aboutsummaryrefslogtreecommitdiffhomepage
path: root/blas
diff options
context:
space:
mode:
authorGravatar Rasmus Munk Larsen <rmlarsen@google.com>2019-02-22 13:56:26 -0800
committerGravatar Rasmus Munk Larsen <rmlarsen@google.com>2019-02-22 13:56:26 -0800
commit6560692c670bcf34fc922474bf37f3c18b8768af (patch)
tree551dd7bd544c2db42f34b92bfaf368744bcc525f /blas
parent0b25a5c431f2764cd46a04f07536d60256ecd256 (diff)
Improve EventCount used by the non-blocking threadpool.
The current algorithm requires threads to commit/cancel waiting in order they called Prewait. Spinning caused by that serialization can consume lots of CPU time on some workloads. Restructure the algorithm to not require that serialization and remove spin waits from Commit/CancelWait. Note: this reduces max number of threads from 2^16 to 2^14 to leave more space for ABA counter (which is now 22 bits). Implementation details are explained in comments.
Diffstat (limited to 'blas')
0 files changed, 0 insertions, 0 deletions