Messages in this thread | | | Subject | Re: [pipe] 3c0edea9b2: lmbench3.PIPE.bandwidth.MB/sec -17.0% regression | From | kernel test robot <> | Date | Wed, 11 Dec 2019 22:52:19 +0800 |
| |
On 12/11/2019 1:52 AM, Linus Torvalds wrote: > On Tue, Dec 10, 2019 at 6:15 AM kernel test robot <rong.a.chen@intel.com> wrote: >> Hi Linus, >> >> Sorry for the inconvenience, indeed, the regression has been fixed. > Woohoo.. And more than fixed, it looks like. Not that I looked at > historical data, I'm not sure how much this number normally > fluctuates. I'm assuming the stddev is just a per-boot "do it a few > times" rather than any long-term thing? > > Linus
Hi Linus,
Yes, the stddev is not very precise, we usually boot more than three times to get the value.
Best Regards, Rong Chen
| |