Degraded 3D Performance with Froyo/Cog 2.2? - Captivate General

Hey everyone,
I was running Quadrant after flashing Cog 2.2/Froyo with OCLF. Everything seems to go faster but the 2D and 3D performance - I can definitely tell that the framerate is lower in those two tests than before. Am I crazy?...

Lencias said:
Hey everyone,
I was running Quadrant after flashing Cog 2.2/Froyo with OCLF. Everything seems to go faster but the 2D and 3D performance - I can definitely tell that the framerate is lower in those two tests than before. Am I crazy?...
Click to expand...
Click to collapse
I noticed that framerates were lower as well, but only about 4 or 5 FPS. A drop this small isn't really all that significant, and 3D performance at 40 FPS rather than 44 FPS is still much better than a phone like the EVO which runs at like 18 FPS, so I wouldn't really worry that much

Yeah, I noticed it too. I'm also noticing other slowdowns, like a lag on the app drawer screen between my finger swipes and corresponding animations on the screen. If I swipe left or right, the actual swiping animation happens a split second later (noticeable lag). Something tells me this isn't the release build of this rom.
I'm also noticing that the battery dies much quicker with this new rom. I'd get about 2 days of use out of JH7 with the amount I use the phone, with JI6 I'm lucky if I get 8-10 hours.

Actually the benchmark result is quite amazing id say
It scores 1900+ in quadrant standarf
Sent from my SAMSUNG-SGH-I897 using XDA App

Lencias said:
Hey everyone,
I was running Quadrant after flashing Cog 2.2/Froyo with OCLF. Everything seems to go faster but the 2D and 3D performance - I can definitely tell that the framerate is lower in those two tests than before. Am I crazy?...
Click to expand...
Click to collapse
You are crazy... Quadrant is for SNAPDRAGON PHONES... STOP USING IT!!!

Use Neocore, that is a better FPS than quadrant. I still get 55.5/6 with Neocore so 3d performance is not degraded, at least not on my phone.

I'd expect some continued optimization of Froyo 2.2. Remember how laggy JF6 was? Then it got way better with JH2/3/7. I expect progress like that with the Froyo builds as well. With OCLF, JI6 is actually fairly impressive speed wise.

kennethpenn said:
I'd expect some continued optimization of Froyo 2.2. Remember how laggy JF6 was? Then it got way better with JH2/3/7. I expect progress like that with the Froyo builds as well. With OCLF, JI6 is actually fairly impressive speed wise.
Click to expand...
Click to collapse
Hmm.. did you do anything after you flashed, like cleared dalvik cache and whatnot? Mine is actually running slower (UI-wise) than JF6

Its funny you posted this because I noticed it as well today when I ran Quadrant just to see what it would get. I think Quadrant is a horrible tool for benchmarking now since it can be so easily manipulated, but I did notice the FPS drop myself.

Gles 2/1 is perfectly fine, i scored 46fps on nenamark, so no need to worry about 3d performance, i do admit 2d does stutter a little tho.

designgears said:
You are crazy... Quadrant is for SNAPDRAGON PHONES... STOP USING IT!!!
Click to expand...
Click to collapse
Take the man's advice! Quadrant is an awful way to measure performance on the Captivate. Quadrant is biased towards Snapdragon processors, and isn't all that accurate.
Sent from my Cognition 2.2 powered Captivate using XDA app.

It's running fine for me.
Sent from my SAMSUNG-SGH-I897 using XDA App

Actually using NenaMark1 which is a benchmark app specifically for OpenGL2.0 I get slightly higher scores than on 2.1.

designgears said:
You are crazy... Quadrant is for SNAPDRAGON PHONES... STOP USING IT!!!
Click to expand...
Click to collapse
dg, may as well save your breath. No matter what is said or done, there will always be people who complain that X is slower than when it was Y and be sure that they are right (and will argue they are right to their dying breath).
Synthetic benchmarks are b***it on a multitasking OS anyway, at any given time any thread can tie up a resource the synthetic benchmark needs or it may not. The only way you could truly benchmark something like this while still allowing multitasking would be with a long range sample like a few hours, the 30 -70 seconds Quadrant runs is a worthless snapshot in time.
They all think they are getting an answer but all they are really doing is using a magnifying glass on a Seurat and wondering why they don't get the big picture. Of course the performance issue couldn't be one of the apps they have installed or that has updated since the last time they benchmarked.
I have 206 apps installed on my captivate with even a ridiculous amount like that and more stuff than the average running I can still get a quadrant score of over 950, its a snapshot in the middle not an effective way to judge a film.

For everyone whinging about reduced "3D" performance, is the actual handset slower in real, day to day use?
This is stock 2.1 Firmware vs stock 2.2 Firmware.
Not Cognition/custom rom with lagfix vs Stock 2.2.

As much as I don't like benchmarks, in this they are only confirming the slow downs I'm seeing in real world use on the leaked froyo rom. (Not cognition)
Sent from my SAMSUNG-SGH-I897 using XDA App

The first time I flashed 2.2 I saw some lag in the form of occasional black screens and delayed responses, even after installing OCLF.
I used Odin One-Click to flash back to JF6 and did a master clear before installing Cognition 2.2. Immediately applied OCLF before doing anything else. Not seeing the black screens and delayed responses this time. Some are suggesting it's the master clear that helps.
As for 3D performance, animations and scrolling between homescreens are a lot smoother than they were in 2.1. Games run awesome too. The only two places I notice some skippyness are in the app drawer (not often though, and very minor) and in the browser (this is due to a known memory leak bug).

1randomtask said:
For everyone whinging about reduced "3D" performance, is the actual handset slower in real, day to day use?
This is stock 2.1 Firmware vs stock 2.2 Firmware.
Not Cognition/custom rom with lagfix vs Stock 2.2.
Click to expand...
Click to collapse
Yeah, I actually never bothered benchmarking, but I'm experiencing noticeable lag between swipes and corresponding animations in the app drawer and launcher.

modest_mandroid said:
Yeah, I actually never bothered benchmarking, but I'm experiencing noticeable lag between swipes and corresponding animations in the app drawer and launcher.
Click to expand...
Click to collapse
I have never experienced that on 2.2, even before applying a lag fix. I suggest One-Clicking back to JF6, doing a master clear, and then installing 2.2 again.

No drop vs stock, the overclock kernel gave the extra fps. it also overclocked the gpu by 11%. 4-5 fps out of 40-45 is right in that area.
Edit :.i did a master clear so maybe that's why I don't see a problem
Not all benchmarks will work because our phones are capped at 56 fps. A benchmark needs to bog the gpu down below that for the whole test to get an accurate measurement. Quadrant does that for only party of the test. Many benchmarks can't do that at all.
Quadrant is a decent bench mark but our phones will not see the score increase of a snapdragon with froyo. linpack shows greatly inflated scores on snapdragon froyo devices. Greene computing goes over this on there website.
http://www.greenecomputing.com/
the full version of quadrant shows that the hummingbird and the processor in the droid 2/x mop the floor with the snapdragon in all areas but processing speed. The hummingbird having an edge over the droid x particularly in the 3d area but losing overall because of the rfs file system lag Which is apparently is weighed heavily and the reason the lagfix blows the score out of proportion and the different lagfixes have big gaps in the score while real world performance is not even noticeable.
http://slideme.org/application/quadrant-advanced
Composite scores don't mean anything and that is my gripe with quadrant, there is no way to decide how to weigh file system benchmarks against processor benchmarks against 3d, each area must be compared separately, and the user should be able to decide what that means.
Edit: linpack shows a huge increase over stock. 13.77 vs the 9.61 I got with voodoo and overclock. But still not at the level of a snapdragon. Wait till the next generating of arm processors come out, if you think a snapdragon/hummingbird hybrid would be awesome that's nothing compared to what's in the works. Higher clock speeds and multi-core cpu's aren't far away.
Sent from my SAMSUNG-SGH-I897 using Tapatalk

Related

Quadrant score

I just did the lag fix mentioned here: http://www.xda-developers.com/android/one-click-lag-fix-for-galaxy-s-i9000/
on my captivate and got a score of 2185... It's my 1st time benchmarking the phone. Is that a good score?
yeah but did your actual perfomance change?
I got 2190 on a reboot after I manually did the fix. Still have all the AT&T stuff, root and side loaded but other than that stock.
The phone does feel a little faster, its not twice as fast as it was but scrolling around and launching programs does seem faster. Placebo effect? Maybe.
Nice
I tried it as well and I can say I have definitely noticed the speed bump. Well worth the three minutes it took me to read and install.
It doesn't really feel that much quicker, more like smoother since the lag is pretty much gone.
SiL3nTKiLL said:
yeah but did your actual perfomance change?
Click to expand...
Click to collapse
yeah it changes all right. things load much faster. webpages, apps etc... you will be amazed
I ran the fix too. The phone finally feels high end. Every app is now responsive. Xda app flies now. The market loads the"downloads" tab in 1 second instead of 7 dolphin is as fast as the stock browser. It's really great.
That said, the quadrant score is bogus. The lag fix sets up a loopback for io writes so quadrant produces a completely bogus and inflated ui score.

Just showin off a bit : )

So after reading nearly 5 hours and spending my time in the wee hours of morning, I finally did all the "stable" mods for the phone... If you haven't been reading, make sure you guys check out the stuff in the development forum.
After all modifications, I was able to get 2701 points in quadrant benchmark. What mods did I do?
-i9000 eclair flash (JM5)
http://forum.xda-developers.com/showthread.php?t=734871
-Alternative mimocans lag fix
(one click installer http://forum.xda-developers.com/showthread.php?t=749495)
-One click root (googled it for i9000)
-Overclock kernel 1.0Ghz to 1.2Ghz
(http://forum.xda-developers.com/showthread.php?t=746343)
This stuff really does help out your phone folks. Bench it now with quadrant, then take a peak at the other stuff and make magic happen. If anyone needs any additional help setting up their captivate, I'm more than happy to help.
I agree those fixes help speed a lot. But the quadrant score is meaningless. the speed hack creates an io loopback. The loopback just tells quadrant what it wants to hear.
Does your BT work on the european ROM. For me all people hear is a gargeled mess on there end?
Sent from my SAMSUNG-SGH-I897 using Tapatalk
Can you post the results of these tests:
Neocore
Linpack
CPU Benchmark
I keep hearing about this quadrant, does it actually improve real world performance? Or is just for the sake of scores?
jhego said:
I keep hearing about this quadrant, does it actually improve real world performance? Or is just for the sake of scores?
Click to expand...
Click to collapse
Quadrant is just benchmarking software that takes the cpu, gpu and memory read/write speeds into account. It runs a series of tests and spits a score number out at the end, so you can compare your device to others (like comparing boner sizes, but less gay).
It doesn't actually do anything to speed up the device though.
modest_mandroid said:
Quadrant is just benchmarking software that takes the cpu, gpu and memory read/write speeds into account. It runs a series of tests and spits a score number out at the end, so you can compare your device to others (like comparing boner sizes, but less gay).
It doesn't actually do anything to speed up the device though.
Click to expand...
Click to collapse
This is correct and to be honest there is way to much weight in the I/O tests. That is the only reason that the stock Droid X bests the Gal X. It has more weight than cpu and gpu so you can't put to much into those scores. They really don't mean anything more than bragging rights. What I am interested in is real world usage.
Real world use with the hack provides amazing speed gains opening and switching apps. Io heavy apps are very much improved while open too. It's finally as fast as the iphone.
Whats your battery life like after the overclock?
The score ended up getting lower and lower every time i used quadrant. 2701 is the highest I was able to get so far, but that's with a fresh install of the rom and all the stuff before I started loading on apps. Everytime I ran the benchmark, I of course killed the apps beforehand.
The battery life is the same- to be honest. This is me comparing a rooted stock ROM to the somewhat fresh install of the eclair i9000. The phone is very snappy. I came from an iPhone 4 and one of the biggest eye sores to me was the less-fluidness of changing programs, response to buttons (homescreen-back button) and pinch to zoom. After all these changes, it's a whole different story. Browsing is very appealing, especially since pinch to zoom isn't jagged or slow. The smoothness of this functionality is on par to an iPhone. And there is no waiting when I press the home button or back button.
True, maybe these numbers aren't considerably accurate (as far as the lag fix and EXT2) but at least it shows raw computing capability in it's current state... meaning, the usage of a virtual EXT2. Never the less, the phone is still all around faster, even if it isn't exactly the proper way of going about it.
The only problem I've seen so far is that it likes to randomly shut off. Won't respond to anything unless if I soft reset it. I haven't really found what causes it, since the consistency of it happening goes about in a non set pattern.
I didn't see any real world increase .. so I reverted back in about 4 hours.... I'd rather have the memory than a number that don't transfer to real world speeds...

Laggy HTC Sensation

Has anyone else had a laggy sensation, was fast when I first got it. Installed apps etc and after about one month became gittery/laggy navigating around or in apps. However when testing with ANTUTU, I got a bench mark score of 4562 with stock ROM and all these apps, running the tests I got results anything from 3200 to 4200 regularly, even running one after the other.
After rooting became available I installed Virtuous Sensation v.1.0.3 with V5 Kernel patch http://forum.xda-developers.com/showthread.php?t=1101249, it was still laggy after installing my apps back. Using ANTUTU, I got one bench mark of 6009, which is pretty good for a sensation, most bench marks about 5600.
I installed 1.1.0 when it came out and it wasn’t much better, bench marks right back to stock between 3200 to 4200.
When testing using Quadrant, my results are at just under 2000, not a whole lot better than the Nexus One 2.2+ which is pathetic, my HTC Desire overclocked to 1.2Ghz on 2.3.4 gets bench marks of around 1700 with a single processor.
It is almost like the 2nd asynchronous CPU is not kicking in?
I must say now with 1.1.2 the lag has all but disappeared but the bench marks are still pretty low I would think.
Run another test without any apps installed except the default install and bench mark tools and the results have been the same.
Has anyone got some tips, or am I doing something wrong?
try moving all of your apps to the sd card and clear cache using cachemate
Android benchmarking is a pretty hit-and-miss affair, and is shockingly over-rated.
As long as you got it smooth for everyday use, 'good' benches are irrelevant for everything but ePeen
Agreed! Take the bench for what it is...day to day device performance should be the deciding factor
juzz86 said:
Android benchmarking is a pretty hit-and-miss affair, and is shockingly over-rated.
As long as you got it smooth for everyday use, 'good' benches are irrelevant for everything but ePeen
Click to expand...
Click to collapse
Thanks so much everyone for your responses, yes I understand bench mark is hit and miss, but it was the only way I could get some form of validation.
The problem was that it was getting laggy for everyday use, which is why I tried the other things to try to understand the issue better.
juzz86 said:
Android benchmarking is a pretty hit-and-miss affair, and is shockingly over-rated.
As long as you got it smooth for everyday use, 'good' benches are irrelevant for everything but ePeen
Click to expand...
Click to collapse
couldnt agree more
danw_oz said:
Thanks so much everyone for your responses, yes I understand bench mark is hit and miss, but it was the only way I could get some form of validation.
The problem was that it was getting laggy for everyday use, which is why I tried the other things to try to understand the issue better.
Click to expand...
Click to collapse
I would do what sweensour says. its the most sensible solution. sounds like you packing the 800 megs with apps. use the external for that. also make sure to copy any back ups to your computer and remove them from your device. back ups take a lot of room

AOSP Lockscreen Mod for GingerClone

I can't take credit for any of the work here, I just modded the framework-res.apk and twframework.apk so that I could use the standard Gingerbread sliding tab lockscreen on my favorite ROM GingerClone. Thanks goes out to the Platypus Dev Team and everyone else involved in making this. The KJ2 Locks Mod does in fact work on GingerClone AS WELL AS OTHER ROMS, but it replaces a large portion of the theming elements of the UI with the ones from Glorious Overdose. I fixed this by modding the frameworks.
Call me crazy, but I'm a big fan of the standard Gingerbread interface, even on my old phone running CM7, I used the standard theme.
***BUGS***
-- Landscape tabs don't match portrait tabs
The only thing I've noticed so far is that the landscape lockscreen tabs do not match the portrait lockscreen tabs, but since the phone screen needs to be OPEN and locked for you to ever see this screen it shouldn't be a big deal. I've torn through the res files and can't seem to find the landscape lock tabs, so unless they're not images and are actually hard coded, I can't figure out where they're at. Maybe the dev could point me in the right direction?
Once again, I can't take credit for ANY of this. I just modded the work created by other people.
***DOWNLOAD***
https://rapidshare.com/files/2888540474/GingerClone_AOSP_Lockscreen_Mod.zip
***SCREENSHOTS***
Uploaded with ImageShack.us
Uploaded with ImageShack.us
Uploaded with ImageShack.us
ENJOY
Thanks man! So I'm assuming we have to flash this? And how are you getting a 2300+ score on quadrant? I tried tegrak and even at 1.1 it was freezing up my phone and doing the 1vib followed by 3vib when it didn't freeze it up.
JonathanBarca10 said:
Thanks man! So I'm assuming we have to flash this? And how are you getting a 2300+ score on quadrant? I tried tegrak and even at 1.1 it was freezing up my phone and doing the 1vib followed by 3vib when it didn't freeze it up.
Click to expand...
Click to collapse
try 1.2, that normally is pretty solid.
See I tried 1.2 but it froze even faster, and that's why I tried to downgrade to the 1.1
My q score suprised me too! First of all, it's the ROM. GingerClone is the only fully de-touchwized ROM available currently and it amazes me how well this device performs on a close to vanilla Android OS (especially for Froyo! I can only imagine that this phone would FLY with Gingerbread or ICS)
I OC'd using Tegrak to 1.125Ghz. Any higher and the system locks up on my phone regardless of ROM or kernel, but if I buy the full version of the app, I'm pretty sure I could get at least 1.4Ghz with overvolt. In the past I've noticed that especially refurbished phones can have issues with overclocking. My first LG Optimus could EASILY overclock up to 800+ Mhz stable, but I bricked it and the replacement they sent me was a refurb that couldn't handle more than 730Mhz and be stable.
But the real magic lies in using the latest version of V6 supercharger. I set mine to Aggressive 2. I had to run Quadrant a couple of times because I thought it was a glitch or a fluke the first time it broke 2000.
JonathanBarca10 said:
Thanks man! So I'm assuming we have to flash this? And how are you getting a 2300+ score on quadrant? I tried tegrak and even at 1.1 it was freezing up my phone and doing the 1vib followed by 3vib when it didn't freeze it up.
Click to expand...
Click to collapse
yeah, I forgot to mention that (figured it was self explanatory). You flash it in recovery.
I just flashed GingerCloneV2, I have not overclocked or supercharged, and my quadrant score was 2025. I ran it twice to be sure it wasnt a fluke. My phone has never even approached 2000 before. The highest score I have gotten prior to GC was 1695. This rom is crazy fast. Never would have imagined a rom based on kd1 could be this fast, considering the fact that supposedly kj2 is the most stable. I just flashed this, so I cant say if this rom is stable or not yet, but I hope it is. If it stays this fast AND its stable, then I wont even worry about overclock or supercharge.
yea same here, ran quadrant on stock gingerclone and got a 1950 and was like "whaaaaaat" its pretty awesome.
have tried supercharging it, with update 9 (and maybe thats where i'm going wrong), and after about an hour it begins to do the 1 buzz followed by 3 quickfire buzzes endlessly, and i'm forced to reinstall Gingerclone. do you guys think its because i should use update 8 for supercharger? its still blindingly fast though
edit: oh and i was wondering if there was some way to take that stupid "update social network" thing from the notification bar with this lockscreen mod? because it takes away the option under the display portion of the settings.
Thanks for the heads up on the lockscreen mod disabling the social status thing. Personally, I always left it on because when you turn it off you lose your power widgets in the drop down too. I know there are widgets that do the same thing, but aside from my launcher and the Google search widget I leave my desktop clean.
I'll have to look into the original framework from gingerclone and merge some of the code with the one from the lockscreen mod. I have some exams to study for this week, but I'll try to get around to it.
yoyowhatup22 said:
try 1.2, that normally is pretty solid.
Click to expand...
Click to collapse
assuming you're running the free version of Tegrak which only changes CPU clock, for the majority of SK4G phones running at 1.2Ghz will actually result in a net loss in performance over 1.1Ghz. Without upping the voltage along with the clock speed, the processor actually begins to encounter errors.
I run into the same thing on my PC. I can safely overclock to 3.0Ghz, and I'm more than capable of overclocking to 3.2, but performance rapidly degrades any higher than 3.01 Ghz without overvolting. At 3.0Ghz, my Geekbench score is around 6500, whereas at 3.02 it falls to under 3000, and at 3.2 it's at about 1200 (if you've never used Geekbench it uses a baseline score of 1000 which is equal to a 2005 PowerMac G5 dual core, and doubling the score means doubling the performance).
If I bump the voltage up from 1.25 to 1.3v, I can overclock even higher than 3.2Ghz, but my computer doesn't have adequate cooling for that and it overheats within about 5 minutes.
So as you can see, faster clock speed doesn't necessarily equate to having a faster phone, the exact opposite can be true. Overvolting would most definitely result in a performance spike, but keep in mind that there IS NO COOLING on a cell phone, not even a heat sink. Excessive heat can very rapidly damage your phone or even break it. Another thing to consider is that not all silicon is created equal especially when you're talking about CPUs.
Look at Intel for example, they sell the EXACT SAME cpu set at different clock speeds as a Celeron, Core2duo and Core2Extreme based on how the individual chip performed on a bench test. This is why overclocking is actually possible. sometimes you get lucky with a well manufactured CPU, and sometimes you end up with a dud that can barely perform at the clock speed it's rated at.
N00b-un-2 said:
Thanks for the heads up on the lockscreen mod disabling the social status thing. Personally, I always left it on because when you turn it off you lose your power widgets in the drop down too. I know there are widgets that do the same thing, but aside from my launcher and the Google search widget I leave my desktop clean.
I'll have to look into the original framework from gingerclone and merge some of the code with the one from the lockscreen mod. I have some exams to study for this week, but I'll try to get around to it.
Click to expand...
Click to collapse
alright, cool. its not a make or break thing for me, i dont mind it that much. but thanks for even saying you'll try!

[Q] Stock ROM much faster processor speeds than custom ROMs?

I'm struggling to get my head around this drastic difference in speed.
Stock ROM on linpack scores about 100Mflops on single thread, and 200Mflops on multi thread similar to this video at 2:50 http://www.youtube.com/watch?v=Zh5quKyypw8
However since flashing CM10 my linpack scores have more than halved! I never get over 80Mflops now on multi thread, and rarely over 40Mflops on single thread. Also in quadrant it shows my cpu scoring really low, but my RAM scoring very high. I was wondering if my the ROMs are effecting my processor speeds?
I am about to go back to the stock ROM to see if linpack scores shoot back up!
If you don't notice a difference who cares about benchmark scores?
Sent from my HTC VLE_U using xda app-developers app
Well I do agree with what you are saying, and it still seems very fast. But I used it for less than a day on the stock ROM, so I don't really have anything to compare it to.
I am wondering more if these benchmarks are giving false readings, or if they are actually picking up on some poor processor optimisations, which whilst may not be noticeable day to day, could have a greater impact on CPU hungry tasks such as making and restoring large backups etc.
But benchmarks for the sake of benchmarks I agree are pointless.
RichardW1992 said:
Well I do agree with what you are saying, and it still seems very fast. But I used it for less than a day on the stock ROM, so I don't really have anything to compare it to.
I am wondering more if these benchmarks are giving false readings, or if they are actually picking up on some poor processor optimisations, which whilst may not be noticeable day to day, could have a greater impact on CPU hungry tasks such as making and restoring large backups etc.
But benchmarks for the sake of benchmarks I agree are pointless.
Click to expand...
Click to collapse
I pretty much stopped using custom roms that aren't based on official sense based roms because of low benchmark numbers, questionable battery life and terrible camera support.
HTC has most if not all drivers for the hardware closed source, so developers are trying their best but I just use the stock ROM
RichardW1992 said:
Well I do agree with what you are saying, and it still seems very fast. But I used it for less than a day on the stock ROM, so I don't really have anything to compare it to.
I am wondering more if these benchmarks are giving false readings, or if they are actually picking up on some poor processor optimisations, which whilst may not be noticeable day to day, could have a greater impact on CPU hungry tasks such as making and restoring large backups etc.
But benchmarks for the sake of benchmarks I agree are pointless.
Click to expand...
Click to collapse
Your not getting false readings... That's just how it is.. Bench marks don't mean anything... Cm will always have a lower score than stock but is just as fast.. If your worried about scores flash a ROM based on sense ota...
Sent from my HTC One S using xda premium

Categories

Resources