update the mg5amcnlo submodule to the latest commit in branch gpucpp by valassi · Pull Request #811 · madgraph5/madgraph4gpu

[gpucpp] go back to codegen logs before regenerating processes - prepare to merg eupstream/master with HIP #801

valassi

committed

Merge remote-tracking branch 'upstream/master' (including HIP changes from PR #801) into gpucpp

valassi

committed

[gpucpp] regenerate all processes - including both #801 and the mg5amcnlo update: no changes except in codegen logs (changes in individual processes have been merged already)

valassi

committed

[gpucpp] rerun 78 tput tests on itscrd90 (with latest gpucpp commit), all ok

valassi

committed

[gpucpp] rerun 18 tmad tests on itscrd90 (with latets gpucpp commit) - almost all ok except for two issues in ggttg and gqttq tests

valassi

committed

[gpucpp] in tmad/madX.sh increase tolerance from 2E-14 to 3E-14 for xsec comparisons (failed for ggttg and gqttq double in PR #811)

valassi

committed

[gpucpp] rerun two tmad tests with increased tolerance, now ok (all 18 tmad tests are now ok on itscrd90)

valassi

committed

[gpucpp] rerun 78 tput tests on LUMI - usual problems in gqttq #806 but otherwise OK

valassi

committed

[gpucpp] rerun 18 tmad tests on LUMI - two test fail for ggttg and gqttq as on itscrd90, will rerun them after increasing the tolerance

valassi

committed

[gpucpp] rerun the two tmad tests on LUMI after increastingtmad tolerance, now "ok" or as good as it gets (gqttq fails with expected #806)

valassi

committed

[gpucpp] COMPLETE GPUCPP go back to itscrd90 logs of tput and tmad tests

valassi

committed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

update the mg5amcnlo submodule to the latest commit in branch gpucpp#811

update the mg5amcnlo submodule to the latest commit in branch gpucpp#811
valassi merged 15 commits into
madgraph5:masterfrom
valassi:gpucpp

Commits on Jan 31, 2024

[gpucpp] update the MG5AMC/mg5amcnlo submodule to the latest commit e428e38c6 in branch gpucpp

[gpucpp] regenerate all processes... there does not seem to be anything major difference in generated c++/fortran code, just version number, setcuts, and a few things around setgrid in myamp.f (man…

[gpucpp] rerun one tput and one tmad test for ggtt, all seems ok - will rerun all tests now (after reverting this commit)

[gpucpp] revert the last commit, prepare to launch all tput and tmad tests

Commits on Feb 1, 2024

[gpucpp] go back to codegen logs before regenerating processes - prepare to merg eupstream/master with HIP #801

Merge remote-tracking branch 'upstream/master' (including HIP changes from PR #801) into gpucpp

[gpucpp] regenerate all processes - including both #801 and the mg5amcnlo update: no changes except in codegen logs (changes in individual processes have been merged already)

[gpucpp] rerun 78 tput tests on itscrd90 (with latest gpucpp commit), all ok

[gpucpp] rerun 18 tmad tests on itscrd90 (with latets gpucpp commit) - almost all ok except for two issues in ggttg and gqttq tests

[gpucpp] in tmad/madX.sh increase tolerance from 2E-14 to 3E-14 for xsec comparisons (failed for ggttg and gqttq double in PR #811)

[gpucpp] rerun two tmad tests with increased tolerance, now ok (all 18 tmad tests are now ok on itscrd90)

[gpucpp] rerun 78 tput tests on LUMI - usual problems in gqttq #806 but otherwise OK

[gpucpp] rerun 18 tmad tests on LUMI - two test fail for ggttg and gqttq as on itscrd90, will rerun them after increasing the tolerance

[gpucpp] rerun the two tmad tests on LUMI after increastingtmad tolerance, now "ok" or as good as it gets (gqttq fails with expected #806)

[gpucpp] COMPLETE GPUCPP go back to itscrd90 logs of tput and tmad tests