Issue metadata
Sign in to add a comment
|
No data received for smoothness.top_25_smooth from win-zenbook since 402383 |
||||||||||||||||||||
Issue descriptionSee the link to graphs below.
,
Jul 11 2016
I'm having a hard time figuring out what caused the ref builds to stop reporting data here. The buildbot stdio has expired, which makes this tricky. It looks like wikipedia isn't producing any results at all for the ref build, but it is producing results for the non-ref build? I wonder why wikipedia would be failing, but only on the ref build... Here's a trace for the non-ref build: https://00e9e64bacd119245edca6ba1259402818fc9acf2c66c46f0b-apidata.googleusercontent.com/download/storage/v1_internal/b/chrome-telemetry-output/o/trace-file-id_11-2016-07-08_14-06-39-25973.html?qk=AD5uMEssRh5GA6uZ6mZ_9zKwqVUVX7oiAXtM88zsW6ePpF27WBS3bMEkkWjRiLLC174yicXdluIP_r6adItKHncX9gepR0ODwu9X-5eb1dZf3y5Sd_GTetkJVRbpRiuTUU1t9cQ_duPqp9-R7SuPIyzAlY3ONOIGMqAqUkNzVtUN6kvqChYFOeCP2Ce7Ua1fZpguyxejN16Jo5oqmfhjrqFbwfzqUaQKh820EKfCLl5yaRNgYSQ2UoyOzwgdtd2AFndlbxWeh__uDL0Osyxn9VyJSiQf9ZTCMNLQeTFQfLAYO0D0ZVP6RYxNRuo93DqnHiM62DZebbWImx9Gn5nm5zDblBZ_VjqEd4-7J5YK-kd1mog2mzmkxdIYUepdyCh5R9Wi_ENj5m45fhrliOAQSp1x96ajXELyng9xCpV0rLuXarz7SqSigjo22khOj3F90puGp9hrSCIpwv4gAqQWrjzoPR4eSbe1ghN4-96hucRtyGg6ah9W6nrRSJ9ggGo8dpIsQr0M5wscQ0q4_Sgy-fGGgQWshrE9-oDZCbzrDn-1WjsXUeQUV9aRMm8ksDxYz10rtP-ufSGp7TLhZLrC9nKOdwEuN-DklnjpY6FB4j98-iohRYB3fkaLe5xveaa85iT4ZYA-k805X4OYLR_HKPk5IyFfQGOn0YSx61iyhp41_ZFpL3XtPNqQ-N9Jjx_XEPGY9qfj0c56lzei2pB2cFGdOd6ZH7ihwDFJYimlrNpTVgA8n61jgaXxgvSq9f_Sn1rMRrlf8YxESYJFnmeZC8gsfZPCFauQ_gph5OapXfyGxlDUJtcYOi0GrrXZ9UkNpQPnMCYvpPMyDe6epv-sou0Vouj1sfvFloBvXGZEVbmrVk6Uy8BxOpEHVB-39asHANbKEzwLar1N And a trace for the ref build: https://00e9e64bac80f684f020b28185b420ca4dd01cb8495faa0bd7-apidata.googleusercontent.com/download/storage/v1_internal/b/chrome-telemetry-output/o/trace-file-id_11-2016-07-08_14-13-10-11602.html?qk=AD5uMEsrbpwtBnpUe7_in0rcI3vVVmnOsUJ0VGy2A_Dvc9N1hswKAI_blxw9QaTqPrfFWNZi8NghAVDqd90djIfLWRC894O7WRaij6iON4EG-DLwOWZ3v7H29RDt4TaDP_Fc5WS4EQlthIrimg0uhocsTfWV0eenWue0f-9bA1e1zTn579La0bnkMqKBNyPShMvlmEYI5VNkvdN0AvgqEierz1itL1TlwvEz93cLpBirF70tmWT_SMgC7JjLM1yOe3vbk_WydIp4mHi3fW7ltSQnW1mgU6Vu14TS3F5gI-BZeKsRJtseXNK4HDh2jKzz-nLl3pjDm1-1aYcwt6ryXMRXBR9eT3QwQHRh_eBAbe4LhGt-iP2dqk8uPDss8gEtkhpBwRYqdxd-cvcSrEwtDSmcz6Z8TG6WHoT7gD0XxReTHI8bs4RIRzFgwtbEhIiCcORbnptieWNYE0FiqcIx8OopmS96FCftPOKftPJ5uj17ERDIQ34TiLQKi1wvSLUspCkoQVVtCsnUT5wFDP7UqFkIgbh7u_zPwHT3AnEH-m8oaTTALd_t6GqA3II4mNfuzUj8FhcZxGktJeooafjhbl5PnC2SIh_WLePOBYUz0AnbsOAtXlzXfRegH2ugGlJkSxLox8B5IlMRL6qPtZziN_liULmuLJvM6DgdxnbjGhMex85PKQPhF5jiEU5qurpLDaLsK8Ol0Xxn2r71TZDkViCm_zXI1aBxmS0etarfdaHgUX50INUCMJcVWnAHH2RGZtfoZMYJpTAjXLGU4aufASrOWrwohPjFBHspmeZuuRBKthNiFrpnEaj3v54_1wTvUkUvHD7alAfRyfbCijBnuQoD0TQTWBhjpcC52Ul8bkX-hDM0C6Jy7l765WpnTt0abApsP_2ys8Ff
,
Jul 11 2016
The ref build trace looks completely broken. Annie, any thoughts on next steps here?
,
Jul 11 2016
What I'm seeing: Data coming back in for Wikipedia on win10 and win7-dual-ati: https://chromeperf.appspot.com/report?sid=74599e89dd763335cdf6d1697d1d4836c0758e76a389b09a4fe1331ed58b903f&rev=400386 https://chromeperf.appspot.com/report?sid=2d6df9e355dd46cd149a1648134ff47b4de848d48f1c0c33c31cad7e674b1866&rev=400386 Broken ref build for news.yahoo.com on zenbook: https://chromeperf.appspot.com/report?sid=1379e4aa2273c44b79f6a5b223dc1938b015b7af88f63d8742a8de2c290feb13 Looking at the log of a recent build on zenbook, news.yahoo.com looks the same as the other sites, but there is a warning on all of them for no importer found: https://build.chromium.org/p/chromium.perf/builders/Win%20Zenbook%20Perf%20%283%29/builds/2115/steps/smoothness.top_25_smooth.reference/logs/stdio [ RUN ] http://news.yahoo.com (INFO) 2016-07-08 17:52:17,630 cache_temperature.EnsurePageCacheTemperature:55 PageCacheTemperature: any (INFO) 2016-07-08 17:52:18,203 chrome_tracing_agent._CreateTraceConfigFile:248 Trace config file string: {"trace_config": {"included_categories": ["webkit.console", "benchmark", "blink.console", "trace_event_overhead"], "record_mode": "record-as-much-as-possible"}} (INFO) 2016-07-08 17:52:18,223 tracing_backend.StartTracing:131 Start Tracing Request: {'params': {'transferMode': 'ReturnAsStream', 'options': 'record-as-much-as-possible', 'categories': 'benchmark,blink.console,trace_event_overhead,webkit.console'}, 'method': 'Tracing.start'} (INFO) 2016-07-08 17:52:24,438 trace._GetTempFileHandle:65 Trace (TraceDataPart("traceEvents")) of size 1315136 bytes saved. (INFO) 2016-07-08 17:52:24,440 trace._GetTempFileHandle:65 Trace (TraceDataPart("tabIds")) of size 40 bytes saved. (INFO) 2016-07-08 17:52:24,444 trace._GetTempFileHandle:65 Trace (TraceDataPart("telemetry")) of size 946 bytes saved. c:\users\chrome~1.lab\appdata\local\temp\tmp56j5nm.html (WARNING) 2016-07-08 17:52:33,563 model._CreateImporters:275 No importer found for TraceDataPart("telemetry") Ned, do you know what that error means? In the chartjson for the ref build, we get: "http://news.yahoo.com": { "description": "First gesture scroll update latency measures the time it takes to process the very first gesture scroll update input event. The first scroll gesture can often get delayed by work related to page loading.", "important": true, "improvement_direction": "down", "name": "first_gesture_scroll_update_latency", "none_value_reason": "Merging values containing a None value results in a None value.", "page_id": 17, "std": null, "type": "list_of_scalar_values", "units": "ms", "values": null },
,
Jul 11 2016
That "No importer found for TraceDataPart("telemetry")" basically mean we don't implement a python importer for telemetry trace, which can be safely ignore because we are pushing everything to TBM2
As you pointed out, the page that which blows up the whole thing is news.yahoo.com. Its first_gesture_scroll_update_latency is null due to some reason.
Trace of news.yahoo.com:
https://console.developers.google.com/m/cloudstorage/b/chrome-telemetry-output/o/trace-file-id_17-2016-07-08_17-55-04-60162.html
,
Jul 12 2016
There are no gesture scroll update latency slices in that trace, but we are seeing a bunch of touch move latency slices.
Win 7 Intel is missing data for wikipedia:
"Wikipedia (1 tab)": {
"description": "First gesture scroll update latency measures the time it takes to process the very first gesture scroll update input event. The first scroll gesture can often get delayed by work related to page loading.",
"important": true,
"improvement_direction": "down",
"name": "first_gesture_scroll_update_latency",
"none_value_reason": "Merging values containing a None value results in a None value.",
"page_id": 11,
"std": null,
"type": "list_of_scalar_values",
"units": "ms",
"values": null
},
but has data for yahoo
"http://news.yahoo.com": {
"description": "First gesture scroll update latency measures the time it takes to process the very first gesture scroll update input event. The first scroll gesture can often get delayed by work related to page loading.",
"important": true,
"improvement_direction": "down",
"name": "first_gesture_scroll_update_latency",
"page_id": 17,
"std": 0.0,
"type": "list_of_scalar_values",
"units": "ms",
"values": [
11.45
]
},
,
Jul 12 2016
This is the wikipedia trace: https://console.developers.google.com/m/cloudstorage/b/chrome-telemetry-output/o/trace-file-id_11-2016-07-08_17-54-51-63667.html It looks very wrong. Maybe the page no longer load & 404 page is showing? :-/
,
Jul 12 2016
,
Jul 12 2016
Moving this nonessential bug to the next milestone. For more details visit https://www.chromium.org/issue-tracking/autotriage - Your friendly Sheriffbot
,
Jul 12 2016
On ChromiumPerf/chromium-rel-win7-gpu-intel/smoothness.top_25_smooth / frame_times / Blogger, we stopped getting ref data at the same time as we started getting non-ref data. That seems kinda suspicious.
,
Jul 12 2016
Issue 627214 has been merged into this issue.
,
Jul 12 2016
I suspect this could be an issue with the WPR recording or telemetry change.
,
Jul 12 2016
Most suspicious thing is this catapult roll: https://chromium.googlesource.com/external/github.com/catapult-project/catapult.git/+log/9f8a0a8c0d45..71bab5766221 Nothing is really jumping out at me though.
,
Jul 22 2016
Ned, do you think there's anything in the catapult roll that could be causing this?
,
Jul 22 2016
No. I don't think the catapult roll is causing this. The best way to diagnose this is to run the benchmark locally on win-zenbook & see what happening on the page.
,
Jul 25 2016
I don't have access to that device - could someone who does take this?
,
Aug 5 2016
Can we get an owner for this? Any volunteers who have access to win-zenbook?
,
Aug 8 2016
Assigning to sullivan@ to delegate for the moment.
,
Aug 10 2016
,
Aug 18 2016
,
Aug 18 2016
,
Sep 30 2016
This has recovered. |
|||||||||||||||||||||
►
Sign in to add a comment |
|||||||||||||||||||||
Comment 1 by tdres...@chromium.org
, Jul 11 2016