Mailinglist Archive: opensuse-bugs (6499 mails)

< Previous Next >
[Bug 1134353] adopting BFQ to control I/O
  • From: bugzilla_noreply@xxxxxxxxxx
  • Date: Tue, 16 Jul 2019 10:32:35 +0000
  • Message-id: <bug-1134353-21960-Zc0awSKcn8@http.bugzilla.suse.com/>
http://bugzilla.suse.com/show_bug.cgi?id=1134353
http://bugzilla.suse.com/show_bug.cgi?id=1134353#c26

--- Comment #26 from Paolo Valente <paolo.valente@xxxxxxxxxx> ---
The anomalies reported by Andreas do not occur in any of my test systems. So I
made a script to repeat one of the offending workloads and take a trace during
the test, and Andreas kindly accepted to execute this script on the system
suffering from these anomalies. He expects to have the system available in the
next weeks.

In the meantime, I have finally run benchmarks on a very fast
NVMe SSD: Samsung SSD 970 PRO. Results, which you find here
https://algo.ing.unimo.it/people/paolo/disk_sched/results.php
are rather good.

Throughput with BFQ is about the same as with the other I/O schedulers
(it couldn't be higher, because this kind of drives just wants the
scheduler to stay as aside as possible, when it comes to throughput).
But in the presence of writes as background workload, start-up times
with BFQ are at least 16 times as low as with the other I/O schedulers.
In absolute terms, gnome-terminal starts in ~1.8 seconds with BFQ, while
it takes at least 28.7 (!) seconds with the other I/O schedulers. As
for the video-playing benchmark, only with BFQ no frame gets lost.

--
You are receiving this mail because:
You are on the CC list for the bug.
< Previous Next >