Closed Bug 660470 Opened 13 years ago Closed 13 years ago

Benchmark: After releng upgrades builders' hard drives, is pymake (plus -j4) more of a win?

Categories

(Release Engineering :: General, defect, P3)

Product:
Release Engineering
Component:
General
Platform:
x86
macOS
Type:
defect

Tracking

(Not tracked)

Status:
RESOLVED FIXED

People

(Reporter: justin.lebar+bug, Unassigned)

References

Details

(Whiteboard: [buildfaster:p1]) 

AIUI many of our machines have very slow hard drives, but releng is in the process of fixing this.  The slow drives may be to blame for the statistic I keep hearing, that pymake (plus -j4) isn't a win on Windows.

If we can prove that pymake is a substantial win, I bet bug 593585 would gain some momentum.
Component: Build Config → Release Engineering
Priority: -- → P3
Product: Core → mozilla.org
QA Contact: build-config → release
Version: unspecified → other
just did a quick run on w32-ix-slave34.

Using http://hg.mozilla.org/mozilla-central/rev/d963476cbf67, our debug mozconfig with mk_add_options MOZ_MAKE_FLAGS=-j4, and
time python -OO build/pymake/make.py client.mk build took 26:13.

A noop build right after that took 6:52.

Since august 8th our build machines have range from 10:21 to 2:15:00 with an average of 50:06.

So this looks like a big win to me!
Whiteboard: [buildfaster:p1]
Status: NEW → RESOLVED
Closed: 13 years ago
Resolution: --- → FIXED
Product: mozilla.org → Release Engineering
You need to log in before you can comment on or make changes to this bug.