Guest fonz93 Posted June 14, 2014 Report Posted June 14, 2014 (edited) It disables GPU acceleration? Only on related parts of apps locked in memory, but it seems that systemUI is still smooth Edited June 14, 2014 by fonz93
Guest ZolaIII Posted June 14, 2014 Report Posted June 14, 2014 Try play Dead Trigger 2 or any game like that for 10-15 minutes and it will crash I play it I am a highest lv. Never had a UI crashed! Played even marathons of several h. I had app crashes when swap whose full on earlier kernels bat that's it. I use it on every test kernel to determine behavior under extensive heavy loads.
Guest Posted June 14, 2014 Report Posted June 14, 2014 In the file /cust/weu/xml/hw_defaults.xml Huawei did following settings: <string apps_set_oomadj="com.android.mms;com.android.contacts;android.process.acore;com.huawei.android.launcher"/> <integer oomadj.com.android.mms="3"/> <integer oomadj.com.android.contacts="3"/> <integer oomadj.android.process.acore="3"/> <integer oomadj.com.huawei.android.launcher="1"/> That's why I disabled it on the ROM side and made the kernel handle it. It was less work for me to undo Huawei's mess that way. How about putting "system.ui" to the oom.adj section above?
Guest H3ROS Posted June 14, 2014 Report Posted June 14, 2014 (edited) I'm not sure. After I discovered that it disabled GPU rendering for some apps I left it that way and worked on other stuff. My guess would be that you'd need to add <integer oomadj.com.android.systemui="-12"/> And I'm suggesting -12 because of this: printWriter.print(" PERSISTENT_PROC_ADJ: "); printWriter.println(-12); if ((9 & applicationInfo.flags) == 9 && (this.isPersistentCustomized(processRecord.processName) || this.isInAppToSetPersAdjArrays(processRecord.processName))) { processRecord.persistent = true; processRecord.maxAdj = -12; } Obviously that might not work because it's not actually set to persistent. Edited June 14, 2014 by H3ROS
Guest thetonyxx Posted June 14, 2014 Report Posted June 14, 2014 (edited) new record on Antutu whit 13/06 release 7853 points :D Edited June 14, 2014 by thetonyxx
Guest Posted June 14, 2014 Report Posted June 14, 2014 My ROM suxx. I never reach more than 7500-7600
Guest thetonyxx Posted June 14, 2014 Report Posted June 14, 2014 (edited) My ROM suxx. I never reach more than 7500-7600 yes, I also 7500-7600 most of the time, but sometimes 7700 and now 7853 records i use SlimStock Final + Zram 128mb Edited June 14, 2014 by thetonyxx
Guest Posted June 15, 2014 Report Posted June 15, 2014 (edited) current state of progress: http://www53.zippyshare.com/v/80470422/file.html add some filesystem patches tryed to fix the msm-cpufrequency hotplug lock lowered some board specific latencies increased the number of zram devices to 2 (for compatibility reasons towards some apps which expect 2 drives being present) RE-ENABLED support for G330 radio driver update will come next Edited June 16, 2014 by Guest
Guest Posted June 15, 2014 Report Posted June 15, 2014 gpu oc plis 320 Mhz should be enough I guess. I'll increase it tomorrow. Why don't you smile on your UserPIC. You look so sad
Guest jor1196 Posted June 15, 2014 Report Posted June 15, 2014 320 Mhz should be enough I guess. I'll increase it tomorrow. Why don't you smile on your UserPIC. You look so sad if formed with 320MHz me, you could do the test a higher Oc? Hahaha if the apology will change someday
Guest Domino2115 Posted June 15, 2014 Report Posted June 15, 2014 Hi, Its seems like that with lastest experimental wifi doesnt work, I'll try it again and let you know
Guest porozex Posted June 15, 2014 Report Posted June 15, 2014 Hi, Its seems like that with lastest experimental wifi doesnt work, I'll try it again and let you know yes i can confirm that. it dosent work
Guest dinotre Posted June 15, 2014 Report Posted June 15, 2014 I'll start an attempt whether an clean 3.4.93 linux kernel is able to run on our device and only merging necessary things from CodeAurora which are MSM specific. any news about this? :)
Guest Posted June 15, 2014 Report Posted June 15, 2014 (edited) Didn't finish this little experiment, yet. Current build: http://www53.zippyshare.com/v/80470422/file.html (updated FM Radio driver) No mentionable changes (except some cache handling optimizations) EDIT: ok...maybe there is one change worth mentioning: I tryed implementing asynchronous fsync, which faux123 extracted from a HTC source code. If it works well I'd prefer it over any "dynamic fsync feature". Edited June 16, 2014 by Guest
Guest juanpablocastillo Posted June 16, 2014 Report Posted June 16, 2014 @moddingg33k First of all, thanks for your work. You are amazing. Second: Sorry if I've been annoying these days. And third: If you have time, take a look in the S Plus changes of the K^Kernel created by Christopher83. In that kernel, there is a change in the memory mapping ION. You can increase overall system RAM. Use a dynamic memory allocation vía CMA(Contigous Memory Allocator)
Guest juanpablocastillo Posted June 16, 2014 Report Posted June 16, 2014 Four: Sorry for my poor english. A mix of my native english and Google Translate. : P
Guest Posted June 16, 2014 Report Posted June 16, 2014 (edited) Updated build: http://www53.zippyshare.com/v/80470422/file.html (FM Radio driver updated) I tryed to smooth out the jittery frequency scaling and hope I didn't decrease performance on this way. First stress-tests looked fine so far. Please help finding issues by playing some games and doing daily stuff you use to do usually. And report back what you experience. Edited June 16, 2014 by Guest
Guest Posted June 16, 2014 Report Posted June 16, 2014 And third: If you have time, take a look in the S Plus changes of the K^Kernel created by Christopher83. In that kernel, there is a change in the memory mapping ION. You can increase overall system RAM. Use a dynamic memory allocation vía CMA(Contigous Memory Allocator) Thx for the info. This might however require some further patching before I might be able to adopt it to our kernel.
Guest Romagnolo1973 Posted June 16, 2014 Report Posted June 16, 2014 Didn't finish this little experiment, yet. Current build: http://www38.zippyshare.com/v/85247627/file.html No mentionable changes (except some cache handling optimizations) EDIT: ok...maybe there is one change worth mentioning: I tryed implementing asynchronous fsync, which faux123 extracted from a HTC source code. If it works well I'd prefer it over any "dynamic fsync feature". good to see that HTC Desire 300 kernel can be usefull in some circumstance
Guest Domino2115 Posted June 16, 2014 Report Posted June 16, 2014 Im using it for 30 minutes, everything runs smooth really nice build, thank you so much :)
Guest squal Posted June 16, 2014 Report Posted June 16, 2014 Im using it for 30 minutes, everything runs smooth really nice build, thank you so much :) how's the game performance?
Guest Domino2115 Posted June 16, 2014 Report Posted June 16, 2014 how's the game performance? sorry I dont play games.
Guest Posted June 16, 2014 Report Posted June 16, 2014 Errrrm....if you prefer an GCC 4.9.1 build take this one: http://www1.zippyshare.com/v/77559056/file.html I suppose that the generated code of this toolchain-build is cleaner and maybe more optimized compared to my 4.8.3-2014-04 linaro one.
Guest squal Posted June 16, 2014 Report Posted June 16, 2014 (edited) Updated build: http://www53.zippyshare.com/v/80470422/file.html (FM Radio driver updated) I tryed to smooth out the jittery frequency scaling and hope I didn't decrease performance on this way. First stress-tests looked fine so far. Please help finding issues by playing some games and doing daily stuff you use to do usually. And report back what you experience.great build...smooth and fast..no lags in subway surfer...the radio is no problem...:) tnxs moddingg33k even in opening the app like facebook...its fast... Edited June 16, 2014 by squal
Recommended Posts
Please sign in to comment
You will be able to leave a comment after signing in
Sign In Now