Related
So I've gotten anywhere between 2.5 to 5.1 MFLOPS using various ROMS and have yet to be able to notice something incredibly different.
710...768...806 - What does it matter? What program other than Linpack shows a sizable difference? Sure, maybe things open quicker? What am I missing here?
I read all this about achieving high MFLOPS and OC Kernels yet I still can't achieve smooth game play on 16 bit emulator on my phone with 5 MFLOPS.
MFLOPS mean jack when there is little way to observe the difference.
Carreno43 said:
So I've gotten anywhere between 2.5 to 5.1 MFLOPS using various ROMS and have yet to be able to notice something incredibly different.
710...768...806 - What does it matter? What program other than Linpack shows a sizable difference? Sure, maybe things open quicker? What am I missing here?
I read all this about achieving high MFLOPS and OC Kernels yet I still can't achieve smooth game play on 16 bit emulator on my phone with 5 MFLOPS.
MFLOPS mean jack when there is little way to observe the difference.
Click to expand...
Click to collapse
Linpack MFLOPS - measures the floating point performance of your phone.
710...768...806 - refers to CPU frequencies
increasing the CPU frequency should equate to better general-case performance, including things opening quicker as you mention, but also other types of general snappiness like moving between screens and so forth.
"I read all this about achieving high MFLOPS and OC Kernels yet I still can't achieve smooth game play on 16 bit emulator on my phone with 5 MFLOPS." - This may have less to do with the performance of your phone and more to do with the emulator itself. Emulation is a surprisingly CPU intensive operation, especially if the emulater isn't well written. Rather than looking a ton into overclocking and JIT, etc, maybe you ought to look for a better piece of software.
Yea,
I've tested most emulators. Wish there was an Atari emulator!
Thanks for the response.
Carreno43 said:
Yea,
I've tested most emulators. Wish there was an Atari emulator!
Thanks for the response.
Click to expand...
Click to collapse
I have run roms with 5.1 MFLOPS and now am running a rom that gets 3. I can honestly say I see no difference.
Spencer_Moore said:
I have run roms with 5.1 MFLOPS and now am running a rom that gets 3. I can honestly say I see no difference.
Click to expand...
Click to collapse
I can see a difference... in battery life! Lolz
g00gl3 said:
I can see a difference... in battery life! Lolz
Click to expand...
Click to collapse
Haha Awesome
it looks like to me that everyone is look at the wrong things.
for example:
I am running a Tom that is getting on a average of 4.9 mflops.
I get smoother screen changes....
streaming videos online is so much faster compared to a 3.0 mflop rom. ...
tubetube and other....... websites.
to me everything I do is faster...
I.don't play game on my phone so I don't know how that is.... but everythng else I do is very much faster.
I love high mflop roms...
I have notice about mflops is that it matters about the kernal that u use.
Isn't it true that the MSM7201 in our phones is already overclocked to get to 528mhz as it is? I see a lot of different places saying Qualcomm chips in general are just not worth overclocking... and since our chip is factory overclocked to begin with... just seems like we're pushing the already-pushed here. But the way this board goes crazy for overclocking... it's contradictory. I don't know what to think, cause I've run Linpack myself and gotten ~4.9 with JIT + OC versus ~2.5 without... but I'm with the OP on this one... only difference I'm seeing is my battery draining faster and my phone getting physically hotter.
xatch said:
Isn't it true that the MSM7201 in our phones is already overclocked to get to 528mhz as it is? I see a lot of different places saying Qualcomm chips in general are just not worth overclocking... and since our chip is factory overclocked to begin with... just seems like we're pushing the already-pushed here. But the way this board goes crazy for overclocking... it's contradictory. I don't know what to think, cause I've run Linpack myself and gotten ~4.9 with JIT + OC versus ~2.5 without... but I'm with the OP on this one... only difference I'm seeing is my battery draining faster and my phone getting physically hotter.
Click to expand...
Click to collapse
I have OC and JIT and getting about 5.1 mflops and haven't had worse battery life or a hotter phone. It could be the battery I'm using but meh (got a replacement one that's 2000 mAh) but I got worse battery life on leak 2.1 than with the rom I'm using now that has OC, JIT, LWP, etc. I can go about 8 hours with heavy texting, moderate internet usage and my lwp's running and it only goes to about 65%
so OC and Jit don't make that big of a difference in gameplay?
Sent from my Eris using the XDA mobile application powered by Tapatalk
What the OP and all the respondents are noting is frankly quite typical of what happens when performance tuning focuses on a single benchmark: the results obtained are essentially meaningless for different kinds of activities on the same device.
That's because there's a whole chain of dependencies that are specific to a given task, any number of which could become the rate-limiting factor; and a different task on the device will have a different set of dependencies and therefore different rate-limiting behaviors.
For instance, let's take writing to an SD card as an example: there's really no way that OC'ing will speed that up in a measurable way - because the CPU isn't the rate limiting factor.
That Linpack benchmark measures floating-point performance using a software library (as the Eris has no hardware FP capability). Most of the apps on the phone do very little FP work at all. But, it's not a bad test of CPU speed, because it performs no I/O. It also may not be very memory bandwidth intensive, either (if the problems it works on stays in the uP cache and there are few page faults).
OTOH, a game emulator needs to write to the graphics display (at a minimum) and possibly also do read I/O from flash.
Different task, different results. Sometimes things can be improved by hardware or firmware; sometimes the software itself needs to be improved.
bftb0
im sorry, but could you just answer in plain english
Sent from my Eris using the XDA mobile application powered by Tapatalk
TheSonicEmerald said:
im sorry, but could you just answer in plain english
Sent from my Eris using the XDA mobile application powered by Tapatalk
Click to expand...
Click to collapse
Lima beans bad.
Pork good.
Slow phone bad.
Fast phone good.
bftb0
Thanks for my laugh of the day on that one.
What I'm trying to get at is -
I should be able to play, at the basic level, Sonic or Mario - Without issues.
At the very least
I prefer roms over market games any day (Sonic, Mario, Zelda, DK-Country) and it cripples the phone, at least in my view, that I cannot enjoy the fruits of old games.
Although, I was able to find some old Atari games - which, thankfully, work without stuttering.
So I'd been using JIT on my Evil Eris ROM (and I'd used it with a few others) and I was getting scores of 5.1~ mflops. Here are some things I noticed.
While I had first enabled JIT it was glitch free. I can safely say at this point that any "speed increase" I'd noticed was likely placebo or at least admissible.
JIT is amazing, but it's amazing for Froyo. The way a JIT compiler works is that it precompiles information into code that is more easily read, causing slight lag to begin and increasing overall performance. The thing they had to do with froyo is optimize just about everything to stop the lag and to make JIT worthwhile.
The other reason is that despite removing sense and eventually using the Universal JIT and basically doing everything to make it stable JIT began to show problems. It was absolutely stable for a very long time but if I left my phone on for a few days or I started installing new programs/ a new launcher it would get glitchy and laggy. This likely goes back to what I said about JIT really being much better for Froyo. It came to the point that I felt slower than before I rooted.
Now that I'm not on JIT I can also overclock higher. Before I was limited to 768mhz and if I went over it was an instant crash. Now with JIT removed I'm up to 787, I can do 806 but it only lasts a few minutes before it freezes up but it's proof to the concept that I can overclock much higher. I get consistently 3.4~ mflops.
I do realize that I while on JIT I got a large increase in mflops but in my experience I am performing better without it. I think that when Froyo ROM's come out JIT will be much better implemented.
tl;dr: JIT ended up causing lag, I can overclock higher without it, I'm just as fast/faster without it.
note: For something like a game JIT may help you much more than just scrolling around since the game will be precompiled and then run more smoothly but in my experience (with Zenonia mostly) there was no improvement and occasionally I would in fact lag more.
Just thought I should say this to everyone.
Kind a sounds like steroids, ya you get muscle and look cool but there are side effects
Yeah and JIT was made for Froyo, not for the other ones. Froyo was heavily optimized, they spent a ton of time just making everything work as well as possible so that the JIT wouldn't actually be detrimental. So while JIT does increase MFLOPS and performance it can cause a lot of bugs and slow downs because it's trying to compile information that isn't optimized for it.
Hm. I think I'll try disabling JIT for a while too. I might prefer the quicker load time over any supposed increase in performance. I haven't had much buggy behavior, but I have had some restarts and wonder if they're caused by JIT or just overclocking in general.
I've found overclocking surprisingly stable without JIT. It's incredible that I can overclock to 806mhz now with glitches and 787 is absolutely stable...
Looking back I realize now that JIT had slowly degraded, I started off overclocking without it to 787 and as I turned it on it began to get very glitchy with 787, eventually freezing immediately with it. I wouldn't be surprised if the restarts you're having are due to JIT in combination with overclocking.
good post +1.
the devs don't say anything about this in their threads, thats if they even know it or have experienced it themselves so i'm not blaming anyone.
good to know so i can at least eval it for myself. thanks.
I think JIT might have been interfering with my Swype. Maybe it's doing too much "just in time" compiling and not enough "ahead of time" compiling. But anyway, yesterday I switched back and forth several times between using JIT and not using JIT, but didn't change anything else, and I definitely noticed that Swype was more responsive without JIT. With JIT enabled, I kept having to retype things because my tracing path would skip over parts; it was choppy. Without JIT it's smooth as silk and therefore accurate again.
I think a lot of the problem is that apps also aren't made to use JIT yet, all of the developers made apps for non-JIT phones first... you see a LOT of problems in phones that have background apps (setcpu, autokillers, etc) when they have JIT enabled. I think anyone who read up on JIT when it was announced will see mentions of the google dev's saying "We really really streamlined the phone to make use of the JIT".... obviously they're remarking that JIT without streamlining and "light" OS is a bit of a waste. JIT on my Froyo ROM (CM6) is very much more stable. If anyone's interested I did some benchmarks for JIT here:
http://forum.xda-developers.com/showthread.php?t=726038
Found this in the Hero forums, and it obviously applies to the Eris as well:
illogic6 said:
http://forum.cyanogenmod.com/topic/880-jit-wont-make-your-phone-super-fast/page__p__7910?#entry7910
Notice this portion of the post: "CPU intensive tasks get faster, but at the cost of RAM."
The way Darchstar explains it is that our phones cannot sacrifice that RAM once our phones have been bogged down with installed apps. He doesn't plan to include JIT by default when CM 6 goes final for HeroC and that's why.
Click to expand...
Click to collapse
Yes well you can decrease or increase the amount of RAM that JIT will use (heap size) and increase stability or decrease it. I understand why it's default off, it makes sense especially since a lot of programs aren't made to use it *cough pandora cough*
I'm running Cyanogenmod 6 froyo, did you test the effects of turning on and off JIT on this ROM?
Yes, like I said JIT was literally made for Froyo ROMs, and you can quite clearly see that when you hear the devs who created it talk about it in google conferences and such, they had to do a lot to get JIT to work well without creating huge slowdown. Compilers are almost ALWAYS a bad idea because of their nature and because if information isn't made for the compiler it ends up doing more harm than good.
Froyo is streamlined to work with JIT without causing slowdown so I would highly recommend using the JIT compiler with Froyo.
Hungry Man said:
Yes, like I said JIT was literally made for Froyo ROMs, and you can quite clearly see that when you hear the devs who created it talk about it in google conferences and such, they had to do a lot to get JIT to work well without creating huge slowdown. Compilers are almost ALWAYS a bad idea because of their nature and because if information isn't made for the compiler it ends up doing more harm than good.
Froyo is streamlined to work with JIT without causing slowdown so I would highly recommend using the JIT compiler with Froyo.
Click to expand...
Click to collapse
What about the Compcahe? Idk what it even does ahah
I turned JIT and surface dithering off. Phone runs smooth and no force closes.
surgeon0214 said:
What about the Compcahe? Idk what it even does ahah
Click to expand...
Click to collapse
Compcache can speed some things up. I enabled it on a SenseUI ROM a while ago and had TERRIBLE, crashing results until I turned it off, at random.
I can confirm that JIT's make Swype crash sometimes. I can do a 100 fast-swype word experiment (which, I do a lot of 'texting', so this experiment happens every day), on days I have JIT on, Swype randomly disappears. (it is set to restart itself after 5.0 seconds, so don't worry). With JIT off, it only does its normal random crashing (when I accidentally hit the Swype key while it's thinking of what words I just swyped, for example -- just the known crash bugs).
JIT day was yesterday, for example (I'm experimenting with JITs because of this thread) - and Swype crashed more than 10 times (I stopped counting by the time I was at the bar for two hours, -- don't drink and swype!)
Non-JIT day was today, and Swype has not crashed 1 time (despite writing a few very long emails in a passenger seat and some serious texting).
Monday is a JIT day, and I'm pretty confident based on my findings, that these have a huge impact on Swype.
I'm just using SenseUI for the weekend (for reliability/etc, because I'm 'On Call' for work), but if my findings prove my theory WRONG, I'll edit this post and note that.
But in general, yes, I agree, JITs mess with Swype. Not to a point of usability, but to a point that, the late person might think Swype is unreliable (unknowingly).
Compcache is your processor compressing the information that is stored in your RAM to increase the total amount of RAM that you can use per program. This can increase performance by letting your programs use more RAM but it can also decrease performance by increasing the time that the CPU has to work on compression. For the Eris I don't suggest Compcache unless you're overclocking to at least 710mhz and if you're on a Froyo ROM I don't suggest it at all since you should have plenty of RAM already.
What VM Heap Size are you using pkopalek? Smaller heap sizes are more unstable, you may find swype is more stable at 24m if you're at 12 or 16. 24m is the point where sense becomes stable (sorta) with JIT.
Hi, my name is Davis Woody. I'm a mechanical engineering student at UTC in Chattanooga, TN. I will be putting together this guide over the next few weeks. I am trying to be as scientific as possible with these scores so I will adhere to several rules for the tests. I would also love to test network strength, battery life, file transfer speed, and network speed but there are simply too many confounding variables. Can't wait to populate this thread.
My testing standards
1. I will allow each ROM to settle over the period of 12-24 hours before testing.
2. Identical applications will be installed on the different ROMs.
3. Identical phone settings (GPS, network, wifi, ect.) will be used during the tests
4. I will test 10 times at 1180mhz and 10 times at highest stable OC. I will note this maximum frequency in the data.
5. I will use a sample size of 10 tests to compute the sample average and standard deviation.
EDIT
If you would like for me to test a specific ROM and kernel combination, please feel free to ask. No request will be ignored.
It has been brought to my attention that Quadrant is not completely accurate for dual core. I am testing with CF Bench as well, the numbers seem to have a strong correlation.
ROM: VirusROM RC1 AOSP 1.13.651.7 2.3.3
Note: This ROM is so tight. It's about as close to CM as I've seen for the 3d. CF Bench is low for this ROM. My lowest quadrant was 3148. :O
Kernel: RCMix 2.6.35.10
Clock Freq.: 1836
CF Bench (mean score, std. dev)
Native: 7610.6, 123.80
Java: 2558.0, 66.50
Total: 4588.4, 82.80
Quadrant Standard (mean score, std. dev)
3650.3, 193.40
Underclock 1188mhz quad. score
2318, 62.64
ROM: Synergy ROM rev 318 w/ Sense 2.3.3
Note: Please note that I had disabled the EXT4 partitioning which can increase quad. scores. This is the fastest Sense ROM I've tested so far. Very stable w/ revision 318.
Kernel: RCMix 2.6.35.10
Clock Freq.: 1836
CF Bench (mean score, std. dev)
Native: 44202, 38.64
Java: 1791.6, 122.90
Total: 5555.4, 66.11
Quadrant Standard (mean score, std. dev)
3417.5, 245.55
Underclock 1188mhz quad. score
2215.2, 91.95
Underclock 1188mhz CF score
Native: 8569.0, 335.85
Java: 2150.3, 576.10
Total: 4660.5, 479.10
ROM: Showdown ROM 1 w/ Sense 2.3.3
Kernel: 2.6.35.14 Helicopter
Clock Freq.: 1836
CF Bench (mean score, std. dev)
Native: 11783, 650
Java: 2825.0, 675.72
Total: 6398.5, 592.3
Quadrant Standard (mean score, std. dev)
2920.3, 210.58
Underclock 1188mhz quad. score
2026.6, 128.68
ROM: Supra 1.5 AOSP 2.3.3
Note: Great ROM but it has its share of issues.
Kernel: SilverNeedles no lights 1.1.3
Clock Freq.: 1728
CF Bench (mean score, std. dev)
Native: 11341, 174.15
Java: 3270.2, 66.92
Total: 6498.2, 106.00
Quadrant Standard (mean score, std. dev)
2902.1, 205.72
Underclock 1188mhz quad. score
2125, 51.64
ROM: viperROM RC1.3 w/ Sense 2.3.3
Note: Great ROM for a daily driver.
Kernel: SilverNeedles no lights 1.1.3
Clock Freq.: 1728
CF Bench (mean score, std. dev)
Native: 11299.0, 117.80
Java: 3279.0, 40.0
Total: 6487.0, 65.71
Quadrant Standard (mean score, std. dev)
2798.2, 159.93
Underclock 1188mhz quad. score
2168.2, 84.35
ROM: Supra ROM 1.5 w/ Sense 2.3.3
Note: Great ROM for a daily driver.
Kernel: SilverNeedles no lights 1.1.3
Clock Freq.: 1728
CF Bench (mean score, std. dev)
Native: 11042, 239.4
Java: 3251, 152.2
Total: 6475, 125.4
Quadrant Standard (mean score, std. dev)
2773.33, 186.17
Underclock 1188mhz quad. score
2061.5,213.34
ROM: Douchless ROM w/ Sense 2.3.3
Note: Tether issues
Kernel: SilverNeedles 1.0
Clock Freq.: 1728
CF Bench (mean score, std. dev)
Native: 10787,192.2
Java: 1816,87.87
Total: 5404.8,120.86
Quadrant Standard (mean score, std. dev)
2647.8, 363.76
Underclock 1188mhz quad. score
2058.8,57.85
Underclock 1188mhz CF score
Native: 8748.4,118.7
Java: 1360.8,79.0
Total: 4321,95.2
ROM: SuperShooter3D(Bl00dy3dition)v.2 w/ Sense 2.3.3
Note:I'm not sure if this was a "clean" install. Got some FCs
Kernel: SilverNeedles no lights 1.1.3
Clock Freq.: 1728
Quadrant Standard (mean score, std. dev)
2239.8, 170.51
Underclock 1188mhz quad. score
1397.1, 269.10
chillfancy said:
Hi, my name is Davis Woody. I'm a mechanical engineering student at UTC in Chattanooga, TN. I will be putting together this guide over the next few weeks. I am trying to be as scientific as possible with these scores so I will adhere to several rules for the tests. I would also love to test network strength, battery life, file transfer speed, and network speed but there are simply too many confounding variables. Can't wait to populate this thread.
My testing standards
1. I will allow each ROM to settle over the period of 12-24 hours before testing.
2. Identical applications will be installed on the different ROMs.
3. Identical phone settings (GPS, network, wifi, ect.) will be used during the tests
4. Overclock speed will be at what I deem to be the highest stable level for each ROM. I will note this frequency in the data.
5. I will use a sample size of 10 tests to compute the sample average and standard deviation.
EDIT
If you would like for me to test a specific ROM and kernel combination, please feel free to ask. No request will be ignored.
It has been brought to my attention that Quadrant is not completely accurate for dual core. I am considering testing with CF bench as well.
Click to expand...
Click to collapse
you may as well leave off quadrant. its easily altered and not really accurate.
Seriously though, quardrant scores don't mean a much. The rom that claims to have the highest quads was not the fastest, but it was the most buggy (in my experience). And the rom I found to be the smoothest/fastest and lag-free had a quadscore 100 below stock lol.
chillfancy said:
...I am trying to be as scientific as possible with these scores so I will adhere to several rules for the tests. I would also love to test network strength, battery life, file transfer speed, and network speed but there are simply too many confounding variables...
My testing standards
1.
2.
3.
4. Overclock speed will be at what I deem to be the highest stable level for each ROM. I will note this frequency in the data.
Click to expand...
Click to collapse
Davis, If you're going to be scientific you've got to compare apples to apples, so ANY variables are unacceptable, including clock speed. To be completely objective, you need to compare all roms at 1.2Ghz (1188mhz) ... otherwise, any comparison is trash. The fact that a given kernel/rom combination can run at 2Ghz is nice, but it must be tested at 1.2Ghz for any comparison of battery life, data transfer rates, benchmarks, etc., to be meaningful.
Your goal is noble, but your method needs to be correct.
Good luck!
Success100 said:
you may as well leave off quadrant. its easily altered and not really accurate.
Click to expand...
Click to collapse
I don't think quadrant is completely useless. It does run full 2d and 3d to test the rendering performance.
oldjackbob said:
Davis, If you're going to be scientific you've got to compare apples to apples, so ANY variables are unacceptable, including clock speed. To be completely objective, you need to compare all roms at 1.2Ghz (1188mhz) ... otherwise, any comparison is trash. The fact that a given kernel/rom combination can run at 2Ghz is nice, but it must be tested at 1.2Ghz for any comparison of battery life, data transfer rates, benchmarks, etc., to be meaningful.
Your goal is noble, but your method needs to be correct.
Good luck!
Click to expand...
Click to collapse
I realized that testing at stock is important to really compare ROMs ,but with that said, a ROM that can run 1.8 stable is going to be a faster daily driver than one that crashes at 1.5. We'll see how the testing goes. And I'm not testing signal, battery life, or data rates.
chillfancy said:
I realized that testing at stock is important to really compare ROMs ,but with that said, a ROM that can run 1.8 stable is going to be a faster daily driver than one that crashes at 1.5. We'll see how the testing goes. And I'm not testing signal, battery life, or data rates.
Click to expand...
Click to collapse
Um, then I'm confused ... what exactly is it that you're testing? Are you just wanting to see which rom/kernel can survive the highest clock speed, regardless of battery life, or temperature (I guess?), or efficiency (however you want to define that)? If that's all you're doing, then why even bother with any "settling in" period at all? Heck, just keep cranking up the clock speed until the machine freezes or reboots, and report back to us, lol!
Carry on...
Quadrant is flawed to begin with. And you are testing with a flawed testing software for accurate results? Try this, run quadrant 5 times...you will see one of those scores will win, the others will vary widly. And this is on your phone!... Give me a break.
life64x said:
Quadrant is flawed to begin with. And you are testing with a flawed testing software for accurate results? Try this, run quadrant 5 times...you will see one of those scores will win, the others will vary widly. And this is on your phone!... Give me a break.
Click to expand...
Click to collapse
The sample size is 10. I'm running 2 different softwares. It's not just picking the highest outlier.
More power, but all your doing is spreading the myth of quadrant being the end all for performance. Not counting what apps are running, any tweaks each rom has..even stock and memory trash cleaning mechanisms, what natural speed of the phone is depending on the bus...yea it sure will be accurate....
So much venom in the forums... Geez.
So quadrants sucks. Oh well, as long as this guy isn't using the I/O tweak that shoots quadrants sky high, does several passes and than takes the average etc.
It sounds like he has a good foundation.
Except the whole overclocking thing. That kindof destroys your whole foundation and everything should be run at stock speed. Your taking all these steps to remove variants and then throwing a giant one in?
"Gentleman. We are going to study the effect of gravity on these rabbits. They will be weighted down at the base of there paws to keep them upright and from floating around.... Except this one. He will be stabbed, gutted and have the weight implanted into his abodmen. For funsies"
Just drop the overclocking part :-D
chillfancy said:
Hi, my name is Davis Woody. I'm a mechanical engineering student at UTC in Chattanooga, TN. I will be putting together this guide over the next few weeks. I am trying to be as scientific as possible with these scores so I will adhere to several rules for the tests. I would also love to test network strength, battery life, file transfer speed, and network speed but there are simply too many confounding variables. Can't wait to populate this thread.
My testing standards
1. I will allow each ROM to settle over the period of 12-24 hours before testing.
2. Identical applications will be installed on the different ROMs.
3. Identical phone settings (GPS, network, wifi, ect.) will be used during the tests
4. Overclock speed will be at what I deem to be the highest stable level for each ROM. I will note this frequency in the data.
5. I will use a sample size of 10 tests to compute the sample average and standard deviation.
Click to expand...
Click to collapse
Fantastic Idea and thanks for this! I'll be looking forward to seeing where they stand.... may I make a suggestion and start the testing after a reboot? Some roms performance varies and that'd make them all equal say .. hit the power button... wait a minute then start the ten tests.
Either way this will be a great reference for test jockeys
Although quadrant is not 100% reliable in determining how fast roms are but it will be interesting to see the results of each rom.
I'm just impressed that you are taking on such a tedious and time consuming test. Best of luck!
felacio said:
Except the whole overclocking thing. That kindof destroys your whole foundation and everything should be run at stock speed. Your taking all these steps to remove variants and then throwing a giant one in?
Click to expand...
Click to collapse
I like the overclocking test. Not including it would be like reading a car review and bashing them for testing a top speed higher than 75 MPH, since that's pretty much the highest speed limit you'll see anywhere.
Sure, you can state the 0-60 time and compare those, but who's gonna be hurt by knowing the top speed?
I want to see how well my rom will work when I crank that stuff up!
Assuming however, he does provide data for both the 1188 speed and the OC speed. Otherwise, he is just being silly. Lol.
BlaydeX15 said:
I like the overclocking test. Not including it would be like reading a car review and bashing them for testing a top speed higher than 75 MPH, since that's pretty much the highest speed limit you'll see anywhere.
Sure, you can state the 0-60 time and compare those, but who's gonna be hurt by knowing the top speed?
I want to see how well my rom will work when I crank that stuff up!
Assuming however, he does provide data for both the 1188 speed and the OC speed. Otherwise, he is just being silly. Lol.
Click to expand...
Click to collapse
Yeah... I don't agree with your analogy. But yeah, the OC should stay. But at the beginning it was forefront and he later decided to add the stock speed.
Your analogy fails because 75mph is 75mph. You can't go faster than 75 without going 76. And so on..
On the phone though. 1.2ghz can swing widely with how efficient the phone runs in that speed. It can be faster or slower depending on how the is is running. Look at Sense vs AOSP.. omg the difference! So the analogy that would work in that case is more "at 75mph. This car guzzles 20mpg. But this one does the same and at 30mpg" but at the end of that. Append ".. but the first car can go 180mph without an issue!"
Use linpack
Sent from my HTC EVO 3d
Root: revolutionary
Recovery: Twrp cwm 4
ROM: Synergy rls318 (I want aosp)
oldjackbob said:
...what exactly is it that you're testing?
Click to expand...
Click to collapse
Still waiting for an answer to this question.
Sorry, but I'm just not seeing the point to this whole effort.
mikedavis120 said:
Use linpack
Sent from my HTC EVO 3d
Root: revolutionary
Recovery: Twrp cwm 4
ROM: Synergy rls318 (I want aosp)
Click to expand...
Click to collapse
I'll consider it but CF Bench measures MFlops too so it might be a bit redundant.. It is taking me about an hour to collect and process the data on a single ROM already.
oldjackbob said:
Still waiting for an answer to this question.
Sorry, but I'm just not seeing the point to this whole effort.
Click to expand...
Click to collapse
The goal is to use testing standards, benchmarking software, and statistics to describe how the different ROMs score. I understand it isn't going to prove that once ROM is "better" than another but it will be very useful for others.
Cool I like it! Stock and OC is key! Also I think the max overclock is good too, if you find a rom that's stable at 2ghz(some people have done it) it'll help point others to those high speed roms. If you do come across one that's stable at 2ghz I would like to still see a 1.8ghz test.
For a battery test could you play a 5 or 10 min vid clip after a fresh reboot and a charge, then just record the percent dropped and voltage. Or just record the amount of batter drained before and after your bench marks. Using a app like battery indicator would be easy too.
I'm debating on taking back my device due to getting such low stock Quadrant scores. Around 850 stock on average.
What's everyone else getting? Am I the only one with such low scores? I appreciate any feedback!
Benchmarks are worthless in general, Quadrant is incredibly worthless because it doesn't break down the scores and is EASILY gamed.
For example, disabling per-file fsync on an ext4 system gains you 600 points or so instantly.
Enabling Stagefright on Samsungs gets you 500-1000 points but breaks media playback.
I'm not talking about using mods at all though. I'm just saying I'm getting half of what I've seen as the average STOCK 1Ghz score for this device. My 800 @ 1Ghz vs. others 1600-ish @ 1GHz. I just want to know if anyone else is getting around my scores (800ish) without mods.
What other devices?
I get about 1300-1400 stock...
Hi,
I just ran it on my (US/rooted) SGP5, and got 947.
I noticed that on the comparison graph, there's Samsung Galaxy S, that had slightly lower number ...
Jim
P.S. FYI, I used Fast Boot, to kill all processes, just before running the Quadrant test. I'm not sure if that affected anything.
BTW, Quadrant gives HEAVY weighting to file I/O scores - so Voodoo Lagfix-enabled devices and native-ext4 ones will smoke RFS devices.
tcb4 said:
I get about 1300-1400 stock...
Click to expand...
Click to collapse
Hi,
Wow! Do you know how you got such scores? What have you done to your device (and what is it)?
As I posted, I just got 947 on my SGP5. It's stock, other than being rooted.
Jim
Entropy,
Your completely missing my point... Here it is again. "My SGP5 is getting around 850~average scores on Quadrant stock, as in, in the state it was when I took it out of the box. Reviewers and other owners of this device are getting much higher quadrant scores with their STOCK devices as well.
When I overclock to 1.3GHz, I get around 1400 average while other SGP5's get around 2k+ at 1.3GHz. I just want to see how many are also experiencing such low scores as I am.
As I said, many of the reviewers online say that the SGP5 gets around 1400-1600 with Quadrant without any modifications or rooting. I get around 850. Please stop with the "Well Quadrant isn't accurate because, if you mod it like this.. you get this", I'll say it again, My. Device. Is. Stock. It's not modified, I didn't do anything to it to boost the score or lower it.
You're missing my point.
Quadrant is unreliable and inconsistent as hell. It can vary by hundreds of points from run to run, it's easily gamed, it will do weird unpredictable things if you look at it the wrong way. Plain and simple, it sucks and its results are utterly and completely worthless.
You ARE benchmarking with the CPU governor set to performance, right? Benchmarks with the governor in play are even more worthless and variable.
I just saw 600 points difference between two runs with the same kernel - 992 once, 1619 the next.
I've gotten consistent bench marking results every time, consistently low, but consistant.... Do you own a SGP and can say this from experience or just because you have a similar Android device? I actually own a SGP5 and haven't seen the "992 once, 1619 next" at all... I've never gotten above 1619 or ever had a jump of more than 100 on the same frequency.
Unless your testing this for yourself on the device in question, please stop once again...
Yes, I own an SGP 5. I unboxed it last night. I actually ran Quadrant again, 1700 the next time around. Just more evidence that it's crappy and meaningless. Benchmarks in general suck, but Quadrant is especially bad. At least Antutu is fairly consistent AND breaks out scores by category.
It's consistent with every other device I've ever owned - run-to-run variances of a few hundred points.
Note that over time, your scores will go down on this device because RFS sucks and becomes slower as it's used, and as I said, I/O performance is weighted HEAVILY.
RFS sucking is a known problem with old Samsungs, that's why Voodoo Lagfix exists.
Returning a device based on Quadrant scores is stupid on an epic scale given what a horrifically worthless and inconsistent benchmark it is...
Around 1300.
Heh, anyways.
What's Voodoo lagfix? And RTS? If you don't mind me asking, not familiar with those two. And so far it performs pretty well @ 1.5GHz with 2000+ quad scores, downloading Antutu now.
ZaIINN said:
Heh, anyways.
What's Voodoo lagfix? And RTS? If you don't mind me asking, not familiar with those two. And so far it performs pretty well @ 1.5GHz with 2000+ quad scores, downloading Antutu now.
Click to expand...
Click to collapse
RFS is ? File System - it's some special Samsung file system that just kinda sucks.
ext4 is the current standard Android filesystem format, which is MUCH better.
(think of it as being like NTFS vs FAT on a PC)
Voodoo Lagfix is a set of initramfs scripts that automatically convert RFS partitions to ext4, improving read/write performance a lot.
Entropy512 said:
RFS is ? File System - it's some special Samsung file system that just kinda sucks.
ext4 is the current standard Android filesystem format, which is MUCH better.
(think of it as being like NTFS vs FAT on a PC)
Voodoo Lagfix is a set of initramfs scripts that automatically convert RFS partitions to ext4, improving read/write performance a lot.
Click to expand...
Click to collapse
Ah interesting, I looked into the Lagfix; Think it would work the same for the SGP as it does for the galaxy?
ZaIINN said:
Ah interesting, I looked into the Lagfix; Think it would work the same for the SGP as it does for the galaxy?
Click to expand...
Click to collapse
Yes - it's on my todo list once I get CWM to play nice.
ANTUTU BENCHMARKING :- The "OPEN" Discussion (11/12/12) 10125 BEST SCORE SO FAR
First things First
I HAVE THE EXTENDED BATTERY
Secondly
TO ME
Benchmarks DO gauge the ACTUAL Performance of MY phone (when pushed)
right your still reading
HELLO
BENCHMARKING
I've had my phone since day one of the uk release, and have flashed many many many roms, and kernels, and hboot's , experimenting with different oc programs, governors, frequencies and voltages, and ui in comparison to smoothness is greatly improved the higher the antutu score
REGARDLESS OF BATTERYLIFE.
that said.
the best concoction I cooked up, with the best ingredients from many experienced chefs (so far) is
Rom:- HTC Stock ICS LATEST
Kernel:-Yoda 10.1
Tweaks:- Pedja kernel tweak app v3.3.3/ Chainfire 3D/ Pimp My Rom
oc'd @ max and min 1.836 ghz
Uv @ 1312 mv
performance CPU Governor
I/0 Scheduler SIO @ 2048kb sd cache
No vsync
my score was]
10125
Please leave detailed Descriptions of
ROM
KERNEL
PATCHES
CPU GOVERNOR
I/O SCHEDULER
SD CLASS/SIZE
OV/UV CHANGES
THANKS IN ADVANCE
mOAr voltz
argument invalid, benchmark numbers does not equal performance. research how to trick benchmarking apps.
gav-collins1983 said:
first things first, I HAVE THE EXTENDED BATTERY. secondly, TO ME, benchmarks DO gauge the performance of MY phone, ......
right your still reading, hello,
benchmarking, I've had my phone since day one of the uk release, and have flashed many many many roms, and kernels, experimenting with different oc programs, governors, frequencies and voltages, and ui in comparison to smoothness is greatly improved the higher the antutu score, regardless of battery life.
that said.
the best concoction I cooked up, with the best ingredients from many experienced chefs (so far) is
leedroids rom v 5.3.0
leedroids kernel v4.1.0
system tuner pro 1.863 Max 192 min (added 0.25mv to 1.863 frequencie because it seemed to struggle)
and leedroids ondemand forced dual core patch
my score was 7004
rant over , lol
Click to expand...
Click to collapse
Jesus Christ is this what this site has came to?
bloodrain954 said:
argument invalid, benchmark numbers does not equal performance. research how to trick benchmarking apps.
Click to expand...
Click to collapse
it's not an argument, I clearly state, that this is to "ME" on "MY" phone,
also, I'm not trying to trick my benchmarking app, what I've seen over 28 weeks is a clear indication, to "ME" on "MY" phone , that ui is smoother and faster on higher benchmark scores, simple as that
Well duh, having an overclocked CPU results in higher benchmarks because the phone can complete things faster. This translates to a smoother UI.
The problem is that benchmarks are affected by so many things that might not reflect accurately the performance of the phone. And comparing different phones is useless because of skins and different hardware performing better in tests but worse in real life.
Faster CPU = higher benchmarks and smoother UI. That doesn't mean that higher benchmarks = smoother UI.
Sent from my HTC Evo 3D using XDA Premium App
SoraX64 said:
Well duh, having an overclocked CPU results in higher benchmarks because the phone can complete things faster. This translates to a smoother UI.
The problem is that benchmarks are affected by so many things that might not reflect accurately the performance of the phone. And comparing different phones is useless because of skins and different hardware performing better in tests but worse in real life.
Faster CPU = higher benchmarks and smoother UI. That doesn't mean that higher benchmarks = smoother UI.
Sent from my HTC Evo 3D using XDA Premium App
Click to expand...
Click to collapse
it does to me from what I've seen, I've made that white clear,
what's with all the hating on evo 3d , we never had this in the desire hd forums.
gav-collins1983 said:
it does to me from what I've seen, I've made that white clear,
what's with all the hating on evo 3d , we never had this in the desire hd forums.
Click to expand...
Click to collapse
Lololol your not in Kansas anymore
Sent from my SPH-D710 using XDA App
one time, at band camp...
sent from Evo
SoraX64 said:
Well duh, having an overclocked CPU results in higher benchmarks because the phone can complete things faster. This translates to a smoother UI.
The problem is that benchmarks are affected by so many things that might not reflect accurately the performance of the phone. And comparing different phones is useless because of skins and different hardware performing better in tests but worse in real life.
Faster CPU = higher benchmarks and smoother UI. That doesn't mean that higher benchmarks = smoother UI.
Sent from my HTC Evo 3D using XDA Premium App
Click to expand...
Click to collapse
No, not, duhhhhh, overclocking also causes instability in some cases, more so if incorrect/not optimized voltages per frequency and governor in kernel, and smooth ness is affected
gav-collins1983 said:
first things first, I HAVE THE EXTENDED BATTERY. secondly, TO ME, benchmarks DO gauge the performance of MY phone, ......
right your still reading, hello,
benchmarking, I've had my phone since day one of the uk release, and have flashed many many many roms, and kernels, experimenting with different oc programs, governors, frequencies and voltages, and ui in comparison to smoothness is greatly improved the higher the antutu score, regardless of battery life.
that said.
the best concoction I cooked up, with the best ingredients from many experienced chefs (so far) is
leedroids rom v 5.3.0
leedroids kernel v4.1.0
system tuner pro 1.863 Max 192 min (added 0.25mv to 1.863 frequencie because it seemed to struggle)
and leedroids ondemand forced dual core patch
my score was 7004
rant over , lol
Click to expand...
Click to collapse
People will hate me for this, and I waited a long time to post this on this thread, but I just have to do it.
Cool Story Bro.
tgruendler said:
People will hate me for this, and I waited a long time to post this on this thread, but I just have to do it.
Cool Story Bro.
Click to expand...
Click to collapse
:thumbup:
From the Beast : Galaxy Note
gav-collins1983 said:
first things first, I HAVE THE EXTENDED BATTERY. secondly, TO ME, benchmarks DO gauge the performance of MY phone, ......
right your still reading, hello,
benchmarking, I've had my phone since day one of the uk release, and have flashed many many many roms, and kernels, experimenting with different oc programs, governors, frequencies and voltages, and ui in comparison to smoothness is greatly improved the higher the antutu score, regardless of battery life.
that said.
the best concoction I cooked up, with the best ingredients from many experienced chefs (so far) is
leedroids rom v 5.3.0
leedroids kernel v4.1.0
system tuner pro 1.863 Max 192 min (added 0.25mv to 1.863 frequencie because it seemed to struggle)
and leedroids ondemand forced dual core patch
my score was 7004
rant over , lol
Click to expand...
Click to collapse
7000 score in Antutu @ 1.86 GHz? That seems low... I right around the same score at 1.65GHz
gav-collins1983 said:
first things first, I HAVE THE EXTENDED BATTERY. secondly, TO ME, benchmarks DO gauge the performance of MY phone, ......
right your still reading, hello,
benchmarking, I've had my phone since day one of the uk release, and have flashed many many many roms, and kernels, experimenting with different oc programs, governors, frequencies and voltages, and ui in comparison to smoothness is greatly improved the higher the antutu score, regardless of battery life.
that said.
the best concoction I cooked up, with the best ingredients from many experienced chefs (so far) is
leedroids rom v 5.3.0
leedroids kernel v4.1.0
system tuner pro 1.863 Max 192 min (added 0.25mv to 1.863 frequencie because it seemed to struggle)
and leedroids ondemand forced dual core patch
my score was 7004
rant over , lol
Click to expand...
Click to collapse
*Sips his Earl Grey* You like high numbers? You clearly did not like high numbers when you tipped me for my "bedroom" services!
gav-collins1983 said:
No, not, duhhhhh, overclocking also causes instability in some cases, more so if incorrect/not optimized voltages per frequency and governor in kernel, and smooth ness is affected
Click to expand...
Click to collapse
Anybody who knows anything about overclocking knows to make sure it's stable..
Seriously, if you think your high quadrants means your phone outperforms the leaked scores of a quad core Tegra 3 phone by almost double, then you're in denial.
Let me put this in words you can understand.
Faster clock speed = faster graphics and calculations.
Faster clock speed = higher benches.
Higher benches != better performance. They're just linked.
Sure, if you're pulling over 4000 in quadrant, your device is performing quite well. But that doesn't make your Evo 3D a super phone twice as good as the rest of ours.
There are so many factors involved that simply saying "my quads are high, and that indicates performance" is ignorant and naive. Once you can explain to me how your phone "performs" better than a quad core device, then maybe I'll spare your thoughts my time.
Sent from my HTC Evo 3D using XDA Premium App
I can't wait to see where this thread goes from here. Also, my phone is completely usable, fast, smooth, and has amazing battery life and I just scored an 1100 on Antutu. Benchmarks are worthless except to raise your eEsteem. See my thread about GeekBench from a few nights ago for proof.
Sent from my PG86100 using Tapatalk
Yes, as I stated in the other thread, quadrant is simply the equivalent of measuring mobile phone penis size and nothing more.
From the Beast : Galaxy Note
7000? Pics or get out.
Twitter: @knowledge561
Blog: knowledgexswag.tumblr.com
my score was OVER 9000!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!
What! Over 9000! THERE'S NO WAY THAT CAN BE RIGHT!