2,652
edits
Eiger: Difference between revisions
→Xeon Phi (Knights Landing, KNL): SNC4
| Line 70: | Line 70: | ||
INTEGRATE: total elapsed wall-clock time 53.1 sec | INTEGRATE: total elapsed wall-clock time 53.1 sec | ||
Back to 16 JOBS and 16 PROCESSORS, but with MCDRAM in ''flat'' mode und <code>numactl --preferred=1 xds_par</code> (thus using all 16GB for arrays, and nothing for cache): | Back to 16 JOBS and 16 PROCESSORS, but with MCDRAM in ''flat'' mode und <code>numactl --preferred=1 xds_par</code> (thus using all 16GB for arrays, and nothing for cache): | ||
INIT.LP: elapsed wall-clock time 29.5 sec | INIT.LP: elapsed wall-clock time 29.5 sec | ||
COLSPOT: elapsed wall-clock time 38.6 sec | COLSPOT: elapsed wall-clock time 38.6 sec | ||
INTEGRATE: total elapsed wall-clock time 53.2 sec | INTEGRATE: total elapsed wall-clock time 53.2 sec | ||
Now setting the KNL to SNC4 mode, and the MCDRAM to cache (using it in flat mode is impractical because the --preferred argument takes only 1 argument, but 4 would have to be specified unless heavy scripting were done): | |||
INIT.LP: elapsed wall-clock time 29.6 sec | |||
COLSPOT.LP: elapsed wall-clock time 37.8 sec | |||
INTEGRATE: total elapsed wall-clock time 50.2 sec | |||
Conclusion: since INIT benefits from more PROCESSORs, one could run XDS twice for fastest turnaround; the first run with JOBS=XYCORR INIT and a high number of processors (99 is maximum). The second run with JOB=COLSPOT IDXREF DEFPIX INTEGRATE CORRECT, and an optimized JOBS/PROCESSORS combination. | Conclusion: since INIT benefits from more PROCESSORs, one could run XDS twice for fastest turnaround; the first run with JOBS=XYCORR INIT and a high number of processors (99 is maximum). The second run with JOB=COLSPOT IDXREF DEFPIX INTEGRATE CORRECT, and an optimized JOBS/PROCESSORS combination. | ||