Re: [Wikitech-l] Facebook's HHVM performance sprint retrospective highlights MediaWiki gains
2015-06-12 18:47 GMT+03:00 Quim Gil q...@wikimedia.org: On Fri, Jun 12, 2015 at 12:59 PM, Gerard Meijssen gerard.meijs...@gmail.com wrote: Hoi, And good news it is :) Thanks Ori :) were people of the WMF involved in this ? It's not that Ori learned about these news by reading Twitter. ;) Tim Starling got dozens of patches merged in HHVM as well. Both are listed at https://github.com/facebook/hhvm/graphs/contributors, and you can learn about the rest of the team that has been working on this at https://www.mediawiki.org/wiki/HHVM See also https://blog.wikimedia.org/2014/12/29/how-we-made-editing-wikipedia-twice-as-fast/ A better question IMO is: have the FB engineers contributed any patches to MW? I'm not sure why is that a better question, considering that HHVM is running in Wikimedia servers and the numbers show the very tangible value it is providing to Wikimedia users. Well, there are a number of reasons that made me consider their contributions more important: 1. Downstream contributions are much less common than upstream ones 2. Intuitively, the number of developers FB has on HHVM is significantly larger that the number of WMF enigineers working on the subject. More eyes on our code is a good thing. 3. MW optimizations for HHVM will bring more to the WMF deployment that generic HHVM optimizations (see the part from the blog where they mention that some HHVM optimizations had to be reverted or reduced in order to have other features work). This is a risky path for the open source MediaWiki though, as sooner or later I expect some announcement that MW will only work on HHVM, which would be a pity. Strainu In any case, as Krenair has shown there are some patches indeed and, well, a Facebook engineer came to work to the WMF offices with Ori and friends during a month or so last Summer, so you see their willingness to help us. For what I have seen, so far this has been an exemplar free software collaboration with different orgs, timezones, and latitudes involved. Beyond the code, I bet we have learned something about how Facebook's free software developers work and vice versa. I wish Wikimedia is capable of attracting and engaging in similar free software partnerships, at this level and with these results! -- Quim Gil Engineering Community Manager @ Wikimedia Foundation http://www.mediawiki.org/wiki/User:Qgil ___ Wikitech-l mailing list Wikitech-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-l ___ Wikitech-l mailing list Wikitech-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-l
Re: [Wikitech-l] Facebook's HHVM performance sprint retrospective highlights MediaWiki gains
Hi, order to have other features work). This is a risky path for the open source MediaWiki though, as sooner or later I expect some announcement that MW will only work on HHVM, which would be a pity. An interesting thought that may not be appreciated or shared by all third-party users or developers. [0] shows that at least a Hack only functionality was considered and the question emerges about future development considering statements like: ... 1.55 times faster than PHP 7 on a MediaWiki workload ... none of these frameworks take advantage of the asynchronous I/O architecture available in HHVM (i.e., async) ... [1] ... advantage of the async capabilities offered by Hack and HHVM, we were able to examine the potential for performance gains through async execution. ... [1] ... able to validate the importance of both JIT compilation and asynchronous execution for optimizing PHP performance ... [1] [0] https://phabricator.wikimedia.org/T99755 [1] http://hhvm.com/blog/9293/lockdown-results-and-hhvm-performance Cheers On 6/13/15, Strainu strain...@gmail.com wrote: 2015-06-12 18:47 GMT+03:00 Quim Gil q...@wikimedia.org: On Fri, Jun 12, 2015 at 12:59 PM, Gerard Meijssen gerard.meijs...@gmail.com wrote: Hoi, And good news it is :) Thanks Ori :) were people of the WMF involved in this ? It's not that Ori learned about these news by reading Twitter. ;) Tim Starling got dozens of patches merged in HHVM as well. Both are listed at https://github.com/facebook/hhvm/graphs/contributors, and you can learn about the rest of the team that has been working on this at https://www.mediawiki.org/wiki/HHVM See also https://blog.wikimedia.org/2014/12/29/how-we-made-editing-wikipedia-twice-as-fast/ A better question IMO is: have the FB engineers contributed any patches to MW? I'm not sure why is that a better question, considering that HHVM is running in Wikimedia servers and the numbers show the very tangible value it is providing to Wikimedia users. Well, there are a number of reasons that made me consider their contributions more important: 1. Downstream contributions are much less common than upstream ones 2. Intuitively, the number of developers FB has on HHVM is significantly larger that the number of WMF enigineers working on the subject. More eyes on our code is a good thing. 3. MW optimizations for HHVM will bring more to the WMF deployment that generic HHVM optimizations (see the part from the blog where they mention that some HHVM optimizations had to be reverted or reduced in order to have other features work). This is a risky path for the open source MediaWiki though, as sooner or later I expect some announcement that MW will only work on HHVM, which would be a pity. Strainu In any case, as Krenair has shown there are some patches indeed and, well, a Facebook engineer came to work to the WMF offices with Ori and friends during a month or so last Summer, so you see their willingness to help us. For what I have seen, so far this has been an exemplar free software collaboration with different orgs, timezones, and latitudes involved. Beyond the code, I bet we have learned something about how Facebook's free software developers work and vice versa. I wish Wikimedia is capable of attracting and engaging in similar free software partnerships, at this level and with these results! -- Quim Gil Engineering Community Manager @ Wikimedia Foundation http://www.mediawiki.org/wiki/User:Qgil ___ Wikitech-l mailing list Wikitech-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-l ___ Wikitech-l mailing list Wikitech-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-l ___ Wikitech-l mailing list Wikitech-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-l
Re: [Wikitech-l] Facebook's HHVM performance sprint retrospective highlights MediaWiki gains
Hoi, And good news it is :) Thanks Ori :) were people of the WMF involved in this ? Thanks, GerardM On 12 June 2015 at 06:08, Ori Livneh o...@wikimedia.org wrote: On Thu, Jun 11, 2015 at 9:02 PM, Ori Livneh o...@wikimedia.org wrote: a parse of the Barack Obama article. A parse of *English Wikipedia's* Barack Obama article. I am trying, I am trying... :) ___ Wikitech-l mailing list Wikitech-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-l ___ Wikitech-l mailing list Wikitech-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-l
Re: [Wikitech-l] Facebook's HHVM performance sprint retrospective highlights MediaWiki gains
I found https://gerrit.wikimedia.org/r/#/c/217157/ and https://gerrit.wikimedia.org/r/#/c/97110/1 On 12 June 2015 at 13:01, Strainu strain...@gmail.com wrote: At the very least, someone from WMF suggested the page: MediaWiki was benchmarked using the Barack Obama page from Wikipedia, as was recommended by an engineer from Wikimedia foundation as representative of their load. A better question IMO is: have the FB engineers contributed any patches to MW? Strainu 2015-06-12 13:59 GMT+03:00 Gerard Meijssen gerard.meijs...@gmail.com: Hoi, And good news it is :) Thanks Ori :) were people of the WMF involved in this ? Thanks, GerardM On 12 June 2015 at 06:08, Ori Livneh o...@wikimedia.org wrote: On Thu, Jun 11, 2015 at 9:02 PM, Ori Livneh o...@wikimedia.org wrote: a parse of the Barack Obama article. A parse of *English Wikipedia's* Barack Obama article. I am trying, I am trying... :) ___ Wikitech-l mailing list Wikitech-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-l ___ Wikitech-l mailing list Wikitech-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-l ___ Wikitech-l mailing list Wikitech-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-l ___ Wikitech-l mailing list Wikitech-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-l
Re: [Wikitech-l] Facebook's HHVM performance sprint retrospective highlights MediaWiki gains
At the very least, someone from WMF suggested the page: MediaWiki was benchmarked using the Barack Obama page from Wikipedia, as was recommended by an engineer from Wikimedia foundation as representative of their load. A better question IMO is: have the FB engineers contributed any patches to MW? Strainu 2015-06-12 13:59 GMT+03:00 Gerard Meijssen gerard.meijs...@gmail.com: Hoi, And good news it is :) Thanks Ori :) were people of the WMF involved in this ? Thanks, GerardM On 12 June 2015 at 06:08, Ori Livneh o...@wikimedia.org wrote: On Thu, Jun 11, 2015 at 9:02 PM, Ori Livneh o...@wikimedia.org wrote: a parse of the Barack Obama article. A parse of *English Wikipedia's* Barack Obama article. I am trying, I am trying... :) ___ Wikitech-l mailing list Wikitech-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-l ___ Wikitech-l mailing list Wikitech-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-l ___ Wikitech-l mailing list Wikitech-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-l
Re: [Wikitech-l] Facebook's HHVM performance sprint retrospective highlights MediaWiki gains
On Fri, Jun 12, 2015 at 12:59 PM, Gerard Meijssen gerard.meijs...@gmail.com wrote: Hoi, And good news it is :) Thanks Ori :) were people of the WMF involved in this ? It's not that Ori learned about these news by reading Twitter. ;) Tim Starling got dozens of patches merged in HHVM as well. Both are listed at https://github.com/facebook/hhvm/graphs/contributors, and you can learn about the rest of the team that has been working on this at https://www.mediawiki.org/wiki/HHVM See also https://blog.wikimedia.org/2014/12/29/how-we-made-editing-wikipedia-twice-as-fast/ A better question IMO is: have the FB engineers contributed any patches to MW? I'm not sure why is that a better question, considering that HHVM is running in Wikimedia servers and the numbers show the very tangible value it is providing to Wikimedia users. In any case, as Krenair has shown there are some patches indeed and, well, a Facebook engineer came to work to the WMF offices with Ori and friends during a month or so last Summer, so you see their willingness to help us. For what I have seen, so far this has been an exemplar free software collaboration with different orgs, timezones, and latitudes involved. Beyond the code, I bet we have learned something about how Facebook's free software developers work and vice versa. I wish Wikimedia is capable of attracting and engaging in similar free software partnerships, at this level and with these results! -- Quim Gil Engineering Community Manager @ Wikimedia Foundation http://www.mediawiki.org/wiki/User:Qgil ___ Wikitech-l mailing list Wikitech-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-l
Re: [Wikitech-l] Facebook's HHVM performance sprint retrospective highlights MediaWiki gains
On Thu, Jun 11, 2015 at 9:02 PM, Ori Livneh o...@wikimedia.org wrote: a parse of the Barack Obama article. A parse of *English Wikipedia's* Barack Obama article. I am trying, I am trying... :) ___ Wikitech-l mailing list Wikitech-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-l
[Wikitech-l] Facebook's HHVM performance sprint retrospective highlights MediaWiki gains
Facebook's HHVM team just completed their first performance lockdown, which they spent focusing on the performance of open-source PHP frameworks under HHVM. The achievement which they chose to highlight in their blog posts is a gain of 19% in their MediaWiki performance benchmark, which is – you guessed it – a parse of the Barack Obama article. https://code.facebook.com/posts/902199373155728/-inside-the-hhvm-lockdown/ http://hhvm.com/blog/9293/lockdown-results-and-hhvm-performance ___ Wikitech-l mailing list Wikitech-l@lists.wikimedia.org https://lists.wikimedia.org/mailman/listinfo/wikitech-l
