Add ppapi perftests infrastructure, plus a performance test for PPP_Messaging.

Add ppapi perftests infrastructure, plus a performance test for PPP_Messaging.

BUG=
TEST=


Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=118888

[dmichael (off chromium), 2012-01-20 17:29:58]

I have another CL to cut elide some string copies in the OOP proxy. It's showing
a modest improvement in testing so far on Debug builds:
https://docs.google.com/a/chromium.org/spreadsheet/ccc?key=0AgmkMxnFvhjmdEp1R...

(It's more of a mixed bag for Release builds so far, but I still have some
things to try.)

[piman, 2012-01-20 17:39:23]

LGTM overall

http://codereview.chromium.org/9195029/diff/5002/ppapi/proxy/ppapi_perftests.cc
File ppapi/proxy/ppapi_perftests.cc (right):

http://codereview.chromium.org/9195029/diff/5002/ppapi/proxy/ppapi_perftests....
ppapi/proxy/ppapi_perftests.cc:1: // Copyright (c) 2011 The Chromium Authors.
All rights reserved.
2012

http://codereview.chromium.org/9195029/diff/5002/ppapi/proxy/ppp_messaging_pr...
File ppapi/proxy/ppp_messaging_proxy_perftest.cc (right):

http://codereview.chromium.org/9195029/diff/5002/ppapi/proxy/ppp_messaging_pr...
ppapi/proxy/ppp_messaging_proxy_perftest.cc:61: PP_Var host_string =
StringVar::StringToPPVar(kTestString);
TBH, if you truly want to measure the cost of sending a string, this should be
included inside the loop, and the PP_Var->string transformation should also be
done in HandleMessage.

[dmichael (off chromium), 2012-01-20 17:53:20]

http://codereview.chromium.org/9195029/diff/5002/ppapi/proxy/ppp_messaging_pr...
File ppapi/proxy/ppp_messaging_proxy_perftest.cc (right):

http://codereview.chromium.org/9195029/diff/5002/ppapi/proxy/ppp_messaging_pr...
ppapi/proxy/ppp_messaging_proxy_perftest.cc:61: PP_Var host_string =
StringVar::StringToPPVar(kTestString);
On 2012/01/20 17:39:23, piman wrote:
> TBH, if you truly want to measure the cost of sending a string, this should be
> included inside the loop, and the PP_Var->string transformation should also be
> done in HandleMessage.
I thought about doing it that way, but I decided to measure the end-to-end
performance of just the proxy. I admit, sending the same string repeatedly is a
bit unrealistic,  If you think it's important to measure the complete end-to-end
operation, I'm happy to do that. But it shouldn't really change anything w.r.t.
comparing two alternate proxy implementations.

[piman, 2012-01-20 18:07:48]

On 2012/01/20 17:53:20, dmichael wrote:
>
http://codereview.chromium.org/9195029/diff/5002/ppapi/proxy/ppp_messaging_pr...
> File ppapi/proxy/ppp_messaging_proxy_perftest.cc (right):
> 
>
http://codereview.chromium.org/9195029/diff/5002/ppapi/proxy/ppp_messaging_pr...
> ppapi/proxy/ppp_messaging_proxy_perftest.cc:61: PP_Var host_string =
> StringVar::StringToPPVar(kTestString);
> On 2012/01/20 17:39:23, piman wrote:
> > TBH, if you truly want to measure the cost of sending a string, this should
be
> > included inside the loop, and the PP_Var->string transformation should also
be
> > done in HandleMessage.
> I thought about doing it that way, but I decided to measure the end-to-end
> performance of just the proxy. I admit, sending the same string repeatedly is
a
> bit unrealistic,  If you think it's important to measure the complete
end-to-end
> operation, I'm happy to do that. But it shouldn't really change anything
w.r.t.
> comparing two alternate proxy implementations.

Well, I'll argue that a proxy implementation might not transmit the string data
with the PP_Var, but only when asked, thus showing a dramatic improvement on
this benchmark when in reality it will likely cause a perf degradation (because
of extra round-trips).

[dmichael (off chromium), 2012-01-20 21:21:18]

Good point. Done.

[commit-bot: I haz the power, 2012-01-20 21:22:18]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/dmichael@chromium.org/9195029/13001

[piman, 2012-01-20 21:24:34]

lgtm

[commit-bot: I haz the power, 2012-01-20 22:16:50]

Try job failure for 9195029-13001 (retry) on linux_clang for step "compile"
(clobber build).
It's a second try, previously, step "compile" failed.
http://build.chromium.org/p/tryserver.chromium/buildstatus?builder=linux_clan...

[commit-bot: I haz the power, 2012-01-24 00:25:36]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/dmichael@chromium.org/9195029/15001

[commit-bot: I haz the power, 2012-01-24 01:37:08]

Try job failure for 9195029-15001 (retry) on mac_rel for step "compile" (clobber
build).
It's a second try, previously, step "compile" failed.
http://build.chromium.org/p/tryserver.chromium/buildstatus?builder=mac_rel&nu...

[commit-bot: I haz the power, 2012-01-24 02:07:18]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/dmichael@chromium.org/9195029/17003

[commit-bot: I haz the power, 2012-01-24 02:43:41]

Try job failure for 9195029-17003 (retry) on mac_rel for step "compile" (clobber
build).
It's a second try, previously, step "compile" failed.
http://build.chromium.org/p/tryserver.chromium/buildstatus?builder=mac_rel&nu...

[commit-bot: I haz the power, 2012-01-24 18:29:27]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/dmichael@chromium.org/9195029/20004

[commit-bot: I haz the power, 2012-01-24 20:29:05]

Change committed as 118888