aboutsummaryrefslogtreecommitdiffhomepage
path: root/bench/perf_monitoring
diff options
context:
space:
mode:
authorGravatar Gael Guennebaud <g.gael@free.fr>2015-02-26 16:04:35 +0100
committerGravatar Gael Guennebaud <g.gael@free.fr>2015-02-26 16:04:35 +0100
commita8ad8887bf80869eb216f6f937a559290984a3f7 (patch)
tree513e9c806e6cc644eab866abb020c1090f22d05c /bench/perf_monitoring
parent400becc591359596b1236ee8d9f2c99c12961a42 (diff)
Implement a more generic blocking-size selection algorithm. See explanations inlines.
It performs extremely well on Haswell. The main issue is to reliably and quickly find the actual cache size to be used for our 2nd level of blocking, that is: max(l2,l3/nb_core_sharing_l3)
Diffstat (limited to 'bench/perf_monitoring')
0 files changed, 0 insertions, 0 deletions