madgraph5/madgraph4gpu
GPU development for the Madgraph5_aMC@NLO event generator software package
Fortran
Pinned issues
Issues
- 0
change name of variable warp/wrap ...
#961 opened by oliviermattelaer - 18
- 8
Reduce madevent 'Fortran overhead' by restricting cudacpp helicity calculation (which scales with SIMD) to only 16 events
#958 opened by valassi - 3
Add support for nprocesses>2 (i.e. beyond mirror processes) in cudacpp to speed up directory handling?
#951 opened by valassi - 5
Understand why CMS sees cross section discrepancy between fortran and cuda/cpp for DY+4 jets
#944 opened by valassi - 0
- 6
DY+3 jets cross section decreases by a factor 10 when changing vector size from 16384 to 32?
#959 opened by valassi - 0
instrument python code in gridpacks to provide timing profiles for event generation
#957 opened by valassi - 0
Add events.lhe comparison in tlau tests
#956 opened by valassi - 1
helicity filtering handling
#915 opened by oliviermattelaer - 5
- 22
- 0
Improve usability of cacche by experiment users?
#954 opened by valassi - 2
nprocesses>1 (again): should cudacpp implement this or not? (if needed, implement it)
#940 opened by valassi - 14
Check if the second helicity filtering in fortran for mirror processes can be removed
#941 opened by valassi - 7
generate_events tries to remove GpuAbstraction.h
#947 opened by valassi - 0
- 3
- 3
master_june24: 'error: unhandled SGPR spill to memory' in gg_ttggg builds on LUMI/HIP
#933 opened by valassi - 2
tput_test in the CI does nothing?
#937 opened by valassi - 6
Observations with the CI scrips
#936 opened by roiser - 2
selected color comparison fails for HIP on LUMI
#931 opened by valassi - 0
- 4
(after master_june24) consider releasing the assumption that two adjacent SIMD vectors have the same channelid in mixed mode
#926 opened by valassi - 0
(after master_june24) add sanity check that channelid is the same in a warp also for GPU?
#929 opened by valassi - 6
master_june24: wrong/bugged and unnecessary SIMD handling of numerators and denominators
#894 opened by valassi - 4
master_june24 add runTest (ME comparison) tests for two warps with different channel
#896 opened by valassi - 2
master_june24: fbridge.inc interface was not modified
#916 opened by valassi - 8
master_june24: major bug in memory access for channelIds in C++ calculate_wavefunctions (missing ieventAccessRecordConst)
#899 opened by valassi - 0
- 2
master_june24: CID_ACCESS::kernelAccessConst should be called only once (it is now called once per diagram)
#895 opened by valassi - 3
- 6
master_june24: missing sanity checks that all channelids are the same in a "warp" (or at least inside each SIMD vector)
#898 opened by valassi - 4
master_june24: wrong memory access for channelIds in C++ sigmakin (missing ieventAccessRecordConst)
#911 opened by valassi - 3
master_june24: add LHE color/helicity comparison test for 2 warps with different channel, fix mixed mode (two SIMD vectors must have same channel)
#924 opened by valassi - 1
- 3
segfault in gg_ttgg runTest.exe in my 'june24' WIP
#917 opened by valassi - 1
issues in gg_ttggg runTest.exe in my 'june24' WIP
#920 opened by valassi - 3
Clarified: gg_ttggg has 1240 diagrams in CPPProcess.cc but only 1122 entries in channel2iconfig (and 945 config)
#919 opened by valassi - 1
- 2
issues in tput (and tmad) tests in my 'june24' WIP
#922 opened by valassi - 1
- 1
Additional validation for gg->ttgg
#921 opened by oliviermattelaer - 5
master_june24: fbridgesequence_nomultichannel is not used
#914 opened by valassi - 1
disable squash merging on the repo
#912 opened by valassi - 3
- 4
- 1
- 2
Fix build errors with clang16 (and clang17)
#904 opened by valassi - 0