Bulk operations performance: juggling external conditions
dl at cs.oswego.edu
Wed Aug 1 05:16:01 PDT 2012
On 08/01/12 06:53, Aleksey Shipilev wrote:
> If anyone interested, I put the new dataset at
> There is only one key difference against *-1 -- it uses fresh FJP
> bootclasspathed. It boosts the CPU utilization dramatically in many
> cases, and some of the quirks are gone.
The remaining utilization issues seem to be in line with
some of my in-progress suggestions for improvements that amount to
(1) reducing the sequential bottlenecks in front of subtask
generation (mainly via an alternative to Spliterator) and
(2) reducing the impact of GC stalls etc on worker availability
(mainly by using CountedCompleters).
Do you see any other systematic issues?
More information about the lambda-dev