The number of nightly tests are 5 tests for video_PlaybackPerf and 6 tests for video_VDAPerf.
Each test of video_PlaybackPerf takes about 6 minutes on most boards.
(I found it takes about 10 minutes on link, but I guess it won't be the case since two tests run on link.)
Each test of video_VDAPef takes about 2 minutes.
If these tests are moved to perbuild, the perbuild run time will increase by 45 minutes at most.
As far as I checked, the run time of executing routing jobs are short enough that adding them doesn't matter.
That is definitely my expectation though.
It seems to be hard to make 100% sure that this moving doesn't cause sticking many tests in the job queue, because the run time depends on much things (user manual job, available DUTs on lab, a test timeout).
Therefore, let us move them experimentally, and move back again only if the problem happens.

The following revision refers to this bug:
  https://chromium.googlesource.com/chromiumos/third_party/autotest/+/81fe1afb25dedf39ce53759b0a49deae538ad5d2

commit 81fe1afb25dedf39ce53759b0a49deae538ad5d2
Author: Hirokazu Honda <hiroh@chromium.org>
Date: Tue Oct 17 13:54:23 2017

autotest: Move video_VDAPerf and video_PlaybackPerf test runs from nightly to per build

video_VDAPerf and video_PlaybackPerf are on dashboard.
The metrics measured by the tests are important, which these regressions
are not allowed on the present ChromeOS.
Let us change their suite settings from crosbolt_perf_nightly to
crosbolt_perf_perbuild to increase run frequency to improve the test coverage.

BUG= chromium:772694 
TEST=None

Change-Id: I3281dd21714f986c809f9d9642513642c985cb0e
Reviewed-on: https://chromium-review.googlesource.com/721223
Commit-Ready: Hirokazu Honda <hiroh@chromium.org>
Tested-by: Hirokazu Honda <hiroh@chromium.org>
Reviewed-by: Ilja H. Friedel <ihf@chromium.org>
Reviewed-by: Owen Lin <owenlin@chromium.org>
Reviewed-by: Ben Cheng <bccheng@chromium.org>

[modify] https://crrev.com/81fe1afb25dedf39ce53759b0a49deae538ad5d2/client/site_tests/video_PlaybackPerf/control.h264.4k
[modify] https://crrev.com/81fe1afb25dedf39ce53759b0a49deae538ad5d2/client/site_tests/video_PlaybackPerf/control.vp9.hfr
[modify] https://crrev.com/81fe1afb25dedf39ce53759b0a49deae538ad5d2/client/site_tests/video_PlaybackPerf/control.vp9.4k.60
[modify] https://crrev.com/81fe1afb25dedf39ce53759b0a49deae538ad5d2/client/site_tests/video_PlaybackPerf/control.vp9.4k
[modify] https://crrev.com/81fe1afb25dedf39ce53759b0a49deae538ad5d2/client/site_tests/video_PlaybackPerf/control.vp8.4k
[modify] https://crrev.com/81fe1afb25dedf39ce53759b0a49deae538ad5d2/client/site_tests/video_VDAPerf/control.vp8
[modify] https://crrev.com/81fe1afb25dedf39ce53759b0a49deae538ad5d2/client/site_tests/video_VDAPerf/control.h264
[modify] https://crrev.com/81fe1afb25dedf39ce53759b0a49deae538ad5d2/client/site_tests/video_VDAPerf/control.vp9.4k
[modify] https://crrev.com/81fe1afb25dedf39ce53759b0a49deae538ad5d2/client/site_tests/video_VDAPerf/control.h264.4k
[modify] https://crrev.com/81fe1afb25dedf39ce53759b0a49deae538ad5d2/client/site_tests/video_VDAPerf/control.vp8.4k
[modify] https://crrev.com/81fe1afb25dedf39ce53759b0a49deae538ad5d2/client/site_tests/video_VDAPerf/control.vp9
[modify] https://crrev.com/81fe1afb25dedf39ce53759b0a49deae538ad5d2/client/site_tests/video_PlaybackPerf/control.vp8

The above CL was landed to ChromeOS 10043.0.0.

Quick update; One day have passed since these tests are moved to perbuild.
Looks any timeout doesn't happen because of this moving.
The number of PlaybackPerf and VDAPerf executed tests are stabilized on all the boards.
I continue monitoring carefully in one week though.

I has been looking for 1 week.
Looks like any timeout hasn't happened by this moving.