I was always wonder if it's true or not that it's possible to get a much higher benchmark score without over clocking just by flashing custom roms?
My phone is a Samsung Infuse i997 , just got it 3-4 days ago.
Personally I do not want or like to over clock the phone or my computer......
Yesterday I made a test with the Q uadrant Benchmark App.
I made two tests, one with stock Froyo 2.2.1 and the other with Infuse 1.5 custom rom. I ran each test several times to get an proper reading.
With the stock Froyo 2.2.1 the benchmark score was 1100 or so and with the Infuse 1.5 custom rom the benchmark score was 1750.
I was told that the Infuse 1.5 custom rom was not by default over clocked and the reason I am getting such a high score in the benchmark test is due to the fact that these custom roms if created properly are a lot faster even though they are not over clocked by default.
So I am just wondering if this is true or are custom roms just made this way to give people false readings ? (to be able to brag )
Has anyone else benchmark tested phones with simular results when comparing stock roms vs. custom roms ?
See ya all on the flip side
Infused is lagfixed to ext4.... it inflates the score.... should read more about whats in it before flashing...
double post. you should put a delete request in the post in the dev section, you may want to read my reply though.
Related
hi guys!
ive been a week trying different roms and using different kernel OC and I could not get more than 3.400 to 1.4 ghz on "quadrant advance program. " install the ROM last night "infin1ty 1.0.0 RC | AOSP | Netarchy 1.3.0.6" ( 1.4 ghz) with the curious fact that descrption says: • IO Quadrant tweak with this rom achievements to 4.400 in quadrant. attached image to see I'm not lying. if anyone knows I need to do to make this happen with other roms, please let me know. I would like to know the reason for this. does anyone happen? excuse my English!
What the "partition hack" does is creating a separate partition in phone memory to trick quadrant to think that it is running the read/write test on SD card while it is actually reading/writing to phone memory thus results in unusually high I/O score.
This hack does not improve the speed of the phone or anything, it's just for showing off to your friends
If you really want to do it, just search for "quadrant hack" and I'm sure you'll find it. (on my phone now so it's not very convenient to find the link for you)
I'm running infin1ty too. But I've overclocked it to 1.3ghz. My score is 3790.
If you want your best score re-run Quadrant several times in a row without exiting the app. Last night on a fresh load of CM7 nightly without overclocking or hacking I was able to hit over 2600.
Cm7 with matrix kernel
My best score was 4536 with quadrant.
I has a 12,000,0000 quadrant on my mogul back in the day.
Sent from my Nexus S 4G using XDA App
mine broke once it hit 9000
sorry to burst bubbles. quadrant really means nothing. it just compares stock phones vs other stock phones at there current state. not am
4539 on mine!
My best was a 4595
Sent from my Htcclay's SuperCM7 NS2 using XDA Premium App
Hi all, first time posting and sorry if this has been asked before. I am running cm7 nightly and have recently tried the quadrant benchmark and am getting some pretty low scores. I know these builds are not stable but even so the scores are low. getting mid 1600s and I have seen other around 2200. is this just because of the cm7 build? thanks.
Sent from my SGH-I897 using Tapatalk
Install this kernel, overclock to 12-1300mhz and try again.
wow sweet! bumped to 1200 and ran a 2000 on quadrant. what's the most the captivate CPU can handle?
Sent from my SGH-I897 using Tapatalk
Can't really say much for Quadrant, as I haven't run it in forever, but I run 1.5ghz stable 24/7 on cm7 w/ glitch. Runs buttery smooth through pretty much any app.
I've kinda been out of the loop in regards to the Captivate, but afaik max clock is 1.6.
I heard something about shao doing a 1.8 kernel, but again, idk how true it is or anything, and I'm lax in time, so I can't really research lol.
yea i see the option to run at 1.6 just didnt know if it was safe to do so. I bet 1.5 runs smooth. Thanks.
Quadrant is nothing of a benchmark, the hummingbird kernel isn't even read correctly. On the CM7 forum I've mentioned this before. Basing everything off a quadrant score is complete in the redundant theory, try using a real benchmark, suggestion.
Smartbench> quadrant
Sent from I'm sure glad Danger went under, else I'd still be on a Mr.Cartoon SK2.
As stated above. Don't put to much into scores. Judge by real world feel. I can show you a Rom that scores almost 3000 on quadrant but runs like crap. Trust your judgement. As for overclocking, it is device dependent. My cap couldn't handle anything over 1.3.
And yes all this has been stated before. A search would have saved you some time.
Thread closed
Hello,
FIRST OFF!
This thread is NOT for discussing whether or not benchmarking is useful. If you feel the need to flame me or anyone here for CARING about benchmarks, then go find another thread to piss in! I made this thread to clean up Faux's thread as it was taking over.
Ever since moving to any of the newer CM7 updates, I have noticed extremely low benchmark scores on CF Bench. When on CM7-1 with one of faux's kernels, everything is fine. (Cf Bench = 7000)
When I say low, Im talking 4000 on CF Bench despite OCing to 1.56Ghz. That is far below what the Sensation STOCK is measured at...
Notes:
1-Scores can be all over the place. At one point, I got a 6400. This would immediately become a 5000, and then back to 4000 range.
2-I have tried using three different methods of controlling the clock. (System tuner by 3c and fauxclock, both recommended by faux for his kernels, and then the CM7 performance settings). Fauxclock had the best outcome, but EXTREMELY unstable!
3-I have done NUMEROUS restarts along with cache clearing and dalvik clearing. I even pulled the A-Bomb out and used the superwipe script to start fresh. After the a-bomb, I DL only what I needed to test with...
4-My phone feels laggy, which goes along with the crappy scores.
At this point, I am restoring to my trusty Alpha 1 + faux kernel. While that works, I feel I will wind up getting left behind.
What could cause this? What are some things I may have overlooked? Others DONT have this problem, at all.
Matt
well i will say benchmarking isnt a totally conclusive. although there isnt any other method of tangible scoreing as of yet. However i will say that if your phone is laggy then your benchmarks are a good way to test why.
scoreing varies when you have active apps in the background unknowingly.
also if you flashed a new rom and kept your old kernel that will dramatically reduce your scores. check out some other kernels for your new rom and see if that makes a difference.
try flashing back to stock and deleting the kernel, then flash the CM7 rom without the kernel in your system and run a test.
theres a lot of things that will change your scoreing.
let me know what you get out of this.
rddocke said:
well i will say benchmarking isnt a totally conclusive. although there isnt any other method of tangible scoreing as of yet. However i will say that if your phone is laggy then your benchmarks are a good way to test why.
scoreing varies when you have active apps in the background unknowingly.
also if you flashed a new rom and kept your old kernel that will dramatically reduce your scores. check out some other kernels for your new rom and see if that makes a difference.
try flashing back to stock and deleting the kernel, then flash the CM7 rom without the kernel in your system and run a test.
theres a lot of things that will change your scoreing.
let me know what you get out of this.
Click to expand...
Click to collapse
I came from ARHD with fauxs kernel. I did a superwipe and then flashed cm7-1 and then flashed fauxs kernel. Everything was great until I tried any of the new CM7's...Just tried cm7-6 and got a WHOPPING 3000!!!!!!!!! YAY!!!!!!!!
I knew it would be low, just playing with the phone. I find it so weird that going back to cm7-1 fixes everything.
Either their new updates kill my performance for whatever reason, or going from one cm7 build to another is a no go. (I did a superwipe just to be sure when going from alpha1 to alpha 5).
Damnit.
If it werent for seeing others with good scores, I would blame cm7 or faux's kernel. Its got to be some little quirk with my phone. (And a few others).
Matt
I've recently tried the Quadrant Standard Benchmark and noticed that my phone's score is severely lower than that of other reviewer's benchmarks? I usually score between 960-1060. Any Idea why I'm not getting into the 1100s? Even my brother's Infuse reached 1300.
Quadrant scores necessarily don't mean squat. But just out of curiosity, what have you done to your phone? Are you running the stock shipped ROM?
Sent from my SGH-I997 using xda premium
I've just rooted it. Bricked it. Used the ultimate Unbrick for Rogers Infuse 4g. Now here I am. One thing I've thought of is space since I have lots of games but eve when I removed almost all of my downloaded Apps. The score hardly improved. I'm wondering whether for those games(mainly Gameloft) where they require a larger download to fully install, when you uninstall those games are the rest of the files from the extra download still on ur phone taking up space?
RazorWare said:
I've just rooted it. Bricked it. Used the ultimate Unbrick for Rogers Infuse 4g. Now here I am. One thing I've thought of is space since I have lots of games but eve when I removed almost all of my downloaded Apps. The score hardly improved. I'm wondering whether for those games(mainly Gameloft) where they require a larger download to fully install, when you uninstall those games are the rest of the files from the extra download still on ur phone taking up space?
Click to expand...
Click to collapse
To be honest, if you're not noticing any lag then all should be fine. I recommend running a custom ROM. pretty much all the infuse roms convert your system to ext4 which is much faster than stock. If you need help with anything don't hesitate to ask, but in the right forum
Sent from my SGH-I997 using xda premium
RazorWare said:
I've just rooted it. Bricked it. Used the ultimate Unbrick for Rogers Infuse 4g. Now here I am. One thing I've thought of is space since I have lots of games but eve when I removed almost all of my downloaded Apps. The score hardly improved. I'm wondering whether for those games(mainly Gameloft) where they require a larger download to fully install, when you uninstall those games are the rest of the files from the extra download still on ur phone taking up space?
Click to expand...
Click to collapse
Apps don't influence ur score much. These scores are based on how ur CPU, GPU and memory works, not based on how apps work.
Repeat the quadrant a couple times in a row. First run always seems to give lower values. After 1-2 runs, take average of next 2-3 and you will get the proper score.
Reboot your phone, give it some time to settle, and then run quadrant.
But like pointed above, stock ROMs give lower scores than custom mainly cos of faster file system. Plus, I enabled GPU based graphics by updating my build.xml and am getting about 1800 with my clock set at max 800.
diablo009 said:
Apps don't influence ur score much. These scores are based on how ur CPU, GPU and memory works, not based on how apps work.
Repeat the quadrant a couple times in a row. First run always seems to give lower values. After 1-2 runs, take average of next 2-3 and you will get the proper score.
Reboot your phone, give it some time to settle, and then run quadrant.
But like pointed above, stock ROMs give lower scores than custom mainly cos of faster file system. Plus, I enabled GPU based graphics by updating my build.xml and am getting about 1800 with my clock set at max 800.
Click to expand...
Click to collapse
apps running in the background can consume resources and affect the score. also rogers roms tend to score lower. it was that way on the captivate too. btw what's a build.xml? do you mean build.prop?
and to the op your score is in line with what you can expect. if you dont have any specific issues i wouldnt worry about it. quadrant is soo sensitive to atleast 2 areas (encoding/decoding media and file system reads) but not others that you can take an infuse to 4200 but it isnt gonna do too much that it didnt do before.
theROMinator said:
Quadrant scores necessarily don't mean squat. But just out of curiosity, what have you done to your phone? Are you running the stock shipped ROM?
Sent from my SGH-I997 using xda premium
Click to expand...
Click to collapse
They really don't!! Neither does Antutu or Linpack.. Gives you a different number every time LOL!! But Definitely look at some of the other roms out there! The few dev's we have left need all the support we can give them!! THANKS INFUSE DEV's!!!!
Hey guys.
On my S8+ I got around 173K antutu score (all scores are in antutu form now on) after i finished setting it all up and with the same apps I have now.
After this I flashed jessie's kernel which OC's the exynos to 2.45 (if im not mistaken) and got from 181K to 182K scores.
Then i flashed CARhd Rom V3 and started getting from 163k to 165k on it's stock kernel and 167K on jessie's, still OC to 2.45. I noticed some performance issues, like lags when i had two apps running on multiscreen so I went back to stock rom, flashed by ODIN.
This said, The benchmarks i got on stock were 163K and with stock rom and jessie's kernel were 167K max.
One should not trust, nor care about benchmarks all that much, but I would like to know if any of you have any idea why i'm scoring 15K lower on the same config WITH A CLEAN INSTALL WITH THE SAME APPS. First answer that came to me was some KNOX bs, but it can not be, as the 182K i got before were on a non stock kernel with knox triggered.
The difference in the scores are, roughly 11K on UX and 3K on CPU. 3D and RAM are the same.
Thank you, guys
badjoras said:
Hey guys.
On my S8+ I got around 173K antutu score (all scores are in antutu form now on) after i finished setting it all up and with the same apps I have now.
After this I flashed jessie's kernel which OC's the exynos to 2.45 (if im not mistaken) and got from 181K to 182K scores.
Then i flashed CARhd Rom V3 and started getting from 163k to 165k on it's stock kernel and 167K on jessie's, still OC to 2.45. I noticed some performance issues, like lags when i had two apps running on multiscreen so I went back to stock rom, flashed by ODIN.
This said, The benchmarks i got on stock were 163K and with stock rom and jessie's kernel were 167K max.
One should not trust, nor care about benchmarks all that much, but I would like to know if any of you have any idea why i'm scoring 15K lower on the same config WITH A CLEAN INSTALL WITH THE SAME APPS. First answer that came to me was some KNOX bs, but it can not be, as the 182K i got before were on a non stock kernel with knox triggered.
The difference in the scores are, roughly 11K on UX and 3K on CPU. 3D and RAM are the same.
Thank you, guys
Click to expand...
Click to collapse
There are many reasons why scores can vary wildly. The main culprit is usually thermal throttling.
I'm running CarHD V4 and score 170k so don't worry about it, the phone is a beast.
I do get the phone is a beast, what confuses me is how it downscores 15K points and 11 of them are on User Experience (UX) on the same phone, same software, same apps after a format =/
On this rom i got this score, with this kernel i got this score, on another rom i got this score, and with that kernel i got this score... just go back to the config that got you the highest score, what's the problem?
peachpuff said:
On this rom i got this score, with this kernel i got this score, on another rom i got this score, and with that kernel i got this score... just go back to the config that got you the highest score, what's the problem?
Click to expand...
Click to collapse
The problem is that he went back to the configuration that have him the best score, but it's 15k lower for some reason.
peachpuff said:
On this rom i got this score, with this kernel i got this score, on another rom i got this score, and with that kernel i got this score... just go back to the config that got you the highest score, what's the problem?
Click to expand...
Click to collapse
The problem is that my ability to find the problem is about the same as your ability to read a post that is as explicit as it can get...
Thread cleaned up.
Please behave civil. There is no need to troll.
Kindly stay on topic and respect each other and their opinions.
Regards
Vatsal,
Forum Moderator.
I was being civil.
Were you doing all of this consistently? testing, flashing, testing, flashing, etc.? Because the phone could have just needed a break.
sireniankyle said:
Were you doing all of this consistently? testing, flashing, testing, flashing, etc.? Because the phone could have just needed a break.
Click to expand...
Click to collapse
Tests on all roms were conducted with the device rested for 10 minutes with no apps running and screen off so it would cool down.
The 182k was tested 4 times in a row, on which the differences maxed at 2k. Same for test 2 (car hd rom), that was conducted 2 days after I installed the rom.
Lastly, the back to settings 1 test was conducted yesterday with the same procedure aswell.
Today I got it to top out at 178k (stock rom, jesse kernel OC) using kernel control to lock all 8 CPU at max speed (2.45 and 1.69, for big and LITTLE), with it on the regular governor it is doing 175 overclocked to the before mentioned speeds.
I don't know what happened to it, but those 162k it first got were the same as when I turned on power saving mode whilst it was able to do 182K, which is odd and the first thing i checked was if it was on, which it wasn't.
All tests were conducted with 1080p screen, RAM cleared, Excess files erased, with everything turned off except for wifi so it can submit the score.
Like I said, benchmarks are worth what they are, and I only conducted antutu testing again because I was noticing a performance deficit. 175 (It's 4k higher than an OP5) is more than fine and I notice no more lags on daily usage.
For reference, it is scoring 2147 - single-core and 6569 multi-core on geekbench 4 aswell, which, in multicore, is 150 points lower than it was when it was doing 182 on antutu.