wow, the difference is crazy. If you can provide the dependency tree that would be nice.
I was looking at the code for the PreDex task and realize we don't parallelize the pre-dexing when there are more than one library which I thought we did. I think dx is making use of all the cores but I'm not sure how much it would help for this case. On Wed, Jan 8, 2014 at 1:27 AM, Jake Wharton <jakewhar...@gmail.com> wrote: > All of these benchmarks are done with nothing in gradle.properties. > > With pre-dexing off, ./gradlew clean assemble --profile --offline: > > :one:one-app 1m13.07s > :one:one-app:dexDogfood 17.612s > :one:one-app:dexRelease 17.261s > :one:one-app:dexDebug 15.849s > > :two:two-v1 1m1.84s > :two:two-v1:dexDogfood 15.553s > :two:two-v1:dexDebug 15.066s > :two:two-v1:dexRelease 14.183s > > :three:three-app 36.907s > :three:three-app:dexDebug 9.278s > :three:three-app:dexDogfood 8.848s > :three:three-app:dexRelease 8.654s > > With pre-dexing on, ./gradlew clean assemble --profile --offline: > > :one:one-app 3m14.97s > :one:one-app:preDexDebug 58.122s > :one:one-app:preDexRelease 51.320s > :one:one-app:preDexDogfood 50.261s > :one:one-app:dexDebug 5.591s > :one:one-app:dexDogfood 4.257s > :one:one-app:dexRelease 4.037s > > :two:two-v1 2m29.22s > :two:two-v1:preDexDogfood 36.434s > :two:two-v1:preDexDebug 36.408s > :two:two-v1:preDexRelease 35.988s > :two:two-v1:dexDogfood 14.891s > :two:two-v1:dexDebug 4.893s > :two:two-v1:dexRelease 4.003s > > :three:three-app 2m14.54s > :three:three-app:preDexRelease 44.191s > :three:three-app:preDexDebug 35.970s > :three:three-app:preDexDogfood 33.390s > :three:three-app:dexDebug 3.956s > :three:three-app:dexRelease 3.575s > :three:three-app:dexDogfood 3.451s > > $ java -version > java version "1.8.0-ea" > Java(TM) SE Runtime Environment (build 1.8.0-ea-b116) > Java HotSpot(TM) 64-Bit Server VM (build 25.0-b58, mixed mode) > > $ ./gradlew --version > > ------------------------------------------------------------ > Gradle 1.9 > ------------------------------------------------------------ > > Build time: 2013-11-19 08:20:02 UTC > Build number: none > Revision: 7970ec3503b4f5767ee1c1c69f8b4186c4763e3d > > Groovy: 1.8.6 > Ant: Apache Ant(TM) version 1.9.2 compiled on July 8 2013 > Ivy: 2.2.0 > JVM: 1.8.0-ea (Oracle Corporation 25.0-b58) > OS: Mac OS X 10.8.5 x86_64 > > Let me know if there's anything else I can provide (e.g., dependency tree?) > > > > --- > Jake Wharton > http://about.me/jakewharton > > > On Tue, Jan 7, 2014 at 9:39 PM, Xavier Ducrohet <x...@android.com> wrote: > >> BTW I'm interested in the time it takes to do predex + dex vs dex only >> (no pre-dexing) for clean builds >> >> It's worth it in incremental builds, but I'm suprised clean builds are so >> much longer. I'd like to send some stats to the VM team (who owns dx) >> >> thanks. >> >> >> >> On Tue, Jan 7, 2014 at 11:36 PM, Xavier Ducrohet <x...@android.com> wrote: >> >>> android { >>> dexOptions { >>> preDexLibraries = false >>> } >>> } >>> >>> >>> On Tue, Jan 7, 2014 at 3:13 PM, Jake Wharton <jakewhar...@gmail.com>wrote: >>> >>>> Well it's definitely pre-dexing. Can this be easily disabled for CI >>>> builds? >>>> >>>> :app1:app1-app 3m14.13s >>>> :app1:app1-app:preDexDebug 1m5.23s >>>> :app1:app1-app:preDexDogfood 51.836s >>>> :app1:app1-app:preDexRelease 41.499s >>>> >>>> 81% of time spent pre-dexing >>>> >>>> >>>> :app2:app2-v1 2m25.18s >>>> :app2:app2-v1:preDexRelease 39.274s >>>> :app2:app2-v1:preDexDogfood 38.637s >>>> :app2:app2-v1:preDexDebug 36.630s >>>> >>>> 78% of time spent pre-dexing >>>> >>>> >>>> :app3:app3-app 2m18.30s >>>> :app3:app3-app:preDexRelease 45.504s >>>> :app3:app3-app:preDexDebug 36.681s >>>> :app3:app3-app:preDexDogfood 34.769s >>>> >>>> 84% of time spent pre-dexing >>>> >>>> >>>> --- >>>> Jake Wharton >>>> http://about.me/jakewharton >>>> >>>> >>>> On Mon, Jan 6, 2014 at 10:37 PM, Jake Wharton <jakewhar...@gmail.com>wrote: >>>> >>>>> They were clean builds, yes. I'll dig deeper tomorrow. It was the end >>>>> of the work day for me when we started the comparisons. >>>>> >>>>> >>>>> --- >>>>> Jake Wharton >>>>> http://about.me/jakewharton >>>>> >>>>> >>>>> On Mon, Jan 6, 2014 at 6:07 PM, Xavier Ducrohet <x...@android.com>wrote: >>>>> >>>>>> incremental dexing has been disabled. There are some issues with it. >>>>>> We have implemented pre-dexing though, but I'm not sure how the >>>>>> performance >>>>>> is impacted. >>>>>> >>>>>> Are those time for clean builds? (which should not be impacted by the >>>>>> removal of the incremental dex). >>>>>> >>>>>> >>>>>> On Mon, Jan 6, 2014 at 4:50 PM, Jake Wharton >>>>>> <jakewhar...@gmail.com>wrote: >>>>>> >>>>>>> I forgot to ask a question... Is anyone else seeing anything like >>>>>>> this? >>>>>>> >>>>>>> I see this on small pet projects too. I can perceive tasks taking >>>>>>> 10x longer than when on 0.6.x. Usually those towards the latter end of >>>>>>> the >>>>>>> processing (preDex, dex, assemble). >>>>>>> >>>>>>> Still digging for more info to pinpoint what is happening. >>>>>>> >>>>>>> >>>>>>> --- >>>>>>> Jake Wharton >>>>>>> http://about.me/jakewharton >>>>>>> >>>>>>> >>>>>>> On Mon, Jan 6, 2014 at 4:44 PM, Jake Wharton >>>>>>> <jakewhar...@gmail.com>wrote: >>>>>>> >>>>>>>> Our project builds have experienced a serious regression in terms >>>>>>>> of performance upon upgrading to 0.7.x. >>>>>>>> >>>>>>>> Without --parallel: >>>>>>>> >>>>>>>> - 0.6.x: 4m 21s >>>>>>>> - 0.7.x: 9m 04s >>>>>>>> >>>>>>>> With --parallel: >>>>>>>> >>>>>>>> - 0.6.x: 3m 36s >>>>>>>> - 0.7.x: 6m 58s >>>>>>>> >>>>>>>> This project has 3 apps (each with three build types, no flavors), >>>>>>>> 4 library modules, and 3 java modules. >>>>>>>> >>>>>>>> --- >>>>>>>> Jake Wharton >>>>>>>> http://about.me/jakewharton >>>>>>>> >>>>>>> >>>>>>> -- >>>>>>> You received this message because you are subscribed to the Google >>>>>>> Groups "adt-dev" group. >>>>>>> To unsubscribe from this group and stop receiving emails from it, >>>>>>> send an email to adt-dev+unsubscr...@googlegroups.com. >>>>>>> For more options, visit https://groups.google.com/groups/opt_out. >>>>>>> >>>>>> >>>>>> >>>>>> >>>>>> -- >>>>>> Xavier Ducrohet >>>>>> Android SDK Tech Lead >>>>>> Google Inc. >>>>>> http://developer.android.com | http://tools.android.com >>>>>> >>>>>> Please do not send me questions directly. Thanks! >>>>>> >>>>>> -- >>>>>> You received this message because you are subscribed to the Google >>>>>> Groups "adt-dev" group. >>>>>> To unsubscribe from this group and stop receiving emails from it, >>>>>> send an email to adt-dev+unsubscr...@googlegroups.com. >>>>>> For more options, visit https://groups.google.com/groups/opt_out. >>>>>> >>>>> >>>>> >>>> -- >>>> You received this message because you are subscribed to the Google >>>> Groups "adt-dev" group. >>>> To unsubscribe from this group and stop receiving emails from it, send >>>> an email to adt-dev+unsubscr...@googlegroups.com. >>>> For more options, visit https://groups.google.com/groups/opt_out. >>>> >>> >>> >>> >>> -- >>> Xavier Ducrohet >>> Android SDK Tech Lead >>> Google Inc. >>> http://developer.android.com | http://tools.android.com >>> >>> Please do not send me questions directly. Thanks! >>> >> >> >> >> -- >> Xavier Ducrohet >> Android SDK Tech Lead >> Google Inc. >> http://developer.android.com | http://tools.android.com >> >> Please do not send me questions directly. Thanks! >> >> -- >> You received this message because you are subscribed to the Google Groups >> "adt-dev" group. >> To unsubscribe from this group and stop receiving emails from it, send an >> email to adt-dev+unsubscr...@googlegroups.com. >> For more options, visit https://groups.google.com/groups/opt_out. >> > > -- > You received this message because you are subscribed to the Google Groups > "adt-dev" group. > To unsubscribe from this group and stop receiving emails from it, send an > email to adt-dev+unsubscr...@googlegroups.com. > For more options, visit https://groups.google.com/groups/opt_out. > -- Xavier Ducrohet Android SDK Tech Lead Google Inc. http://developer.android.com | http://tools.android.com Please do not send me questions directly. Thanks! -- You received this message because you are subscribed to the Google Groups "adt-dev" group. To unsubscribe from this group and stop receiving emails from it, send an email to adt-dev+unsubscr...@googlegroups.com. For more options, visit https://groups.google.com/groups/opt_out.
