All graphs for this bug:
  https://chromeperf.appspot.com/group_report?bug_id=797404

(For debugging:) Original alerts at time of bug-filing:
  https://chromeperf.appspot.com/group_report?sid=5becea355283ee4b3ea05cd4822185ad3151ba21d457beb7cfb62c0b17718683


Bot(s) for this bug's original alert(s):

chromium-rel-mac-retina
chromium-rel-mac11-pro
linux-release

📍 Pinpoint job started.
https://pinpoint-dot-chromeperf.appspot.com/job/14a76a51040000

📍 Found a significant difference after 1 commit.
https://pinpoint-dot-chromeperf.appspot.com/job/14a76a51040000

[heap] fix regression on thread_times.key_silk_cases
By ofrobots@google.com · Wed Dec 20 14:15:08 2017
v8 @ aefc8a315343b124f16a2cfdaab3b086629e145d

Understanding performance regressions:
  http://g.co/ChromePerformanceRegressions

📍 Pinpoint job started.
https://pinpoint-dot-chromeperf.appspot.com/job/14978ce9040000

📍 Found a significant difference after 1 commit.
https://pinpoint-dot-chromeperf.appspot.com/job/14978ce9040000

[heap] fix regression on thread_times.key_silk_cases
By ofrobots@google.com · Wed Dec 20 14:15:08 2017
v8 @ aefc8a315343b124f16a2cfdaab3b086629e145d

Understanding performance regressions:
  http://g.co/ChromePerformanceRegressions

The pin pointed commit (aefc8a3) in question was a fix for regression  https://crbug.com/795323  on andriod devices. While the original android regression is fixed, it seems that we have a different regression as pointed out by this bug.

Interesting things to note from the charts:
* The culprit for 795323 (4623b82) had actually improved the scavenge times on these particular benchmarks (while regressing some on Android). Of course, with the culprit rectified, we have regressed back to the original performance. In other words, the current pin pointed commit (aefc8a3) shouldn't be considered a regression as far as scavenger time is concerned.
* However, the system-memory:private_dirty_size chart does not have a corresponding improvement when the culprit for 795323 landed (4623b82).
* The 'ref' lines on the charts seem to show a significant improvement as a result of the pin pointed commit (aefc8a3). What does the 'ref' line represent?

Issue 797905 has been merged into this issue.

I compared memory dumps in the good trace and the bad trace (linked below).

The only difference in "private dirty" memory is in
/private/var/folders/p4/py6nlvcx4fs5s3fg51_pzqmc0000gm/0/com.apple.LaunchServices-175-v2.csstore 131,072.0 KiB
/private/var/folders/p4/py6nlvcx4fs5s3fg51_pzqmc0000gm/0/com.apple.LaunchServices-175-v2.csstore 28,028.0 KiB

This shows up in the bad trace but does not show up in the good trace. This seems unrelated to Ali's change.


The good trace:
https://00e9e64bac2324df39f7f3d9fb353ed9dd4f00c185a0ebc206-apidata.googleusercontent.com/download/storage/v1/b/chrome-telemetry-output/o/https___mail_google_com_mail__2017-12-20_13-09-53_21110.html?qk=AD5uMEt-PKbNl88Bjwpy_3zX__yva5-LGDeyvzXv_IYuP0xzZOcknwVG05qenml5c4hj2SdpL4CGlRJXpgkbVsaq4R43YdKq-nbP2oU3qbd0s9EgjT4mhto6t3L-VewJglBSg0tChTifO-mxUwu1pK7i38fC8dVPeQUPcHgScS32ASiqnOuk4PldvlX7i4rGDoUwK_zUUqI0WfrCbd5tTV7Q8--dXoh4IoXwFp6VJicQqAHqOoABVM8k2Vsvqvi-dgFWZQShP8SW0rMvTzKP-Gc9au9HLG3FQBSV3fWEN125fk1KQ9rIIWQ9YbhtaDky7EXnV5swoKpQpga8hygpAjceUpHe0h6Yuf7Xi0QGctDC7EuRxKbimf0Yk64CPRi7PLDvfASNqTf-dlZowrF9gAXkX5AqIhCh7bCQVTm8ufO7zSYclQ381JwfJs3p1dEf8fVYSKG1WQ4l2O1CytNNjtPdIodtkCanDH4sgmq55UPHJLG4cDa0fDAFP2iJs4tTdFQkJw0O_Mgdqke-Te7Yw3GK0QwPlFFQ3NEaPXy5XLhhhvu10NscDRhoZuxbW7hkXGtPCwO_f6ONysSQBPJG0o1uCBkTP1Fg8mrTyZsBeeGWDo-MihAyV8L5iZNmTCkeRkSaVIC85lf4qfIdP06F8g6-hamCtnzUhK7oTX24u-ojWAWrJcsA3SZZ1qgABqMT1Ap0N58UwhJNVxXFiW2tFzQmOp7gWQefLBdBfnh6jnZvRrp-Mjd2ngsOm9JDPkn75BurTsj7nTAXpCwPII2sM7MtysO5HSwy1yO_9Qf3BrLrevzQ97ccMYlCZpCLn7swBVxsWdx4DPct

The bad trace:
https://00e9e64bacb898dfb14394e860ebf7950a7eddc4762e84e3cd-apidata.googleusercontent.com/download/storage/v1/b/chrome-telemetry-output/o/https___mail_google_com_mail__2017-12-20_18-35-44_87859.html?qk=AD5uMEtBGkm-KNf0Wj79bwuwoXihcJS--FnyYXXCCyFGnMBniIFuiO5qxc6MBtAOlg9LkbStFjN51fcNAU2HryP4kMnJYHa2fEE4lmuk3TUCXVAIuM8jR9SbDfpvS-aBY9B6S18nWpyHJ3QO9liVExCxeKfXEgr_YR4WmwP73mCxFs50xdlmrXhsFOAHCfW_In0TJ5jpS--TYxD1AR7yj751MFh6M_ReKZCZVFRr6usv9dF04XY3GK5FZsbFMSgGFGTrTTyC7tHu28aMLeB5uO6_TI6yqJH9rTSt5jYUN7lp2tsyD9lKhtB5MKXkmcaK-YXdkXGAhdrou6CYve-PGEKicYTrj_K4zyqgrHrsgnzvpsLslmdWFMxWwOlGA3pf00YOWlG1Fk5a4K1bCqEKGcLIOAMhBu2A_55YcHQNmX0jEcJlZLMoGICUQDr4G0HqaGm9Px945DnFPXp2oami7erhUa7WLrBjcdI-7kFR1ySRd7zYnSNKbXPtvz504oKA74vBaLRPR8noqzlOVYTS3mgSktigNVUNmsHhgDaJ6_v5MFdOAO1SCx50NTFhG3hgO5iv8PEXvdY7TeQSaDSmr_oQ25HMVmTcDGqwGE8SPqvKaS4MkkXo001F3pOWnlmNhhxmpIrpz8P2-dsUDev0bHvYPS5BhxlP7XUDolgqxv9j9XCFTsM8LjqJi-uchE8ALAKRng_7UgeBE28xp4o8jIuT6Um0cG2IZakHR7yBWDEuX2lECboj2W1KrCmm0r7AEf-LjjE3ZevIa6NZdHeRj_VCZS86SSATffwUjOxWjWskXfDMBsS0wVi4P6jlL5YooLRlO89RAK67

The ref lines changed at the same point, which suggests that the environment infra changed at the same time.

Can this issue be closed as invalid?

+1 for "won't fix"-ing it.

Closing as 'wont-fix'. 

* the scavenge time delta is expected and acceptable as per #6.
* the private dirty memory delta is unrelated.