[PATCH 1/1] blk-mq: map all HWQ also in hyperthreaded system

Christoph Hellwig hch at lst.de
Wed Jun 28 07:15:24 PDT 2017


On Wed, Jun 28, 2017 at 03:44:40PM +0300, Max Gurtovoy wrote:
> This patch performs sequential mapping between CPUs and queues.
> In case the system has more CPUs than HWQs then there are still
> CPUs to map to HWQs. In hyperthreaded system, map the unmapped CPUs
> and their siblings to the same HWQ.
> This actually fixes a bug that found unmapped HWQs in a system with
> 2 sockets, 18 cores per socket, 2 threads per core (total 72 CPUs)
> running NVMEoF (opens upto maximum of 64 HWQs).
> 
> Performance results running fio (72 jobs, 128 iodepth)
> using null_blk (w/w.o patch):

Can you also tests with Sagi's series to use the proper IRQ
level mapping?




More information about the Linux-nvme mailing list