ray test seed merging code on NormalStool

test seed merging code on NormalStool

Open sebhtml opened this issue 11 years ago • 13 comments

Oct 01 '13 12:10 sebhtml

/mnt/lustre03/corbeil/corbeil_group/nne-790-ab/projects/ANL

Oct 01 '13 12:10 sebhtml

the new code in code/SpuriousSeedAnnihilator/ (code/SpuriousSeedAnnihilator/SeedMergingWorkflow.cpp ) has still some issues. it won't ship in 2.3.0

Oct 01 '13 12:10 sebhtml

Works with MPI_IO=n on mp2

616d2a26cc1e39f59325a0e632af46262edaa12c

Oct 02 '13 20:10 sebhtml

(without seed merging)

[boisver1@ip03-mp2 ANL]$ qsub NormalStool-9.sh 45367.mp2.m [boisver1@ip03-mp2 ANL]$ pwd /mnt/lustre03/corbeil/corbeil_group/nne-790-ab/projects/ANL

Jan 15 '14 19:01 sebhtml

with -merge-seeds on

[boisver1@ip03-mp2 ANL]$ qsub NormalStool-10.sh 45369.mp2.m

Jan 15 '14 19:01 sebhtml

Problem with Message.cpp 533 with routing

will rerun with core generation

[boisver1@ip03-mp2 ANL]$ qsub NormalStool-11.sh 45534.mp2.m

Jan 15 '14 22:01 sebhtml

with routing:

[boisver1@ip03-mp2 ANL]$ pwd /mnt/lustre03/corbeil/corbeil_group/nne-790-ab/projects/ANL [boisver1@ip03-mp2 ANL]$ qsub NormalStool-12.sh 49061.mp2.m

Jan 20 '14 22:01 sebhtml

new location (lustre03 is read-only and will be decommissioned.)

/mnt/scratch_mp2/corbeil/corbeil_group/nne-790-ab/projects/ANL

Feb 20 '14 19:02 sebhtml

Rank 245 has 173 seeds >= 1000 nt.

but has 3500 seeds >= 1 nt.

a lot of them are just deadends...

Feb 22 '14 02:02 sebhtml

[boisver1@ip03-mp2 ANL]$ qsub NormalStool-14.sh 26991.mp2.m

Feb 22 '14 02:02 sebhtml

-14 hanged at the start, maybe because of hardware, who knows.

Feb 22 '14 12:02 sebhtml

-15 failed because 3 checkpoint files are corrupted on mp2.

[boisver1@ip03-mp2 ANL]$ qsub NormalStool-16.sh
27065.mp2.m

Feb 22 '14 12:02 sebhtml

with raw code (-13): Rank 1 has 3583 seeds to register. with new patch: xxx

Feb 22 '14 12:02 sebhtml