On Wednesday, November 14, 2012 10:53:37 AM UTC-8, Alex Keybl wrote:
> Discussions are ongoing as to whether disabling the test is our
> best path forward here, given engineering opposition to disabling
> PGO.
I strongly recommend disabling the test for 32-bit Linux PGO and moving on. Bug
79929
On 11/13/12 5:22 PM, Ehsan Akhgari wrote:
Let me try to be more clear. Assuming that the assertion that the bug
in question is not caused by the PGO compiler miscompiling, turning off
PGO in order to move on would be the wrong thing to do. If it's only
affecting a single test, then that test s
> The majority of our Linux users also do not use any of our builds or even use
> the same compiler as we do. By above logic we should completely stop doing
> Linux builds, but that's clear not a good idea :)
An attempt at getting closer to building what our Linux users use on a daily
basis is
On Tue, Nov 13, 2012 at 05:22:22PM -0800, Ehsan Akhgari wrote:
> On 2012-11-13 5:03 PM, Justin Dolske wrote:
> >On 11/13/12 4:34 PM, Ehsan Akhgari wrote:
> >
> >>But the point here is that unless we know for sure that we're dealing
> >>with a compiler bug, disabling Linux PGO builds may just wallpa
On 11/13/2012 3:53 PM, Alex Keybl wrote:
If the Snappy initiative (or any other group of Mozillians) has short-term
plans to evangelize the perf wins of PGO to Linux distros, I agree that we
should leave PGO builds and testing enabled on Linux and further investigate
the mysterious crash in bu
On 2012-11-13 5:03 PM, Justin Dolske wrote:
On 11/13/12 4:34 PM, Ehsan Akhgari wrote:
But the point here is that unless we know for sure that we're dealing
with a compiler bug, disabling Linux PGO builds may just wallpaper over
the problem.
That's quite possible, and I'm sure there are other
On 11/13/12 4:34 PM, Ehsan Akhgari wrote:
But the point here is that unless we know for sure that we're dealing
with a compiler bug, disabling Linux PGO builds may just wallpaper over
the problem.
That's quite possible, and I'm sure there are other currently-used ways
to exercise the code tha
On 2012-11-13 3:30 PM, Justin Dolske wrote:
On 11/13/12 10:27 AM, Ehsan Akhgari wrote:
Agreed. Actually, reading the bug closely, there's nothing which says
someone has tried to debug this (it's not even clear if it's
reproducible locally), and it seems like the only evidence that we have
abou
If the Snappy initiative (or any other group of Mozillians) has short-term
plans to evangelize the perf wins of PGO to Linux distros, I agree that we
should leave PGO builds and testing enabled on Linux and further investigate
the mysterious crash in bug 799295. Otherwise, our builds/testing sho
On 11/13/12 10:27 AM, Ehsan Akhgari wrote:
Agreed. Actually, reading the bug closely, there's nothing which says
someone has tried to debug this (it's not even clear if it's
reproducible locally), and it seems like the only evidence that we have
about this being PGO related is that it happens o
On 2012-11-13 9:56 AM, Jonathan Kew wrote:
On 12/11/12 15:47, Alex Keybl wrote:
Bug 799295 [1], the driver for this thread, is still an open issue for
FF18 (shipping in 6 weeks). The JS team's recommendation remains to
disable PGO on Linux. According to Taras, the major benefits of PGO on
Linux
On 12/11/12 15:47, Alex Keybl wrote:
Bug 799295 [1], the driver for this thread, is still an open issue for FF18 (shipping in
6 weeks). The JS team's recommendation remains to disable PGO on Linux. According to
Taras, the major benefits of PGO on Linux are for a "starry-eyed-future". Given
al
On 13.11.2012 00:47, Alex Keybl wrote:
>almost nobody uses Mozilla Firefox builds(and no Firefox disributors do pgo)
We should really get the latter fixed. Disabling PGO for our builds
seems like a step in the wrong direction; the numbers collected in this
thread suggest that it's a major los
On Mon, Nov 12, 2012 at 03:47:32PM -0800, Alex Keybl wrote:
> Bug 799295 [1], the driver for this thread, is still an open issue for
> FF18 (shipping in 6 weeks). The JS team's recommendation remains to
> disable PGO on Linux. According to Taras, the major benefits of PGO on
> Linux are for a "star
Bug 799295 [1], the driver for this thread, is still an open issue for FF18
(shipping in 6 weeks). The JS team's recommendation remains to disable PGO on
Linux. According to Taras, the major benefits of PGO on Linux are for a
"starry-eyed-future". Given
> almost nobody uses Mozilla Firefox buil
On 12/10/12 01:01 PM, Chris Peterson wrote:
> On 10/11/12 7:36 PM, Justin Lebar wrote:
>>> 2. Linux is the foundation of B2G and Firefox for Android, where we
>>> *definitely* must deliver
>>> the fastest product we can
>>
>> I totally agree, but it's not clear to me whether continuing to do PGO
>>
On 10/11/2012 1:34 PM, Mike Hommey wrote:
On Thu, Oct 11, 2012 at 02:26:33PM -0400, Rafael Ávila de Espíndola wrote:
On 10/11/2012 02:33 AM, Mike Hommey wrote:
On Wed, Oct 10, 2012 at 05:57:53PM -0400, Justin Lebar wrote:
By "turning off Linux PGO testing", you really mean "stop making and
dis
David Anderson wrote:
> It's still unclear to me what our Linux PGO builds mean. Do
> distributions use them? If not, are they using the exact same
> compiler version and PGO environment data? If not, then they have a
> different configuration that we haven't tested.
I agree that we should make su
On 10/11/12 7:36 PM, Justin Lebar wrote:
2. Linux is the foundation of B2G and Firefox for Android, where we
*definitely* must deliver
the fastest product we can
I totally agree, but it's not clear to me whether continuing to do PGO
on desktop Linux has any effect on our ability to potentially
On Fri, Oct 12, 2012 at 12:28:43AM -0700, David Anderson wrote:
> On Thursday, October 11, 2012 5:49:07 PM UTC-7, Brian Smith wrote:
> > I think it is important to give Linux users the fastest browser we
> > can give them, because:
>
> It's still unclear to me what our Linux PGO builds mean. Do
>
On Thursday, October 11, 2012 5:49:07 PM UTC-7, Brian Smith wrote:
> I think it is important to give Linux users the fastest browser we can give
> them, because:
It's still unclear to me what our Linux PGO builds mean. Do distributions use
them? If not, are they using the exact same compiler ver
> 2. Linux is the foundation of B2G and Firefox for Android, where we
> *definitely* must deliver
> the fastest product we can
I totally agree, but it's not clear to me whether continuing to do PGO
on desktop Linux has any effect on our ability to potentially do PGO
on Android/B2G. If we were to
Zack Weinberg wrote:
> Link-time optimization is described as an experimental new feature in
> the GCC 4.5.0 release notes[1]. The 4.6.0 release notes[2] say that
> it has now "stabilized to the point of being usable", and the 4.7.0
> release notes[3] describe it as still further improved both in
On 10/11/2012 03:49 PM, Ehsan Akhgari wrote:
Don't both of these proposals map to tons of manual work? I'm not
convinced that doing either of those would necessarily be easier than
finding and fixing the PGO bug at hand.
The problem is that fixing this one bug might take only a few days, but
On 2012-10-11 4:34 PM, Mike Hommey wrote:
On Thu, Oct 11, 2012 at 02:26:33PM -0400, Rafael Ávila de Espíndola wrote:
On 10/11/2012 02:33 AM, Mike Hommey wrote:
On Wed, Oct 10, 2012 at 05:57:53PM -0400, Justin Lebar wrote:
By "turning off Linux PGO testing", you really mean "stop making and
dis
On Wednesday, October 10, 2012 11:33:31 PM UTC-7, Mike Hommey wrote:
> That being said, PGO on Linux is between 5 and 20% improvement on our
> various talos tests. That's with the version of gcc we currently use,
> which is 4.5. I'd expect 4.7 to do a better job even, especially if we
> added lto t
These Dromaeo improvements will in part be because IonMonkey is not fully
JIT'ing these paths yet (a regression we're tracking from Firefox 17).
-David
On Thursday, October 11, 2012 11:26:49 AM UTC-7, Rafael Ávila de Espíndola
wrote:
> On 10/11/2012 02:33 AM, Mike Hommey wrote:
>
> > On Wed, O
On Thu, Oct 11, 2012 at 02:26:33PM -0400, Rafael Ávila de Espíndola wrote:
> On 10/11/2012 02:33 AM, Mike Hommey wrote:
> >On Wed, Oct 10, 2012 at 05:57:53PM -0400, Justin Lebar wrote:
> >>By "turning off Linux PGO testing", you really mean "stop making and
> >>distributing Linux PGO builds," right
On 2012-10-11 3:12 PM, Anthony Jones wrote:
On 11/10/12 19:33, Mike Hommey wrote:
That being said, PGO on Linux is between 5 and 20% improvement on our
various talos tests. That's with the version of gcc we currently use,
which is 4.5. I'd expect 4.7 to do a better job even, especially if we
add
On 11/10/12 19:33, Mike Hommey wrote:
> That being said, PGO on Linux is between 5 and 20% improvement on our
> various talos tests. That's with the version of gcc we currently use,
> which is 4.5. I'd expect 4.7 to do a better job even, especially if we
> added lto to the equation (and since we ar
I filed bug 800471 for considering using Clang on Linux.
-Gary
This also suggests another option: using clang on linux too. This would
have the added benefit of using the same compiler for OS X and Linux,
which would remove most of the argument of developers spending time on
linux only issues.
On 10/11/2012 02:33 AM, Mike Hommey wrote:
On Wed, Oct 10, 2012 at 05:57:53PM -0400, Justin Lebar wrote:
By "turning off Linux PGO testing", you really mean "stop making and
distributing Linux PGO builds," right?
The main reason I'd want Linux PGO is for mobile. On desktop Linux,
most users (I
> 5-20%, if it were a general slowdown, is _huge_. We have people who work
>
> for months to get speedups of 1 or 2%.
Yes, I know, that is pretty much all I do at Mozilla ;) I don't think scattered
Talos wins of 5-20% are so valuable and important that we should keep
sacrificing developer time
On 11/10/12 08:54, David Anderson wrote:
Keep in mind that debug builds are probably at least an order of
magnitude slower (or a large factor), whereas PGO is a very small
factor. (After all, we do not PGO on Mac and it doesn't seem to be a
problem.)
5-20%, if it were a general slowdown, is _hu
Tim Taubert wrote:
Nobody likes running debug builds because they're slower
I always run debug builds. What does that make me? ;-)
--
Warning: May contain traces of nuts.
___
dev-platform mailing list
dev-platform@lists.mozilla.org
https://lists.moz
Keep in mind that debug builds are probably at least an order of magnitude
slower (or a large factor), whereas PGO is a very small factor. (After all, we
do not PGO on Mac and it doesn't seem to be a problem.)
-David
On Thursday, October 11, 2012 12:05:35 AM UTC-7, Tim Taubert wrote:
> On 10/10
Right, exactly. I am arguing that testing PGO, which is a buggy optimization
pass, incurs too much developer cost to justify a "5-20%" talos improvement on
select benchmarks. On Linux, which is a very small percentage of our market
share, and where distributions make their own builds anyway.
Wh
On 10/11/2012 09:32 AM, Boris Zbarsky wrote:
> The suggestion, as far as I can tell, is to drop Linux PGO completely.
> We woudln't have it in nightly, Aurora, Beta, or releases. Compiling
> with PGO on Linux would be an unsupported configuration that we'd
> probably advise distros against, becaus
On 10/11/12 3:05 AM, Tim Taubert wrote:
Also, I'm not sure how this affects Telemetry results. In terms of perf
measurements we'd probably need to completely ignore everything from
non-release builds as the results might differ heavily for some use
cases.
I'm not following.
The suggestion, as
On 10/10/2012 11:57 PM, Justin Lebar wrote:
> The main reason I'd want Linux PGO is for mobile. On desktop Linux,
> most users (I expect) don't run our builds, so it's not a big deal if
> they're some percent slower. (Unless distros commonly do PGO builds
> of Firefox?) But we're not doing mobil
On Wed, Oct 10, 2012 at 05:57:53PM -0400, Justin Lebar wrote:
> By "turning off Linux PGO testing", you really mean "stop making and
> distributing Linux PGO builds," right?
>
> The main reason I'd want Linux PGO is for mobile. On desktop Linux,
> most users (I expect) don't run our builds, so it
On Wed, Oct 10, 2012 at 5:57 PM, Justin Lebar wrote:
> By "turning off Linux PGO testing", you really mean "stop making and
> distributing Linux PGO builds," right?
>
> The main reason I'd want Linux PGO is for mobile. On desktop Linux,
> most users (I expect) don't run our builds, so it's not a
Yeah, if we're not testing them I guess we don't have to make or distribute
them at all.
-David
On Wednesday, October 10, 2012 2:58:18 PM UTC-7, Justin Lebar wrote:
> By "turning off Linux PGO testing", you really mean "stop making and
>
> distributing Linux PGO builds," right?
>
>
>
> The m
By "turning off Linux PGO testing", you really mean "stop making and
distributing Linux PGO builds," right?
The main reason I'd want Linux PGO is for mobile. On desktop Linux,
most users (I expect) don't run our builds, so it's not a big deal if
they're some percent slower. (Unless distros commo
44 matches
Mail list logo
