Re: CLJ-703 - 10x compilation time decrease after applying one-line patch, no downsides.
Luc Prefontaine writes: > There's a better 'fix'. > Run your builds on SSDs or on a memory based file system. "Performance problems are solved best by throwing more hardware at them." Seriously?! No, I completely agree with Bozhidar in that regular bugfix releases would be a good thing. Clojure releases already have a minor version, so we might as well use it. Bye, Tassilo -- You received this message because you are subscribed to the Google Groups "Clojure" group. To post to this group, send email to clojure@googlegroups.com Note that posts from new members are moderated - please be patient with your first post. To unsubscribe from this group, send email to clojure+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/clojure?hl=en --- You received this message because you are subscribed to the Google Groups "Clojure" group. To unsubscribe from this group and stop receiving emails from it, send an email to clojure+unsubscr...@googlegroups.com. For more options, visit https://groups.google.com/d/optout.
Re: CLJ-703 - 10x compilation time decrease after applying one-line patch, no downsides.
There's a better 'fix'. Run your builds on SSDs or on a memory based file system. With the RAM available these days, any build can be done on a memory based file system. It speeds up every step by a huge factor. Not only compilation. You only need to copy your targets to a 'permanent' place. We have been doing this for years to shrink build times. We AOT most of our code base. Luc P. > I'll just say one more time that the team should really consider doing > bug-fix releases in the future. This problem sounds serious enough to be > handled as quickly as possible. Clojure 1.6 was released over one year ago, > so I'm guessing 1.8 is more than a year away from now. Waiting for major > releases for a bug fix in not exactly the greatest user experience. > > On 9 May 2015 at 05:18, Alexander Hudek wrote: > > > I'd like to chime in here in support of this, our company has been running > > a modified clojure build because of this for over a year now. > > > > Alex > > > > > > On Friday, May 8, 2015 at 2:12:50 PM UTC-4, Martin Raison wrote: > >> > >> Hi all, > >> > >> This issue has been around for a while without much activity, although a > >> very simple fix is already there: > >> http://dev.clojure.org/jira/browse/CLJ-703 > >> > >> We have a pretty big Clojure project that we compile on machines with > >> slow hard-drives, on CentOS 6. We noticed that compilation had become > >> extremely slow, however CPU usage was always below 10%. > >> > >> Simply removing the line "cfs.getFD().sync();" inside > >> "src/jvm/clojure/lang/Compiler.java" > >> dramatically improved the compilation time (Something like 10x, although at > >> this level of increase, measurement almost doesn't make sense anymore :)). > >> > >> We've been using the suggested fix (remove-sync-only.patch) for a few > >> weeks and haven't observed any issues. > >> > >> It seems we should really figure out whether this line is actually > >> necessary, or if it was added there "just in case". Even if some situations > >> require it, the speed bump is so huge that we might want some special-case > >> logic. > >> > >> Martin > >> > > -- > > You received this message because you are subscribed to the Google > > Groups "Clojure" group. > > To post to this group, send email to clojure@googlegroups.com > > Note that posts from new members are moderated - please be patient with > > your first post. > > To unsubscribe from this group, send email to > > clojure+unsubscr...@googlegroups.com > > For more options, visit this group at > > http://groups.google.com/group/clojure?hl=en > > --- > > You received this message because you are subscribed to the Google Groups > > "Clojure" group. > > To unsubscribe from this group and stop receiving emails from it, send an > > email to clojure+unsubscr...@googlegroups.com. > > For more options, visit https://groups.google.com/d/optout. > > > > -- > You received this message because you are subscribed to the Google > Groups "Clojure" group. > To post to this group, send email to clojure@googlegroups.com > Note that posts from new members are moderated - please be patient with your > first post. > To unsubscribe from this group, send email to > clojure+unsubscr...@googlegroups.com > For more options, visit this group at > http://groups.google.com/group/clojure?hl=en > --- > You received this message because you are subscribed to the Google Groups > "Clojure" group. > To unsubscribe from this group and stop receiving emails from it, send an > email to clojure+unsubscr...@googlegroups.com. > For more options, visit https://groups.google.com/d/optout. > -- Luc Prefontaine sent by ibisMail! -- You received this message because you are subscribed to the Google Groups "Clojure" group. To post to this group, send email to clojure@googlegroups.com Note that posts from new members are moderated - please be patient with your first post. To unsubscribe from this group, send email to clojure+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/clojure?hl=en --- You received this message because you are subscribed to the Google Groups "Clojure" group. To unsubscribe from this group and stop receiving emails from it, send an email to clojure+unsubscr...@googlegroups.com. For more options, visit https://groups.google.com/d/optout.
Re: CLJ-703 - 10x compilation time decrease after applying one-line patch, no downsides.
I'll just say one more time that the team should really consider doing bug-fix releases in the future. This problem sounds serious enough to be handled as quickly as possible. Clojure 1.6 was released over one year ago, so I'm guessing 1.8 is more than a year away from now. Waiting for major releases for a bug fix in not exactly the greatest user experience. On 9 May 2015 at 05:18, Alexander Hudek wrote: > I'd like to chime in here in support of this, our company has been running > a modified clojure build because of this for over a year now. > > Alex > > > On Friday, May 8, 2015 at 2:12:50 PM UTC-4, Martin Raison wrote: >> >> Hi all, >> >> This issue has been around for a while without much activity, although a >> very simple fix is already there: >> http://dev.clojure.org/jira/browse/CLJ-703 >> >> We have a pretty big Clojure project that we compile on machines with >> slow hard-drives, on CentOS 6. We noticed that compilation had become >> extremely slow, however CPU usage was always below 10%. >> >> Simply removing the line "cfs.getFD().sync();" inside >> "src/jvm/clojure/lang/Compiler.java" >> dramatically improved the compilation time (Something like 10x, although at >> this level of increase, measurement almost doesn't make sense anymore :)). >> >> We've been using the suggested fix (remove-sync-only.patch) for a few >> weeks and haven't observed any issues. >> >> It seems we should really figure out whether this line is actually >> necessary, or if it was added there "just in case". Even if some situations >> require it, the speed bump is so huge that we might want some special-case >> logic. >> >> Martin >> > -- > You received this message because you are subscribed to the Google > Groups "Clojure" group. > To post to this group, send email to clojure@googlegroups.com > Note that posts from new members are moderated - please be patient with > your first post. > To unsubscribe from this group, send email to > clojure+unsubscr...@googlegroups.com > For more options, visit this group at > http://groups.google.com/group/clojure?hl=en > --- > You received this message because you are subscribed to the Google Groups > "Clojure" group. > To unsubscribe from this group and stop receiving emails from it, send an > email to clojure+unsubscr...@googlegroups.com. > For more options, visit https://groups.google.com/d/optout. > -- You received this message because you are subscribed to the Google Groups "Clojure" group. To post to this group, send email to clojure@googlegroups.com Note that posts from new members are moderated - please be patient with your first post. To unsubscribe from this group, send email to clojure+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/clojure?hl=en --- You received this message because you are subscribed to the Google Groups "Clojure" group. To unsubscribe from this group and stop receiving emails from it, send an email to clojure+unsubscr...@googlegroups.com. For more options, visit https://groups.google.com/d/optout.
Re: CLJ-703 - 10x compilation time decrease after applying one-line patch, no downsides.
I'd like to chime in here in support of this, our company has been running a modified clojure build because of this for over a year now. Alex On Friday, May 8, 2015 at 2:12:50 PM UTC-4, Martin Raison wrote: > > Hi all, > > This issue has been around for a while without much activity, although a > very simple fix is already there: > http://dev.clojure.org/jira/browse/CLJ-703 > > We have a pretty big Clojure project that we compile on machines with slow > hard-drives, on CentOS 6. We noticed that compilation had become extremely > slow, however CPU usage was always below 10%. > > Simply removing the line "cfs.getFD().sync();" inside > "src/jvm/clojure/lang/Compiler.java" > dramatically improved the compilation time (Something like 10x, although at > this level of increase, measurement almost doesn't make sense anymore :)). > > We've been using the suggested fix (remove-sync-only.patch) for a few > weeks and haven't observed any issues. > > It seems we should really figure out whether this line is actually > necessary, or if it was added there "just in case". Even if some situations > require it, the speed bump is so huge that we might want some special-case > logic. > > Martin > -- You received this message because you are subscribed to the Google Groups "Clojure" group. To post to this group, send email to clojure@googlegroups.com Note that posts from new members are moderated - please be patient with your first post. To unsubscribe from this group, send email to clojure+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/clojure?hl=en --- You received this message because you are subscribed to the Google Groups "Clojure" group. To unsubscribe from this group and stop receiving emails from it, send an email to clojure+unsubscr...@googlegroups.com. For more options, visit https://groups.google.com/d/optout.
Re: CLJ-703 - 10x compilation time decrease after applying one-line patch, no downsides.
I'm glad to hear that. Thanks! Le vendredi 8 mai 2015 14:23:03 UTC-7, Alex Miller a écrit : > > Yes. > > > On Friday, May 8, 2015 at 3:09:06 PM UTC-5, Andy Fingerhut wrote: >> >> Just to clarify, Alex, since your idea of "next release" might be >> different than those who aren't following Clojure 1.7 alphas/betas/etc. >> carefully. >> >> Do you mean "hoping to look at it for Clojure 1.8, since Clojure 1.7 is >> to the point of primarily fixing critical bugs and regressions" ? >> >> Andy >> >> On Fri, May 8, 2015 at 12:07 PM, Alex Miller >> wrote: >> >>> I'm hoping to look at it for the next release. >>> >>> Alex >>> >>> -- >>> You received this message because you are subscribed to the Google >>> Groups "Clojure" group. >>> To post to this group, send email to clo...@googlegroups.com >>> Note that posts from new members are moderated - please be patient with >>> your first post. >>> To unsubscribe from this group, send email to >>> clojure+u...@googlegroups.com >>> For more options, visit this group at >>> http://groups.google.com/group/clojure?hl=en >>> --- >>> You received this message because you are subscribed to the Google >>> Groups "Clojure" group. >>> To unsubscribe from this group and stop receiving emails from it, send >>> an email to clojure+u...@googlegroups.com. >>> For more options, visit https://groups.google.com/d/optout. >>> >> >> -- You received this message because you are subscribed to the Google Groups "Clojure" group. To post to this group, send email to clojure@googlegroups.com Note that posts from new members are moderated - please be patient with your first post. To unsubscribe from this group, send email to clojure+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/clojure?hl=en --- You received this message because you are subscribed to the Google Groups "Clojure" group. To unsubscribe from this group and stop receiving emails from it, send an email to clojure+unsubscr...@googlegroups.com. For more options, visit https://groups.google.com/d/optout.
Re: CLJ-703 - 10x compilation time decrease after applying one-line patch, no downsides.
Yes. On Friday, May 8, 2015 at 3:09:06 PM UTC-5, Andy Fingerhut wrote: > > Just to clarify, Alex, since your idea of "next release" might be > different than those who aren't following Clojure 1.7 alphas/betas/etc. > carefully. > > Do you mean "hoping to look at it for Clojure 1.8, since Clojure 1.7 is to > the point of primarily fixing critical bugs and regressions" ? > > Andy > > On Fri, May 8, 2015 at 12:07 PM, Alex Miller > wrote: > >> I'm hoping to look at it for the next release. >> >> Alex >> >> -- >> You received this message because you are subscribed to the Google >> Groups "Clojure" group. >> To post to this group, send email to clo...@googlegroups.com >> >> Note that posts from new members are moderated - please be patient with >> your first post. >> To unsubscribe from this group, send email to >> clojure+u...@googlegroups.com >> For more options, visit this group at >> http://groups.google.com/group/clojure?hl=en >> --- >> You received this message because you are subscribed to the Google Groups >> "Clojure" group. >> To unsubscribe from this group and stop receiving emails from it, send an >> email to clojure+u...@googlegroups.com . >> For more options, visit https://groups.google.com/d/optout. >> > > -- You received this message because you are subscribed to the Google Groups "Clojure" group. To post to this group, send email to clojure@googlegroups.com Note that posts from new members are moderated - please be patient with your first post. To unsubscribe from this group, send email to clojure+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/clojure?hl=en --- You received this message because you are subscribed to the Google Groups "Clojure" group. To unsubscribe from this group and stop receiving emails from it, send an email to clojure+unsubscr...@googlegroups.com. For more options, visit https://groups.google.com/d/optout.
Re: CLJ-703 - 10x compilation time decrease after applying one-line patch, no downsides.
Just to clarify, Alex, since your idea of "next release" might be different than those who aren't following Clojure 1.7 alphas/betas/etc. carefully. Do you mean "hoping to look at it for Clojure 1.8, since Clojure 1.7 is to the point of primarily fixing critical bugs and regressions" ? Andy On Fri, May 8, 2015 at 12:07 PM, Alex Miller wrote: > I'm hoping to look at it for the next release. > > Alex > > -- > You received this message because you are subscribed to the Google > Groups "Clojure" group. > To post to this group, send email to clojure@googlegroups.com > Note that posts from new members are moderated - please be patient with > your first post. > To unsubscribe from this group, send email to > clojure+unsubscr...@googlegroups.com > For more options, visit this group at > http://groups.google.com/group/clojure?hl=en > --- > You received this message because you are subscribed to the Google Groups > "Clojure" group. > To unsubscribe from this group and stop receiving emails from it, send an > email to clojure+unsubscr...@googlegroups.com. > For more options, visit https://groups.google.com/d/optout. > -- You received this message because you are subscribed to the Google Groups "Clojure" group. To post to this group, send email to clojure@googlegroups.com Note that posts from new members are moderated - please be patient with your first post. To unsubscribe from this group, send email to clojure+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/clojure?hl=en --- You received this message because you are subscribed to the Google Groups "Clojure" group. To unsubscribe from this group and stop receiving emails from it, send an email to clojure+unsubscr...@googlegroups.com. For more options, visit https://groups.google.com/d/optout.
CLJ-703 - 10x compilation time decrease after applying one-line patch, no downsides.
I'm hoping to look at it for the next release. Alex -- You received this message because you are subscribed to the Google Groups "Clojure" group. To post to this group, send email to clojure@googlegroups.com Note that posts from new members are moderated - please be patient with your first post. To unsubscribe from this group, send email to clojure+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/clojure?hl=en --- You received this message because you are subscribed to the Google Groups "Clojure" group. To unsubscribe from this group and stop receiving emails from it, send an email to clojure+unsubscr...@googlegroups.com. For more options, visit https://groups.google.com/d/optout.
CLJ-703 - 10x compilation time decrease after applying one-line patch, no downsides.
Hi all, This issue has been around for a while without much activity, although a very simple fix is already there: http://dev.clojure.org/jira/browse/CLJ-703 We have a pretty big Clojure project that we compile on machines with slow hard-drives, on CentOS 6. We noticed that compilation had become extremely slow, however CPU usage was always below 10%. Simply removing the line "cfs.getFD().sync();" inside "src/jvm/clojure/lang/Compiler.java" dramatically improved the compilation time (Something like 10x, although at this level of increase, measurement almost doesn't make sense anymore :)). We've been using the suggested fix (remove-sync-only.patch) for a few weeks and haven't observed any issues. It seems we should really figure out whether this line is actually necessary, or if it was added there "just in case". Even if some situations require it, the speed bump is so huge that we might want some special-case logic. Martin -- You received this message because you are subscribed to the Google Groups "Clojure" group. To post to this group, send email to clojure@googlegroups.com Note that posts from new members are moderated - please be patient with your first post. To unsubscribe from this group, send email to clojure+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/clojure?hl=en --- You received this message because you are subscribed to the Google Groups "Clojure" group. To unsubscribe from this group and stop receiving emails from it, send an email to clojure+unsubscr...@googlegroups.com. For more options, visit https://groups.google.com/d/optout.