On 2015/02/25 07:41:25, Sven Panne wrote:
On 2015/02/24 22:33:45, michael_dawson wrote:
> In terms of turning on the build bots and getting test coverage would
> this be in simulated PPC mode ?
As a first step, yes. I don't know what the plans are for bots with real
PPC
HW,
but even if/when we have such bots, the PPC simulator bots would still
stay,
otherwise the simulator itself would definitely bit-rot.
> I've been running in the simulated
> mode to validate the commits accepted so far and the tests pass
> other than a number of timeouts. (we also run in native mode in
> our internal testing were results are btter) If so, would the
> first step be to create a review that brings the ppc directories
> up to date based on changes in the common code since our last update
> along with disabling the tests which timeout ? I see that there are
> excludes for timeouts in simulated mode on other platforms, although
> the number I see in the PPC simulated runs is higher so this would
> just be a first step.
That's perfectly fine. Having a relatively high number of tests marked as
SLOW
and/or TIMEOUT for a new platform is the usual way to get a port working.
Of
course this number should be reduced over time, but that's in the hands of
people caring about the platform port.
> For us, getting the performance opts in are important, otherwise we
have to
> maintain the deltas in our repo, but I can see how getting the build
bots
> and test coverage enabled is a good step.
So what needs to be done? The steps I see so far are:
* Reduce this CL at hand to the kBootCodeSizeMultiplier change and
land it
(that's harmless for the rest of the system and non-controversial).
* Remove any use of V8_OOL_CONSTANT_POOL from your code and even from
your
"internal deltas".
* Remove the PPC platform-dependent hack in serialize.cc, perhaps by
adding
a
new RelocInfo::Mode.
* Any other changes you still need to get things running and
integrated.
Are
there any?
What I've been looking at today is the tests which timeout. I upped the
timeout
to 120s and the number dropped from 38 to 13 so I'm seeing what 180s does.
I'm
assuming its better to give the tests a bit more time to complete so that
we get
better coverage. It takes a while to run the full suite in simulated mode
to it
will take me a little while to find the right combination of larger
timeout/test
exclusions.
https://codereview.chromium.org/882263003/
--
--
v8-dev mailing list
v8-dev@googlegroups.com
http://groups.google.com/group/v8-dev
---
You received this message because you are subscribed to the Google Groups "v8-dev" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to v8-dev+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.
