Hi Kristian, Axel Schwenke wrote:
Kristian Nielsen wrote:
The short story is that for this kind of configuration, it will be necessary to configure @@slave_domain_parallel_threads to be at most the value of @@slave_parallel_threads divided by the number of domain ids used. Eg. with 32 databases/domains, --slave-parallel-threads=64 --slave-domain-parallel-threads=2. Otherwise the domains will starve each other for threads, leading to poor parallelism.
OK, that is something a can easily add to the configuration. I'll schedule two more runs of the benchmark for that.
And here we go. I have run two nearly identical benchmarks, #18 and #19 with the only difference being the setting of slave-domain-parallel-threads which is max(1, slave-parallel-threads/number_domains) for #18 and the default for #19. I have made a comparison spreadsheet for the "safe" slave. It shows that #19 (default settings, left on that sheet) gives better relative and absolute slave performance in almost all cases, compared to #18. BR, XL