MemShrink progress, final

I was due to write a MemShrink progress report today, but I’ve decided that after almost 2.5 years, my reserves of enthusiasm for these regular reports has been exhausted.  Sorry!

I do still plan to write posts when significant fixes relating to memory consumption are made.  (For example, when generational GC lands, you’ll hear about it here.)  I will also continue to periodically update the MemShrink “big ticket items” list.  And MemShrink meetings will continue, so MemShrink-tagged bugs will still be triaged.  And for those of you who read the weekly Platform meeting notes, I will continue to write MemShrink updates there.  So don’t despair — good things will continue to happen, but they’ll just be marginally less visible.

23 Responses to MemShrink progress, final

  1. I’m sad to hear this, but it’s understandable since MemShrink is in its mature phase and doesn’t have so many obvious big ticket items to track at any one time now.

    I can’t begin to stress how important MemShrink is/has been to my continued usage of all the Gecko browsers irrespective of how much the UX team test my sanity and patience on the Firefox side of things. This, e10s (when it’s fully implemented), continued improvements in JS performance and support of web standards is what I really find crucial in a browser. Good luck with MS in the future and may the good things continue to happen.

  2. Could someone else step up and do these reports in your place? I’m not volunteering, but I’m sure we could find someone. ;)

  3. I’m sad to see the reports go. Seeing my quick response, you can tell I was looking forward to this day ;)

    However, I do understand that it’s been hard work for you to keep up, while you obviously were busy doing awesome stuff for Firefox (or otherwise, these reports wouldn’t even be here :) ). So in my name and probably all other readers: thank you for your hard work and thank you to keep us up to date for this long. Keep up the great work!

    I have one final request for you (if it is possible): could you add a URL to the link section of your blog which always points to fixed bugs in the past week with the [MemShrink] tag? It’ll be easier for us to still keep up a little bit, I personally hate the Bugzilla search tool and just can’t get the hang of it :(

    • Nicholas Nethercote

      The link is a good idea, but I can’t work out how to add a new link there O_o

      • Why not just add the link to the platform meeting minutes? I’ve been doing the same for fixed orange bugs since I started participating in the meeting.

  4. It was great to follow the memshrink progress here with you through the years. Thank you!

  5. Andrej Hribernik

    Thanks for posting all this time, was very interesting following the teams progress!

  6. Thanks for this valuable reports. They were a pleasure to read!

  7. Thanks for posting them for so long, and for all the hard work on Memshrink. You and all the other contributors have made (and continue to make) a massive difference to Firefox. I’m seriously impressed with the scale of the changes you peeps have brought about :-)

  8. Thanks a lot for all your reports. I read all of them from the beginning and that’s been a fantastic way to know what was going on on the memory side.
    That also helped to keep hope in Firefox in the difficult days of Firefox 4-6 which had terrible memory issues which really impacted performance.

  9. First, a question: Where are the weekly platform notes you mentioned located at?

    Second, an observation: Between your stopping regular memshrink updates and Taras’s snappy update blog posts having gone dark a year ago; I suspect it won’t be that long until the reduced visibility leads to renewed accusations that no-one at Mozilla cares about performance. If my assumption that the increased overhead of the many-process architecture E10S will impose will increase memory use at and shortly after start up (even if it results in an improvement over longer periods of time) probably sometime late next year when that major upgrade is expected to be released.

  10. Let me echo what the others have said here. Thank you for all your work.

    In the dark days after the 4.0 release, this blog was really started momentum in the other direction. It gave those who love FF hope.

    It is understandable that the need for these posts are diminished as the perception the “FF is a hog” days are over. However I am sad to see them go. I cannot begin to describe the excitement as your enthusiasm was contagious. When you were posting every week, this blog was what I looked forward to most out of all my other readings.

    Thank you again and good luck in future.

    I will see you all again for ggc! I cant wait.

  11. Thought much about new projects beyond the scope of MemShrink that could save memory, increase performance, and so on?

    What about ideas for an HTML version 6? HTML/XML/SGML really is quite poor in many ways, with perhaps the worst idea being the overly verbose and redundant closing tag. If I recall, the purpose of doing it that way was to increase human readability. Turns out it didn’t, it just added clutter to the HTML. While we have one shortcut, the trailing slash to indicate a tag is opened and immediately closed, it’s not enough. If something like “” could be added to HTML as a universal closing tag, that would be a step in the right direction. If the verbose closing tag could then be eliminated from the language, that should simplify and speed up parsing ever so slightly. We’d lose one shortcut, the ability to close several tags at once, like this: …, but that’s no longer proper HTML anyway, if it ever was.

    Not sure how important it would be to make improvements of that sort to HTML, but I’m guessing it may not matter too much. These days the bulk of the data the Internet transmits is probably video. HTML traffic accounts for, what, less than 10% of the data? Maybe less than 1%?

  12. Thanks for your work on Memshrink and for this blog, and also congratulations for getting Firefox’s memory usage to the point where it is no longer necessary to publicise the Memshrink team’s efforts to such an extent.
    Following your link to the weekly meeting notes led me to this article about Memshrink – I don’t recall seeing it mentioned here though I may have missed it. I particularly like the statement “Eternal Persistence is the Price of Excellence”. Hopefully now the Memshrink approach and the appropriate testing is now fully embedded in the Firefox development culture.
    http://aosabook.org/en/posa/memshrink.html

  13. Another thank you for your work and your reports!

  14. Thanks Nick for all the hardwork you and others have done over the years to make Firefox better. I, like many others, have thoroughly enjoyed your progress reports. They will be missed. I was particularly looking forward to some comments about the recent TBPL memory issues and how they were resolved.

    On a completely different note, are you aware that the desktop version of Are We Slim Yet is not showing any new memory profiles past Nov 22, 2013? The mobile version has updates up to an including Nov 25, 2013 and then nothing again for all of the last week.

    • Nicholas Nethercote

      I asked John Schoenick (AWSY’s maintainer) about the hiatus, he said: “Mozilla’s Mountain View office had a power shutdown that friday, and the test daemon was stopped to prevent weirdness. I used the occasion to have it go back and archive old test data, which takes many hours and has to be done offline to avoid machine load causing noise, and hadn’t re-started everything. I’m having it archive a few more months then it will restart, http://areweslimyet.com/status/ will update when tests are running.” Thanks for noticing and asking!

  15. I last asked this a few months ago, are either GGC or E10S mature enough to do AWSY runs yet? I’ve read that as long as you don’t use addons E10S builds are currently usable (ETA mid/late next year for general release); I haven’t seen anything recent on the status of GGC.