Re: [locking/rwsem] 10a59003d2: unixbench.score -25.5% regression

[Waiman Long <longman@redhat.com>]

On 11/23/20 10:53 AM, kernel test robot wrote:
>
> Greeting,
>
> FYI, we noticed a -25.5% regression of unixbench.score due to commit:
>
>
> commit: 10a59003d29fbfa855b2ef4f3534fee9bdf4e575 ("[PATCH v2 5/5] locking/rwsem: Remove reader optimistic spinning")
> url: https://github.com/0day-ci/linux/commits/Waiman-Long/locking-rwsem-Rework-reader-optimistic-spinning/20201121-122118
> base: https://git.kernel.org/cgit/linux/kernel/git/tip/tip.git 932f8c64d38bb08f69c8c26a2216ba0c36c6daa8
>
> in testcase: unixbench
> on test machine: 16 threads Intel(R) Xeon(R) E-2278G CPU @ 3.40GHz with 32G memory
> with following parameters:
>
> 	runtime: 300s
> 	nr_task: 30%
> 	test: shell8
> 	cpufreq_governor: performance
> 	ucode: 0xde
>
> test-description: UnixBench is the original BYTE UNIX benchmark suite aims to test performance of Unix-like system.
> test-url: https://github.com/kdlucas/byte-unixbench
>
> In addition to that, the commit also has significant impact on the following tests:
>
> +------------------+---------------------------------------------------------------------------+
> | testcase: change | fio-basic: fio.write_iops -29.9% regression                               |
> | test machine     | 192 threads Intel(R) Xeon(R) CPU @ 2.20GHz with 192G memory               |
> | test parameters  | bs=4k                                                                     |
> |                  | cpufreq_governor=performance                                              |
> |                  | disk=1SSD                                                                 |
> |                  | fs=xfs                                                                    |
> |                  | ioengine=sync                                                             |
> |                  | nr_task=32                                                                |
> |                  | runtime=300s                                                              |
> |                  | rw=randwrite                                                              |
> |                  | test_size=256g                                                            |
> |                  | ucode=0x4003003                                                           |
> +------------------+---------------------------------------------------------------------------+
> | testcase: change | aim7: aim7.jobs-per-min 952.6% improvement                                |
> | test machine     | 96 threads Intel(R) Xeon(R) Platinum 8260L CPU @ 2.40GHz with 128G memory |
> | test parameters  | cpufreq_governor=performance                                              |
> |                  | disk=4BRD_12G                                                             |
> |                  | fs=f2fs                                                                   |
> |                  | load=100                                                                  |
> |                  | md=RAID0                                                                  |
> |                  | test=sync_disk_rw                                                         |
> |                  | ucode=0x4003003                                                           |
> +------------------+---------------------------------------------------------------------------+

A performance drop in some benchmark is expected. However, there are 
others that can show improvement. Will take a look to see if we can 
reduce the performance regression.

Thanks,
Longman