated clang-format * Should be correct to print everything in power model * addressing concerns & errors * Automated clang-format * add m_stats_pw in operator+ * Automated Format --------- Co-authored-by: Justin Qiao <[email protected]> Co-authored-by: Justin Qiao <[email protected]> Co-authored-by: Tim Rogers <[email protected]> Co-authored-by: JRPan <[email protected]> Co-authored-by: purdue-jenkins <[email protected]>
2023-06-20fix most c warningsAhmad Alawneh
2020-04-07Merge remote-tracking branch 'localpub/dev' into devtgrogers
2019-09-13Big reformat change using clang-format-6.0Nick
2019-09-13Revert "Add src/ director formatting"Nick
This reverts commit 26ca8de4a6ec9bfe422a14cbe325a5f257df453b.
2019-09-13Add src/ director formattingNick
2019-07-27fix event done implboyealkim
2019-07-26fix indentationboyealkim
2019-07-25fix stream_merge front algoboyealkim
2019-07-10Move g_globals and g_constantsMengchi Zhang
Signed-off-by: Mengchi Zhang <[email protected]>
2019-05-15make gpu_tot_cycle local variable not global variableMahmoud
2018-05-29implements null stream case on cudaStreamWaitEvent and updates warning on ↵Jonathan
null event to reflect api reference
2018-05-29clarifying commentsJonathan
2018-05-29fixes seg fault on concurrentKernels, speeds up concurrent_streams_empty, ↵Jonathan
adds check for null stream case on cudaStreamWaitEvent
2018-05-28better print statements for debugging stream_wait_event opJonathan
2018-05-25seg fault fixJonathan
2018-05-25tentative impl of cudaStreamWaitEventJonathan
2016-07-06ADD: add kernel launching latency from stream to distributorJin Wang
2016-07-06ADD: print kernel parameter size footprint. BUG: concurrent kernels on same ↵Jin Wang
shader, should use hw_cta_id to store shared mem info
2016-07-06ADD: add stats for kernel launching and complete cycleJin Wang
2016-07-05MOD: schedule one child kernel each cycleJin Wang
2016-07-05BUG: kernels should return to stream if not pushed to concurrent kernel poolJin Wang
2016-07-05BUG: multiple child kernels finishJin Wang
2016-07-05MOD: add child kernel stream and scheduling supportJin Wang
2015-06-05Fixing bug with max cycle/instruction/cta + bug with C++ name de-mangling ↵Tayler Hetherington
with spaces (e.g., using templates)
2014-08-14Review: 46001. Fixing a source of non-determinism in GPGPU-Sim (Bug 174). ↵Tayler Hetherington
Not sure if this is the only issue, but this was definitely a source of non-determinism (checked with many of the <rodinia>-FT benchmarks). In gpgpu_sim_thread_concurrent() (gpgpusim_entrypoint.cc), the inner do{...}while(active); only breaks when the gpu is not active. As a result, the gpu is only initialized when the gpu becomes momentarily inactive. If one kernel completes while another kernel is currently pending in the stream's queue, the next kernel will start running immediately causing the GPU to not be reinitialized (which includes resetting per-kernel stats). The fix simply recognizes that a kernel has completed and breaks from the loop prior to starting the next operation. [git-p4: depot-paths = "//depot/gpgpu_sim_research/fermi/distribution/": change = 16463]
2014-08-14 - Adding support for cudaFuncSetCacheConfig API, that allows changing theAhmed El-Shafiey
L1 Cache and Shared Memory configurations across kernels. The support enable the user to specify two more configurations (Preferred L1) or (Preferred Shared Memory) besides the default config. If the cudaFuncSetCacheConfig API is used to set the cache configuration of a specific kernel to either of these configuration (cudaFuncCachePreferShared, cudaFuncCachePreferL1), the simulator will change the cache configuration at kernel launch accordingly, if there is no alternative configurations provided to the simulator it will use the default configurations with a warning message display [git-p4: depot-paths = "//depot/gpgpu_sim_research/fermi/distribution/": change = 15816]
2014-08-14remove a warning for unused variableAhmed El-Shafiey
[git-p4: depot-paths = "//depot/gpgpu_sim_research/fermi/distribution/": change = 14317]
2014-08-14-being more conservative in locking the code of the kernel finish detection, ↵Ahmed El-Shafiey
solves the problem of underminitstic kernel finish detection [git-p4: depot-paths = "//depot/gpgpu_sim_research/fermi/distribution/": change = 14206]
2014-08-14Undoing a change that should't be in this branch.Ayub Gubran
[git-p4: depot-paths = "//depot/gpgpu_sim_research/fermi/distribution/": change = 10697]
2014-08-14My last work added to the branch, wokring on the functional simulator in ↵Ayub Gubran
cudasim. [git-p4: depot-paths = "//depot/gpgpu_sim_research/fermi/distribution/": change = 10695]
2014-08-14Assigned debug level 3 to stream manager output.Inderpreet Singh
[git-p4: depot-paths = "//depot/gpgpu_sim_research/fermi/distribution/": change = 10013]
2014-08-14change copyright notice to include authorsTor Aamodt
[git-p4: depot-paths = "//depot/gpgpu_sim_research/fermi/distribution/": change = 9872]
2011-06-29changing copyright to BSDTor Aamodt
[git-p4: depot-paths = "//depot/gpgpu_sim_research/fermi/distribution/": change = 9687]
2011-01-20Integration change. Bug fixes from AMD-CMU trace gen branch. Wilson Fung
[git-p4: depot-paths = "//depot/gpgpu_sim_research/fermi/distribution/": change = 8389]
2010-12-28- parameter memory and active threads now part of kernel_info_t:Tor Aamodt
Parameters are finalized at kernel launch, which means the contents of parameter memory are initialized. Kernel arguement names have a fixed order, hence same address should be assigned on subsequent kernel launches of same kernel in other streams provided the data size param_t::size of arguments for each kernel launch is identical (an assertion has been added to check this is true). - passing regression [git-p4: depot-paths = "//depot/gpgpu_sim_research/fermi/distribution/": change = 8303]
2010-12-28- Checkpointing new support for concurrent kernel execution (CUDA only, not ↵Tor Aamodt
OpenCL) This changelist adds full support for streams supported by a new class, stream_manager and enables concurrent execution of kernels from different streams. - fast_regression.sh fails for simpleMultiCopy, simpleStreams (other tests passing) ** Known issues ** - Kernel parameter passing is not done correctly for concurrent kernel execution (somehow concurrentKernels is not affected by this): the parameters are stored inside function_info, which is shared among parallel kernel launches so that the values passed into the launch are likely to get overwritten if multiple grids are launched in parallel streams. - Statistics are printed out whenever the simulation thread runs out of cuda commands (doesn't make sense to print out when a kernel ends during concurrent kernel execution). This will probably require further tweaking so as to be more compatible with data collection scripts. [git-p4: depot-paths = "//depot/gpgpu_sim_research/fermi/distribution/": change = 8302]
generated by cgit v1.3 (git 2.53.0) at 2026-07-10 21:22:03 +0000