Here, I prove that synthetic benchmarks fail.
{
"lightbox_close": "Close",
"lightbox_next": "Next",
"lightbox_previous": "Previous",
"lightbox_error": "The requested content cannot be loaded. Please try again later.",
"lightbox_start_slideshow": "Start slideshow",
"lightbox_stop_slideshow": "Stop slideshow",
"lightbox_full_screen": "Full screen",
"lightbox_thumbnails": "Thumbnails",
"lightbox_download": "Download",
"lightbox_share": "Share",
"lightbox_zoom": "Zoom",
"lightbox_new_window": "New window",
"lightbox_toggle_sidebar": "Toggle sidebar"
}
My god.... How the heck did you do it?
Benchmarks are becoming less and less meaningful as people customize and mod their systems more and more.
If I could find a USB port on a cocker spaniel and install Android on it, I could tweak the poor little guy to get 2500 on Quadrant if I really wanted to (and if I didn't have to give the dog back afterwards).
The highest my SGS go is 1517 Quadrant. How you guys mannaged to get 2000 and above.
Using Darky 9.5, dark core 1.4
I got 3147 quadrant with miui rc4 and oc to 1300MHz.
Sent from my GT-I9000 using XDA App
progene81 said:
The highest my SGS go is 1517 Quadrant. How you guys mannaged to get 2000 and above.
Using Darky 9.5, dark core 1.4
Click to expand...
Click to collapse
so called lag fix for start, and some other tweaks including OC.
But lag fix itself gives a huge boost in quadrant.
I do agree synthetic benchmarks suck especially concerning phones since there are many chips, and many other optimizations for particular CPUs, and they're not equal in either implementation or HW is so much different that comparisons are pointless.
For instance snapdragon might lack with raw power, and our hummingbird is better, but when it comes to SIMD instruction ALUs snapdragon wins + quadrant's optimization for snapdragon et viola...
rschenck said:
Benchmarks are becoming less and less meaningful as people customize and mod their systems more and more.
If I could find a USB port on a cocker spaniel and install Android on it, I could tweak the poor little guy to get 2500 on Quadrant if I really wanted to (and if I didn't have to give the dog back afterwards).
Click to expand...
Click to collapse
Kinda like this?:
theduckking said:
Kinda like this?:
View attachment 565756
Click to expand...
Click to collapse
Exactly, but with a dorsal fin and ramjets.
rschenck said:
Exactly, but with a dorsal fin and ramjets.
Click to expand...
Click to collapse
sorry, that's the turbo ARDog (Android Ready Dog) I got. If you want to customize it further you can have mine, PM me with your address if you're interested so I can organize the shipment (or well, fuell and program the GPS of the ARDog, it can fly so why not utilize it^^)
theduckking said:
sorry, that's the turbo ARDog (Android Ready Dog) I got. If you want to customize it further you can have mine, PM me with your address if you're interested so I can organize the shipment (or well, fuell and program the GPS of the ARDog, it can fly so why not utilize it^^)
Click to expand...
Click to collapse
Isn't there an App for that?
I made this score
Using GingerCriskelo V27 and CF-Root 3.0
Sent from my GT-I9000 using XDA Premium App
My score:
+ The feeling of a pretty darn fast ROM
Of course the usage was pretty low over the weekend but still^^
Pretty easy to get 4500-4600 scores on Quadrant with CM7 and a few kernel/system tweaks..
So which tweaks did u set?
Sent from my GT-I9000 using XDA Premium App
Tk-Glitch said:
Pretty easy to get 4500-4600 scores on Quadrant with CM7 and a few kernel/system tweaks..
Click to expand...
Click to collapse
So you 've overclocked to 1200 MHz or still Using 1000?
Sent from my GT-I9000 using XDA Premium App
w7nt3rmut3 said:
So which tweaks did u set?
Sent from my GT-I9000 using XDA Premium App
Click to expand...
Click to collapse
1 - Use a RAMdisk for Quadrant storage;
2 - Use old graphics driver from NS;
3 - Overclock.
Overclocking the bus will lead to faster RAM, so faster I/O score from the RAMdisk.
w7nt3rmut3 said:
So you 've overclocked to 1200 MHz or still Using 1000?
Sent from my GT-I9000 using XDA Premium App
Click to expand...
Click to collapse
Yes, 1.2GHz. Sorry^^.
At 1GHz, ~4000.
And with 1.4GHz, 4800-4900.
1.5GHz here :
Now 1.4GHz only, but with a faster bus speed, resulting in a boost on I/O and memory scores :
EDIT : Added *lololz* screenshot.
EDIT2 : Replaced *lololz* screenshot by a 1.5GHz one for even more LULZ.
EDIT3 : Added faster bus speed to bring them all and in the darkness bind them.
edited
10 chars
coldflid said:
Photoshop...
Click to expand...
Click to collapse
Absolutely not. Quadrant is just very easy to fool on the I/O score, as I said : with a ramdisk.
The normal I/O score on CM7 is ~1300, resulting in about 3000 final score @ 1.5GHz. Everyone here with an i9000 can have ~4900 with a light CM7 and 1.4GHz OC using a RAMdisk. It's a dirty hack for sure.
If you take the I/O score away, all the others are genuine. In fact, except the CPU score, they are all higher @ 1.4GHz (because of a faster bus frequency).
If my phone was stable at a higher bus, 5200-5300 should be possible at 1.5GHz, and even more at 1.6GHz (unstable for now on my device until I find a way to stabilize it).
Tk-Glitch said:
Absolutely not. Quadrant is just very easy to fool on the I/O score, as I said : with a ramdisk.
The normal I/O score on CM7 is ~1300, resulting in about 3000 final score @ 1.5GHz. Everyone here with an i9000 can have ~4900 with a light CM7 and 1.4GHz OC using a RAMdisk. It's a dirty hack for sure.
If you take the I/O score away, all the others are genuine. In fact, except the CPU score, they are all higher @ 1.4GHz (because of a faster bus frequency).
If my phone was stable at a higher bus, 5200-5300 should be possible at 1.5GHz, and even more at 1.6GHz (unstable for now on my device until I find a way to stabilize it).
Click to expand...
Click to collapse
No, i didn't mean yours, i totally believe yours, i meant the one on the first post, and now that i look it more closely it doesn't seem fake.
Sorry
Related
{
"lightbox_close": "Close",
"lightbox_next": "Next",
"lightbox_previous": "Previous",
"lightbox_error": "The requested content cannot be loaded. Please try again later.",
"lightbox_start_slideshow": "Start slideshow",
"lightbox_stop_slideshow": "Stop slideshow",
"lightbox_full_screen": "Full screen",
"lightbox_thumbnails": "Thumbnails",
"lightbox_download": "Download",
"lightbox_share": "Share",
"lightbox_zoom": "Zoom",
"lightbox_new_window": "New window",
"lightbox_toggle_sidebar": "Toggle sidebar"
}
My latest score ;-) Had a couple apps running in the background also.
Running bionix Final rom W/ King Klicks Kernel @ 1.2ghz
Ryans Lag Fix
Sent from my SGH-T959 using XDA App
Is The Nexus One CPU THAT Good?
BTW...Thats Awesome!!
Erickomen27 said:
Is The Nexus One CPU THAT Good?
BTW...Thats Awesome!!
Click to expand...
Click to collapse
The Nexus One has JIT which gives you a a major boost in CPU! JIT is only available on Android 2.2 devices.
SGS phones in the U.S. are only running 2.1 @ the moment. So when 2.2 comes around I'm sure 3000+ Quadrant scores and 30MLOPS will be possible.
player4lifeov said:
The Nexus One has JIT which gives you a a major boost in CPU! JIT is only available on Android 2.2 devices.
SGS phones in the U.S. are only running 2.1 @ the moment. So when 2.2 comes around I'm sure 3000+ Quadrant scores and 30MLOPS will be possible.
Click to expand...
Click to collapse
SGS with jit hits around 14.4 on linpack. Look it up though. The processors are different...linpack is optimized for the type of chip in the n1.
Sent from my Fusion powered Vibrant.
player4lifeov said:
The Nexus One has JIT which gives you a a major boost in CPU! JIT is only available on Android 2.2 devices.
SGS phones in the U.S. are only running 2.1 @ the moment. So when 2.2 comes around I'm sure 3000+ Quadrant scores and 30MLOPS will be possible.
Click to expand...
Click to collapse
Yeah I'm Running 2.2 Major Spped Overhaul But I'm Getting Like 13.7MFLOPS And 1030 Quadrant
Yes Running JIT
(Not Optimized)
player4lifeov said:
The Nexus One has JIT which gives you a a major boost in CPU! JIT is only available on Android 2.2 devices.
SGS phones in the U.S. are only running 2.1 @ the moment. So when 2.2 comes around I'm sure 3000+ Quadrant scores and 30MLOPS will be possible.
Click to expand...
Click to collapse
from what i have read, there are 2 problems with benchmark scores.
1. they favor the processors other then hummingbird (meaning they give false high numbers to other devices and false low numbers to hummingbird processors)
2. bench marks dont always translate into real world/daily use performance of the phone
So if it feels fast and great, who cares what your bench marks are
Do you see that huge Green Part of your score, yup. It is inflating your score by almost 800 points on quandrant.
Vibrant + xda app
Vibrant + xda app
Wtf.
Won't upload my score. Tells me wrong file format ....how? Its PNG.
No lag fix. Stock w/ update jl6
Vibrant + xda app
j56bp said:
No lag fix. Stock w/ update jl6
Vibrant + xda app
Click to expand...
Click to collapse
Nice!
Sent from my SGH-T959 using XDA App
PaiPiePia said:
Do you see that huge Green Part of your score, yup. It is inflating your score by almost 800 points on quandrant.
Click to expand...
Click to collapse
Haha, it is isn't it? So I guess that's my way off falsifying information through this benchmark. Since the benchmark favors snap dragon cpu. ;~}
Sent from my SGH-T959 using XDA App
@jb56- Are you running an optimized kernel? That's a really good score
Sent from my SGH-T959 using XDA App
My Linpack Is Number One On The SGS Devices
It's not the benchmark app, its Hummingbird not using the full potential of JIT
kboater said:
from what i have read, there are 2 problems with benchmark scores.
1. they favor the processors other then hummingbird (meaning they give false high numbers to other devices and false low numbers to hummingbird processors)
2. bench marks dont always translate into real world/daily use performance of the phone
So if it feels fast and great, who cares what your bench marks are
Click to expand...
Click to collapse
I don't agree with your statements. When strictly running Eclair 2.1, how come almost every benchmark showed that SGS was the fastest 2.1 phone around?
Then throw in Froyo 2.2, and the Dalvik JIT compiler, and suddenly the SGS is no longer the fastest phone.
So if these benchmark apps favor chips other than Hummingbird, how come they showed SGS was the fastest 2.1 phone? Are you saying all these benchmark developers have a conspiracy to show SGS as the fastest 2.1 phone, but the slowest 2.2 phone?
Come on, it's not the benchmarks. IT is the actual Hummingbird CPU being UNABLE to take advantage of the Dalvik JIT compiler.
The benchmarks are pretty accurate, and they are showing that Hummingbird is unable to take advantage of the huge performance gains of Dalvik JIT.
If your argument was to be accurate, then all benchmarks should have favored non hummingbird CPU's running 2.1 also.
I just don't buy it, it's not the benchmark, its Hummingbird not utilizing JIT while other CPU's like Scorpion do.
player4lifeov said:
The Nexus One has JIT which gives you a a major boost in CPU! JIT is only available on Android 2.2 devices.
SGS phones in the U.S. are only running 2.1 @ the moment. So when 2.2 comes around I'm sure 3000+ Quadrant scores and 30MLOPS will be possible.
Click to expand...
Click to collapse
Actually SGS phones that are already running 2.2, and SGS Tablets running Froyo 2.2 have not been able to reach 30MFLOPS. The Hummingbird will NOT gain the huge cpu performance gains from 2.2 Dalvik JIT compiler. Unfortunate but seems to be true.
Just got these results with Eugene's Froyo rom JPA Mancnut R2 with Kernel #69
qreffie said:
Just got these results with Eugene's Froyo rom JPA Mancnut R2 with Kernel #69
Click to expand...
Click to collapse
Ummm......that's an ok Linpack score I guess
~SB
PaiPiePia said:
Do you see that huge Green Part of your score, yup. It is inflating your score by almost 800 points on quandrant.
Click to expand...
Click to collapse
That's because it's lag fixed.
Ji6 core kernel and oclf
Team Whiskey has been successful in overclocking the T-Mobile Vibrant to 1.6 GHz, according to this thread: http://forum.xda-developers.com/showpost.php?p=8844947&postcount=41
Pretty incredible stuff. They're working on incorporating voodoo into this kernel right now, then they'll release it.
It sounds too good to be true.
omg!
(yeah, that's about all I have to say)
I believe I read somewhere that they were gonna do a captivate specific ROM via bionix. Pretty sure it is for 2.1, linpack is @12 MFLOPS and open GL site shows it on 2.1.-impressive? Yes . Can't say that I wouldnt take it for a spin though. Hopefully 2.2 will soon follow
Sent from my SAMSUNG-SGH-I897 using XDA App
kennethpenn said:
Team Whiskey has been successful in overclocking the T-Mobile Vibrant to 1.6 GHz, according to this thread: http://forum.xda-developers.com/showpost.php?p=8844947&postcount=41
Pretty incredible stuff. They're working on incorporating voodoo into this kernel right now, then they'll release it.
It sounds too good to be true.
Click to expand...
Click to collapse
Damn, with the back panel on the Captivate being metal, you might burn your hand with this setup.
kennethpenn said:
Team Whiskey has been successful in overclocking the T-Mobile Vibrant to 1.6 GHz, according to this thread: http://forum.xda-developers.com/showpost.php?p=8844947&postcount=41
Pretty incredible stuff. They're working on incorporating voodoo into this kernel right now, then they'll release it.
It sounds too good to be true.
Click to expand...
Click to collapse
Its real...
Sent from my SAMSUNG-SGH-I897 using XDA App
designgears said:
Its real...
Sent from my SAMSUNG-SGH-I897 using XDA App
Click to expand...
Click to collapse
I came buckets
Dg do you have your hand in this cookie jar as well? Can we expect an overclocked version of cog 10? Lol
Sent from my SAMSUNG-SGH-I897 using XDA App
I'll pretty sure they're gonna make z4mod their base. They're also going to implement led button notifications (;
Sent from my SGH-T959 using XDA App
Yup its real his name is morfic. I briefly saw him and unhelpful chatting about it. I cant believe this thing is stable... only bad side is that it only runs eclair till we get source for froyo kernel.
Man I've been wanting to push this badboy past 1.2 ghz
Stable and tentative... Could see 1800! I was just thinking today (with the droid oc and all) when another one would come our way. I guess it was just a matter of time. And boy did it ever come!! Great thing is there is already plans to bring it our way
Sent from my SAMSUNG-SGH-I897 using XDA App
I wanna see the screen shot with them running "quadrant advanced" to see the cpu score!!
{
"lightbox_close": "Close",
"lightbox_next": "Next",
"lightbox_previous": "Previous",
"lightbox_error": "The requested content cannot be loaded. Please try again later.",
"lightbox_start_slideshow": "Start slideshow",
"lightbox_stop_slideshow": "Stop slideshow",
"lightbox_full_screen": "Full screen",
"lightbox_thumbnails": "Thumbnails",
"lightbox_download": "Download",
"lightbox_share": "Share",
"lightbox_zoom": "Zoom",
"lightbox_new_window": "New window",
"lightbox_toggle_sidebar": "Toggle sidebar"
}
This one is Cognition 8.1 with z4mod.
Damn!!! I just read it as well, cant wait to give it a spin some day
if you read the current setup required for that 3k+ quadrant score, he said he was running OCLF, so drop between 500-700 points just for that to get a more realistic voodoo score. it's impressive, don't get me wrong, but it isn't as astounding because a lot of that score is still from OCLF (hell you can get 2500 with 1.2GHz and OCLF)...
as for a *STABLE* 1.6 GHz OC, I'll believe it when I see it.
Kaik541 said:
as for a *STABLE* 1.6 GHz OC, I'll believe it when I see it.
Click to expand...
Click to collapse
Yeah my phone doesn't run well with a simple 1.2 GHz OC... so I'm skeptical of a stable 1.6 OC kernel.
Quadrant is just terrible. Any comparison scores should be taken using Advanced with the I/O test subtracted. But, even then, it only scores Qualcomm MSM processors "correctly."
Kaik541 said:
if you read the current setup required for that 3k+ quadrant score, he said he was running OCLF, so drop between 500-700 points just for that to get a more realistic voodoo score. it's impressive, don't get me wrong, but it isn't as astounding because a lot of that score is still from OCLF (hell you can get 2500 with 1.2GHz and OCLF)...
as for a *STABLE* 1.6 GHz OC, I'll believe it when I see it.
Click to expand...
Click to collapse
It also gets 78fps in neocore versus ~55fps stock and it removes the fps cap. Issues with 1.2ghz seem to be due to undervolting rather than overclocking to 1.2ghz. In my experience it has worked fine at 1.2ghz without undervolt.
http://phandroid.com/2010/10/27/lat...-unlock-achieving-78-0-breaks-3k-on-quadrant/
WickedStyx said:
It also gets 78fps in neocore versus ~55fps stock and it removes the fps cap. Issues with 1.2ghz seem to be due to undervolting rather than overclocking to 1.2ghz. In my experience it has worked fine at 1.2ghz without undervolt.
http://phandroid.com/2010/10/27/lat...-unlock-achieving-78-0-breaks-3k-on-quadrant/
Click to expand...
Click to collapse
yes, but I never questioned the FPS cap, I have no idea what affects that so I can't question it. besides, it's not like our eyes can see much more than 30 FPS, so the difference between 60 FPS and 78 FPS means little to me (in day to day usage). also, I never said anything about undervolting either or made a comment about the 1.2 GHz overclock. Mine runs fine with a 1.2 GHz overclock and the hummingbird processor has already been theorized to be stable up to 1.6 GHz, so this isn't really news, just the first time someone is advertising it as fully implemented. and, what stable means, is it runs across all the devices without things like the "sleep of death" effect occuring (which still happened for SOME with the 1.2 GHz OC).
so until I can personally see 1.6 GHz on my phone without any ill effects on battery life (which should only be minimal if the CPU is being scaled dynamically) or sleep death effects, I won't make a judgment... which is what I already said
This is awesome though! Woot woot!
Who needs a Galaxy S 2 with this?
2GHZ OC here we come!
Will hopefully find out soon a dl link for the vibrant ROM is in new thread "Manhattan project." Viral black posted that they have a Captivate but will try to port to every sgs variant possible. Captivate will be 2nd I would suppose. Sombionix did say they were working on a 2.2 ROM, so hopefully this will meet froyo in the not so distant future. It seemed to be working well from the posts that I read. Then again it is still early.
Sent from my SAMSUNG-SGH-I897 using XDA App
Can be downloaded now:
"Manhattan Project"
http://www.teamwhiskey.com/home/downloads
I don't know how I got this, I usually only get 4100 and 4200 max but some how I ended up getting 4451! This defiantly competes with the SGS2. Please post any score you have here!
{
"lightbox_close": "Close",
"lightbox_next": "Next",
"lightbox_previous": "Previous",
"lightbox_error": "The requested content cannot be loaded. Please try again later.",
"lightbox_start_slideshow": "Start slideshow",
"lightbox_stop_slideshow": "Stop slideshow",
"lightbox_full_screen": "Full screen",
"lightbox_thumbnails": "Thumbnails",
"lightbox_download": "Download",
"lightbox_share": "Share",
"lightbox_zoom": "Zoom",
"lightbox_new_window": "New window",
"lightbox_toggle_sidebar": "Toggle sidebar"
}
LOL, and our OC isn't even near as high! 1.1 vs 1.5 ghz!!
SICK! Wow I cant f**king wait!
Chubby_Skunk said:
SICK! Wow I cant f**king wait!
Click to expand...
Click to collapse
LOL if we break 1.1... that would be amazing. I wouldn't be surprised to see wayy over 5000!
Do yal not understand how out of date quadrant is? Its results are worthless on any dualcore phone period.
Sent from my Xoom using Tapatalk
jug6ernaut said:
Do yal not understand how out of date quadrant is? Its results are worthless on any dualcore phone period.
Sent from my Xoom using Tapatalk
Click to expand...
Click to collapse
Haha, yeah I'm pretty sure every one knows that. It's just fun being silly with it
This should be in general, not themes and apps.
Why is that ?... post about a quadrant score , which is an app .
Is there any good benchmark for dual core ?
ofek said:
Why is that ?... post about a quadrant score , which is an app .
Is there any good benchmark for dual core ?
Click to expand...
Click to collapse
Yeah, it's called real lief performance.
If everything works with no lag = epic bench mark
If something lags = crappy software
xploited said:
Yeah, it's called real lief performance.
If everything works with no lag = epic bench mark
If something lags = crappy software
Click to expand...
Click to collapse
Lol +0+0+0+0+0+1= +1
Use Chain Fire (CF) Bench... much better for dual core
CaelanT said:
This should be in general, not themes and apps.
Click to expand...
Click to collapse
Well this is sort of where this kind of benchmark stuff has always been.
CaelanT said:
This should be in general, not themes and apps.
Click to expand...
Click to collapse
Quadrant is an App
1. How are you doing a quadrant benchmark? Mine errors out at "decoding video". (10/12)
2. I am working on a 1.2 ghz overclock kernel. So far a lot of bugs, maybe that's why quadrant failed... if I succeed, I will try 1.4... if I can even out the major bugs I will release it.
Also, just want to add that on said kernel I got a 6752 overall in Chainfire...
Kthxbai
Sent from my friend's Atrix
trvstw said:
1. How are you doing a quadrant benchmark? Mine errors out at "decoding video". (10/12)
2. I am working on a 1.2 ghz overclock kernel. So far a lot of bugs, maybe that's why quadrant failed... if I succeed, I will try 1.4... if I can even out the major bugs I will release it.
Also, just want to add that on said kernel I got a 6752 overall in Chainfire...
Kthxbai
Sent from my friend's Atrix
Click to expand...
Click to collapse
You're likely getting an error at "decoding video" because you have used your camera. In order to get quadrant to work properly, you have to run a benchmark before you've opened the camera and caused it to force close.
thegregulator said:
You're likely getting an error at "decoding video" because you have used your camera. In order to get quadrant to work properly, you have to run a benchmark before you've opened the camera and caused it to force close.
Click to expand...
Click to collapse
Thank you.
Sent from my friend's Atrix
Smartbench 2011
Awesome Benchmark...
does graphics tests
whats the point hight quadrant or any benchmark if the phone stills laggy and can't play 720p lose the fingerprint captabilities and webtop?!
if you want hight quadrant without webtop and fingerprint buy another phone xD lol
just saying.....please don't interpret me wrong
or, if you want all those features, don't install a pre-beta version of a ROM?
highest quadrant score i've gotten so far is 3400-3500, what are you guys doing with your settings to get so high?
Alcapone263 said:
or, if you want all those features, don't install a pre-beta version of a ROM?
highest quadrant score i've gotten so far is 3400-3500, what are you guys doing with your settings to get so high?
Click to expand...
Click to collapse
I'm not sure how many times you've tried the test, but I did it about five times in a row and each time my score raised by about 150 until it steadied at 4200 or so. Make sure to free as much RAM and extra processor usage as possible using a task manager or similar. Also, are you at 1.1Ghz? That's what I'm at with the kernel that came with the CM7 pre-beta install. I believe it is faux's kernel.
vinay427 said:
I'm not sure how many times you've tried the test, but I did it about five times in a row and each time my score raised by about 150 until it steadied at 4200 or so. Make sure to free as much RAM and extra processor usage as possible using a task manager or similar. Also, are you at 1.1Ghz? That's what I'm at with the kernel that came with the CM7 pre-beta install. I believe it is faux's kernel.
Click to expand...
Click to collapse
Lol, and unmount your storage (just plug it into the computer, mass storage)
Evo Shift 4G benchmark CONTEST - lets get this "best kernel" debate out of the way
Who can beat my Antutu (benchmark program available in market) record?
CM7 nightly, min cpu 1900 mhz, max cpu 1900mhz, Toasty gov, hardware version 0004.
Scarystable compcache disabled:
RAM 612
CPU integer 1221
CPU float-point 518
2d graphics 276
3d graphics 793
database io 535
sdcard write 9.9MB/s 99
sdcard read 17.5 MB/s 175
total score: 4229
Will post more detailed specs later.
If you're getting over 4000, you're doing it right.
Linpack single thread score = 62.33 MFLOPS, Time 1.35 Seconds, Norm Res 5.68, Precision 2.22044604...
Linpack multi thread score = 57.961 MFLOPS, Time 2.91 seconds Norm Res, 3.24, Precision 2.22044604...
P.S. You should use Toasty (or performance gov depending on kernel) for benchmarking and run benchmarks on a fresh boot and close as many apps as possible.
I actually just beat my record.
{
"lightbox_close": "Close",
"lightbox_next": "Next",
"lightbox_previous": "Previous",
"lightbox_error": "The requested content cannot be loaded. Please try again later.",
"lightbox_start_slideshow": "Start slideshow",
"lightbox_stop_slideshow": "Stop slideshow",
"lightbox_full_screen": "Full screen",
"lightbox_thumbnails": "Thumbnails",
"lightbox_download": "Download",
"lightbox_share": "Share",
"lightbox_zoom": "Zoom",
"lightbox_new_window": "New window",
"lightbox_toggle_sidebar": "Toggle sidebar"
}
I haven't ran it yet with the new pershoot, i get better scores with scary. But I'm still a pershoot fan but for all who care I do (ill test the new pershoot later) get better scores with scary
Teleported from my Shifty Speedy using Tapatalk
we now bow to the master
leadfoot99 said:
we now bow to the master
Click to expand...
Click to collapse
Sorry, you are limited to five thanks per day.
BTW you are running a really old version of Antutu. Latest version is 2.3.1
that was an old score
246 views, 2 replies. Sad.
I thought you all had more balls than that.
Well there is no real best kernel. Its preference and they also act different on different phones.
Sent from my Supreme Shift Minimally!
Just hearing excuses, no numbers.
You want numbers dude I'm hitting 2300+ on Supreme how's them for numbers while your running Round here bragging about barely hitting 4k overclocked to 1.9 on a stripped down aosp ROM I'm doing 2300 on Supreme with 27 widgets set @ STOCK SPEEDS of 806, if you haven't noticed all us big kids don't care about the numbers because their pointless, this isn't a (maybe I shouldn't have used that word) measuring contest
Sent from my PG06100 using Tapatalk
..........................double post
strapped365 said:
You want numbers dude I'm hitting 2300+ on Supreme how's them for numbers while your running Round here bragging about barely hitting 4k overclocked to 1.9 on a stripped down aosp ROM I'm doing 2300 on Supreme with 27 widgets set @ STOCK SPEEDS of 806, if you haven't noticed all us big kids don't care about the numbers because their pointless, this isn't a (maybe I shouldn't have used that word) measuring contest
Sent from my PG06100 using Tapatalk
Click to expand...
Click to collapse
I care. And that's why I made the thread. Want to see how I can get the best performance out of my phone. Obviously, this cannot be done on a sense 3.0 rom until the kernel source gets released.
VICODAN said:
I care. And that's why I made the thread. Want to see how I can get the best performance out of my phone. Obviously, this cannot be done on a sense 3.0 rom until the kernel source gets released.
Click to expand...
Click to collapse
What where trying to say is the title is misleading the purpose of the thread (or your being misleaded yourself) "the best kernel" is not based on how high the numbers can get on a benchmark score. I've had sense run as fast as AOSP and not get close to its points, For me I get the best performance (when using AOSP) using pershootv9 its very smooth, scary can give numbers in the 4,000's but with my phone its pershoot that responds the best.
High numbers, do NOT measure performance
Sent from my Supreme Shift through a minimal tunnel
notsointeresting said:
What where trying to say is the title is misleading the purpose of the thread (or your being misleaded yourself) "the best kernel" is not based on how high the numbers can get on a benchmark score. I've had sense run as fast as AOSP and not get close to its points, For me I get the best performance (when using AOSP) using pershootv9 its very smooth, scary can give numbers in the 4,000's but with my phone its pershoot that responds the best.
High numbers, do NOT measure performance
Sent from my Supreme Shift through a minimal tunnel
Click to expand...
Click to collapse
Pershoot responds like crap on my phone. I realize there are different hardware versions. This is more of a benchmarker/tweaker thread. I want to see who can get the best benchmarks and with what kernel.
Linpack does a benchmark of raw megaflops. If you don't want to use antutu fine, linpack is good enough for me:
In computing, FLOPS (or flops or flop/s, for floating point operations per second) is a measure of a computer's performance, especially in fields of scientific calculations that make heavy use of floating point calculations, similar to the older, simpler, instructions per second. Since the final S stands for "second", conservative speakers consider "FLOPS" as both the singular and plural of the term, although the singular "FLOP" is frequently encountered. Alternatively, the singular FLOP (or flop) is used as an abbreviation for "FLoating-point OPeration", and a flop count is a count of these operations (e.g., required by a given algorithm or computer program). In this context, "flops" is simply the plural rather than a rate.
Click to expand...
Click to collapse
stock
Stock rooted 2.2 @806 also. 2331
VICODAN said:
Pershoot responds like crap on my phone. I realize there are different hardware versions. This is more of a benchmarker/tweaker thread. I want to see who can get the best benchmarks and with what kernel.
Linpack does a benchmark of raw megaflops. If you don't want to use antutu fine, linpack is good enough for me:
Click to expand...
Click to collapse
Ok that's fine, if thats what your aiming for though the title of this thread is a little misleading.
and when I say each phone functions different its not just by hardware version, yes the shift has 2 (0005 for the newer ones) but theres many other factors that makes the phone respond differently. there not all built the same.
and just to share a linpack score for the hell of it- 42 on Supreme Sense
notsointeresting said:
Ok that's fine, if thats what your aiming for though the title of this thread is a little misleading.
and when I say each phone functions different its not just by hardware version, yes the shift has 2 (0005 for the newer ones) but theres many other factors that makes the phone respond differently. there not all built the same.
and just to share a linpack score for the hell of it- 42 on Supreme Sense
Click to expand...
Click to collapse
Very good for Sense. Thanks for sharing. Was that multi or single threaded?
doublepost
VICODAN said:
Very good for Sense. Thanks for sharing. Was that multi or single threaded?
Click to expand...
Click to collapse
Single, since multi-thread is meant more for dual-core phones
notsointeresting said:
Single, since multi-thread is meant more for dual-core phones
Click to expand...
Click to collapse
Yeah I thought the Shift had a multi threaded processor, dual core or not. Interesting. Thanks for your numbers.
I'm curious to know if anyone has incredible results
I use quadrant and cf- bench, for now I have a quadrant screenshot, this is also my average score:
Uploaded at Picoodle.com
I use last Mysam rom with abysskernel 2.5
theendfear said:
I'm curious to know if anyone has incredible results
I use quadrant and cf- bench, for now I have a quadrant screenshot, this is also my average score:
Uploaded at Picoodle.com
I use last Mysam rom with abysskernel 2.5
Click to expand...
Click to collapse
At what cpu speed, 1600-1664-1704?
Use Antutu or CF-Bench...Quadrant is a random number generator.
Mine is fully stock and antutu gives something around 6200, what about others?
Sent from my GT-N7000 using XDA App
this is my score with cf-bench:
Uploaded at Picoodle.com
always with abyssnote kernel 2.5 and mysam rom at 1.66ghz
CPU at 1.7GHz @ Rocket ROM v11
{
"lightbox_close": "Close",
"lightbox_next": "Next",
"lightbox_previous": "Previous",
"lightbox_error": "The requested content cannot be loaded. Please try again later.",
"lightbox_start_slideshow": "Start slideshow",
"lightbox_stop_slideshow": "Stop slideshow",
"lightbox_full_screen": "Full screen",
"lightbox_thumbnails": "Thumbnails",
"lightbox_download": "Download",
"lightbox_share": "Share",
"lightbox_zoom": "Zoom",
"lightbox_new_window": "New window",
"lightbox_toggle_sidebar": "Toggle sidebar"
}
@vessk0 wow!! at antutu test 7528 its exellent! I tried now, and don't go up 6900.. I think Its time to change rom =)
Are the 3D benchmarks limited to ±58fps (the refresh rate of the screen)? Is there any v-sync off setting that could unleash the power and give better results?
Yup, as far as i know it's limited to 60 fps. (Correct me if i'm wrong)
vessk0 said:
CPU at 1.7GHz @ Rocket ROM v11
Click to expand...
Click to collapse
Do I saw something wrong?
RAM: 1205
how you do that? can you teach me?
Even better with v12, second in world N7000 chart
Sent from my GT-N7000 using XDA App
Hi vessk0, I have the same rom now, but the score is too low
what's your configuration? I tried with 1.7ghz, aggressive task killer in sistem tuner app, 2048mb cache on sd and noop setted in I/O scheduler section.. the score is around 6900-7000 points
For benchmark I use 1.7GHz @ Performance governer, but internal SD card in settings of Antutu. That's it.
vessk0 said:
For benchmark I use 1.7GHz @ Performance governer, but internal SD card in settings of Antutu. That's it.
Click to expand...
Click to collapse
every time I use 1.7GHz and performance gov. my phone allways freezes.
I tried some different roms and kernels but same result
any advice?
k-12 said:
every time I use 1.7GHz and performance gov. my phone allways freezes.
I tried some different roms and kernels but same result
any advice?
Click to expand...
Click to collapse
Check your voltages, try to reduce it to 1325-1350 for 1.7 (default is 1450). That'll help it run cooler, and make it less likely to lock up.
But that said, you may just have an Exynos chip that is not happy at that speed, there is no guarantee that they can all do 1.7.
Rocket ROM V12/Abyss 2.6 @1.7GHz Performance Governor:
Quadrant
Linpack
Antutu
CF-Bench
SunSpider
BrowserMark
Vellamo
And here's a fairly recent benchmark of the Droid Razr with some other phones...you can use this as rough basis of comparison.
http://www.anandtech.com/show/5198/motorola-droid-razr-review-a-better-clad-bionic/8
TL;DR: The Note really kicks ass when you crank up the clocks!
@croak I have tested your same combination rom/kernel and for now seems the best! with performance gov and cpu set to 1.7ghz the score is very higher in all tests
our note is one of the most promising device, estimating that the development is just at the beginning point
What's going to be amazing is how fast the upcoming A15-based Exynos will be. Clock for clock, it's supposed to be 20% faster than A9 cores. And they will be shipping at 2Ghz!
I expect that the A15 Exynos dual-core will outperform a quad-core Tegra 3 and use a lot less power doing it. And it'll be cheaper to make.
Croak said:
I expect that the A15 Exynos dual-core will outperform a quad-core Tegra 3 and use a lot less power doing it. And it'll be cheaper to make.
Click to expand...
Click to collapse
One of the 3 can not be right. If it outperform a quadcore hardly will be needing less power. If it needs less power hardly will outperform quadcores and at end if both are true then no way to be cheaper!!!!
From quadlogic!!!
Sent from my GT-N7000 using XDA App