How about we start an official database for all benchmarking we can do on the Fascinate? Here is every free benching app I could find. My phone is stock 2.1 for now. After FroYo I plan on overclocking the piss outta this beast!
System info:
SGH-I500 Fascinate
non-root 2.1update1 Eclair
OS: 2.6.29
BogoMips: 997.7
Bone stock outta the box. Sorry no screengrabs since I haven't rooted the phone.
*An3DBench - 6581
*Quadrant - 878
*Neocore - 55.9fps
*Linnpack - MFLOPS 8.181 / Time 10.025 sec / Norm Res 5.68
*DroidBooster - 21.7218% increased performance
*CPU Benchmark - 799ms
*GPU Benchmark - Absolute - 34155 Relative - 31999
*NBench -
INTEGER INDEX : 13.977
FLOATING-POINT INDEX : 1.395
MEMORY INDEX : 2.915
INTEGER INDEX : 3.990
FLOATING-POINT INDEX : 0.774
*NENA - 50.4fps
*Caffeine Mark labs with DroidBooster - Raw 12777, 4.5 Stars, Rank - 9
*FPS2d - 1000 iterations. avg-55fps stdev-4.97fps
*CaffeineMark - 1805
*RLBenchmark
SQLite - SCH-I500 (SDK:7)
Benchmark Results:
1000 INSERTs - 63.337 sec
25000 INSERTs in a transaction - 3.852 sec
25000 INSERTs into an indexed table in a transaction - 4.612 sec
100 SELECTs without an index - 0.211 sec
100 SELECTs on a string comparison - 0.086 sec
Creating an index - 2.82 sec
5000 SELECTs with an index - 2.116 sec
1000 UPDATEs without an index - 9.648 sec
25000 UPDATEs with an index - 8.047 sec
INSERTs from a SELECT - 7.55 sec
DELETE without an index - 5.85 sec
DELETE with an index - 6.041 sec
DROP TABLE - 6.866 sec
Overall - 121.036 sec
*BenchmarkPI - 2702ms
*Speed PI - Calculated to 1,000,000 digits: 659ms act.
*SunSpider JavaScript Benchmark 0.9.1 - Best Score = 12300.1ms +/-2% (Best browser I benched was Dolphin HD, believe it or not. I benched it along with Skyfire Beta(close behind at 13983.5ms+/-6.6%), the newest FireFox Beta(way slower score over 26600ms), Opera Mobile(not too far off at 14050.0ms +/-4.6%), and NetFront Life(dismal at over 36000ms). And at least on my machine, today, with my Un-rooted stock 2.1, Dolphin HD has the fastest java based score among them all in SunSpider.)
*Eyes+Chrome - 22 Asterisk**
**Including thread title.
I fail to see the point of this thread. A forum thread is most certainly not a database for storing benchmark data, and that data itself is useless without any context. Unless you're comparing multiple ROMs, of the SF to other phones, the numbers are meaningless, and I see no comparisons being made here.
Now, if you file that data away, and present it again later in comparison with the numbers from the same tests on the official Froyo ROM (once it's released), then I will be somewhat interested. ;-)
Sent from my SCH-I500 using XDA App
Haters...
Stock 2.1 Rooted with Voodoo lag fix
Quadrant: 1780
Instead of tearing down, why not contribute? Post up benchmarks of other ROMs.
Then we can compare.
I didn't say this thread was only for box stock SF devices.
My fascinates screen gave me a boner. That seems to be a common occurence. Slap that on there with benchmarks. BTW, not calling u a liar but I have NEVER seen neocore above 55.7. And I have a captivate as well,( well had). Odd. I'm all for it. I don't see why not, put up all the benchmarks from all the different ROMS.
Sent from my SCH-I500 using XDA App
benchmarks mean very little on this phone. you can get huge quadrant scores from one of the old overclock voodoo kernels but they will not stand up to the newer non oc kernels in everyday use.
Related
I did some testing. Here are numbers for those of you curious about performance. All were tested on the same Nexus One, more to come soon!
Paulobrian 2.2 /w cyano 2.6.34 kern @ 1113 Mhz
Set CPU
Short - 206 ms
Long - 345 ms
CPU Benchmark
660 ms 1113 mhz
Benchmark
Graphics
Total Graphics Score - 30.36
Draw Opacity Bitmap - 10.80 Mpixels per sec
Draw Transparent bitmap - 7.77 Mpixels per sec
CPU
Total CPU Score - 2491.91
MWIPS DP - 166.94
MWIPS SP - 187.62
MFLOPS DP - 19.03
MFLOPS SP - 26.79
VAX MIPS DP - 142.033
VAX MIPS SP - 141.58
Memory
Total Memory Score - 317.97
Copy Memory - 288.93 Mb/sec
Filesystem
Total Filesystem Score - 104.72
Creating 1000 empty files - 4.502 sec
Deleting 1000 empy files - 3.252 sec
Write 1M into file - 2.351 M/sec
Read 1M into file - 203.333 M/sec
Paulobrian 2.2 @ 998 MHZ
Set CPU
Short - 243 ms
Long - 378 ms
CPU Benchmark
750 ms 998 mhz
Benchmark
Graphics
Total Graphics Score - 28.53
Draw Opacity Bitmap - 10.14 Mpixels per sec
Draw Transparent bitmap - 7.30 Mpixels per sec
CPU
Total CPU Score - 2415.09
MWIPS DP - 157.48
MWIPS SP - 182.48
MFLOPS DP - 28.37
MFLOPS SP - 26.79
VAX MIPS DP - 138.70
VAX MIPS SP - 137.20
Memory
Total Memory Score - 367.81
Copy Memory - 334.22 Mb/sec
Filesystem
Total Filesystem Score - 116.88
Creating 1000 empty files - 5.95 sec
Deleting 1000 empy files - 4.393 sec
Write 1M into file - 2.59 M/sec
Read 1M into file - 232.56 M/sec
CM 5.0.7 2.1 @ 1113 Mhz
Set CPU
Short - 248
Long - 848ms
CPU Benchmark
855ms 1113 mhz
Benchmark
Graphics
Total Graphics Score - 31.505909
Draw Opacity Bitmap - 11.191257 Mpixels per sec
Draw Transparent bitmap - 8.078896 Mpixels per sec
CPU
Total CPU Score - 794.6353
MWIPS DP - 60.86
MWIPS SP - 60.75
MFLOPS DP - 9.26
MFLOPS SP - 10.01
VAX MIPS DP - 40.39
VAX MIPS SP - 39.51
Memory
Total Memory Score - 335.01
Copy Memory - 304.41 Mb/sec
Filesystem
Total Filesystem Score - 114.35
Creating 1000 empty files - 5.30 sec
Deleting 1000 empy files - 3.99 sec
Write 1M into file - 2.79 M/sec
Read 1M into file - 227.27 M/sec
I'm not getting how the graphics are worse on 2.2 over CM.
halorin said:
I'm not getting how the graphics are worse on 2.2 over CM.
Click to expand...
Click to collapse
Yea I found that kind of odd too...
Well, the 2.1 kernel is overclocked by about 10%, and the difference in total graphics score is about 11%, so this leads me to believe that graphics speeds would remain the same if these tests used the same clock speed.
Re: [Benchmarks] Paulobrian 2.2 vs CM 5.0.7 2.1
So basically 2.2 will have stock phones almost where we have all been for months with 2.1
deprecate said:
So basically 2.2 will have stock phones almost where we have all been for months with 2.1
Click to expand...
Click to collapse
It's funny you say that, but my coworker has a N1 unrooted and I will do something for him on his phone and you can definitely tell a difference. Mine is SOOO much smoother and faster...Once you go root, you never go back!
You should also consider that the 2.2 we have is not the full update and that Cy's is HEAVILY optimized and overclocked. This tells me that when we get the real 2.2 and optimize it, it will blow Cy's out of the water
Kind off a lousy comparison. Click 'em both to 998 min/max and then see what the differences are.
Mi|enko said:
Kind off a lousy comparison. Click 'em both to 998 min/max and then see what the differences are.
Click to expand...
Click to collapse
I wanted to just show the performance of the best of each rom.
It's an acknowledgment of cyanogen's excellent work, to be sure.
But if you want flash and tethering, similar performance all with a stock OS and kernel, well Froyo is the cat's ass.
tengtou said:
I wanted to just show the performance of the best of each rom.
Click to expand...
Click to collapse
You actually showing the performance of the kernel not the rom, and no even mentioning the kernels.
JCopernicus said:
You actually showing the performance of the kernel not the rom, and no even mentioning the kernels.
Click to expand...
Click to collapse
Im not sure what kernals they are, just the roms itself, and whichever kernals came with it...
custom374 said:
You should also consider that the 2.2 we have is not the full update and that Cy's is HEAVILY optimized and overclocked. This tells me that when we get the real 2.2 and optimize it, it will blow Cy's out of the water
Click to expand...
Click to collapse
Or perhaps, once Cy gets his hands on 2.2 (when published to aosp), we will see just how amazing froyo can be when unrestricted.
deprecate said:
Or perhaps, once Cy gets his hands on 2.2 (when published to aosp), we will see just how amazing froyo can be when unrestricted.
Click to expand...
Click to collapse
I cant wait!
A while back I posted about how I managed to get a ridiculous score of 2597 on Quadrant using Dameon87 and noobnl's ramdrive hack, and achieved the highest (and it's still standing) stock clock Quadrant score on SmartphoneBenchmarks.com.
After I posted the score, I complained about it on SmartphoneBenchmark's forums, and was informed by an admin that they were aware of the skewed scores in Quadrant and working on a benchmark tool of their own.
Well that tool was released and today I gave it a spin. Its called Smartbench 2010 and is currently free on the Android Market. It runs a series of tests and splits the results into a Productivity Index (CPU) and a Games Index (GPU). The Epic performs VERY well on this benchmark; I scored 1178 and 2610 in Productivity and Games respectfully, while the HTC G2 scores 1045 and 1396.
But here's the real test; we all know that phones with the RFS file system perform very poorly on Quadrant... the benchmark doesn't get along with the file system well and the result is bad scores. I had a couple guys in the freenode #samsung-epic IRC chat run scores, and DRockstar with RFS got 1133 and 2521. It looks like this benchmark actually scores I/O on RFS accurately!
However, I'm still skeptical, and I'd like to see some more scores first. So download it, run some benchmarks, and post back here with your results.
Be sure to post what ROM you're running and if you're on EXT4 or RFS. Don't forget to kill your background apps before testing, I forgot to and got a pretty bad score the first time.
I'd recommend running the benchmark at least twice to be sure your phone is running at its best.
Me: Quantum ROM 2.7, EXT4
Productivity Index: 1178
Games Index: 2610
EDIT - I'd also like to give a shout out to a couple other benchmark apps that I think are pretty good. GLBenchmark (not on market, has to be downloaded from their website and installed) and 0xbench, which the guys in IRC introduced me to, a very comprehensive benchmark tool.
EDIT2 - I also would like to make it clear that I don't condone the use of any benchmark score as "proof" of any piece of hardware or ROM as being better than any other. Every phone is different and will perform slightly differently depending upon dozens of different reasons. In addition, ALL benchmark tools are susceptible to error and manipulation, particularly when you consider that our OS is running on top of a VM over the hardware and thus the hardware is not being natively tested by the benchmark tool (AFAIK). Lastly, what a benchmark cannot test is your own experience with the phone or ROM, how well it "feels" to perform for you, how much you enjoy using it, etc. Employ common sense when citing benchmarks!
UPDATE - 1/13 - An update is available for the application with more devices in the scores comparison list.
Bonsai 1.1.3 , EXT4 , No-Journal mod
Prod. Index: 947
Games Index: 2207
Stock DI18 Rom unrooted. Why is my game index so high?
tphillips78 said:
Stock DI18 Rom unrooted. Why is my game index so high?
Click to expand...
Click to collapse
Because our GPU is currently the best available in current smartphones.
Thanks for posting a stock score BTW, I was really hoping to see one.
EXT4 Quantum Rom
1178 Productivity
2715 Games
Ran it again:
1182 Productivity
2719 Games
And again:
1190 Productivity
2736 Games
Electrofreak said:
Because our GPU is currently the best available in current smartphones.
Thanks for posting a stock score BTW, I was really hoping to see one.
Click to expand...
Click to collapse
That's what I thought I just needed someone else to say it so I could believe it more lol. Thanks
1001 and 2304
Running supernova 1.04 dk28
Rfs filesystem
Sent from my SPH-D700 using XDA App
My Epic running Bonsai 1.1.3 ext4
1139 productivity
2689 gaming
My wife's Epic stock di18
700 productivity
2297 gaming
Sent from my SPH-D700 using Tapatalk
tphillips78 said:
That's what I thought I just needed someone else to say it so I could believe it more lol. Thanks
Click to expand...
Click to collapse
No probs, if you're interested in hardware, I wrote an article last April that goes into some detail about the Galaxy S hardware and compares to other modern smartphone hardware. The article is getting a little old but most of it is still relevant: http://alienbabeltech.com/main/?p=17125
1053 and 2484 using ext4 running Midnight 2.7
Electrofreak said:
No probs, if you're interested in hardware, I wrote an article last April that goes into some detail about the Galaxy S hardware and compares to other modern smartphone hardware. The article is getting a little old but most of it is still relevant: http://alienbabeltech.com/main/?p=17125
Click to expand...
Click to collapse
I am definitely going to check that out thank you.
ROM: quantium 2.7 ext4
productivity 742
games index 2505
I ran the test again after i cleared all running apps "no reboot"
productivity 1130
games index 2631
That just shows how much the phone boggs down after a day of normal usage.
737 productivity
2698 games
Devoid Angel ROM
Ran again
755 p
2732 g
Runnin CM6 with ext4 and got...
Productivity - 1151
Games - 2643
Sent from my Samsung-SPH-D700 using XDA App
Midnight Rom 2.4 ext4
Productivity :1081
Game :2383
My scores
ROM: quantium 2.7 ext4
productivity 1193
games index 2692
This is a full day booted phone just held down home and cleared all apps then ran Smartbench
Another stock DI18 not rooted.
First run....
Productivity Index: 714
Games Index: 2468
Second run....
Productivity Index: 719
Games Index: 2403
Stock DI18 not rooted 736 and 2585.
Second run, without killing any background apps:
Productivity - 1195
Games - 2694
Running Nebula ROM 1.0.6 (Twin Jets) with EXT4.
Interestingly, successive tests gave lower and lower scores.
1132 and 2750. On bonsai with no journal mod
Sent from my Evo Killer!
I've tried to overclock, lagfix, change roms and kernels, but I didn't see any big difference.
I would like to stay with real FPS benchmarks like GLBenchmark or others not like quadrant.
I've seen many people using ext2 instead of ext4, other say rfs is better, so I'm getting very confused. And also many people around here.
Each one of us has got a personalized SGS. Let's run each one of us GLBenchmark OpenGl 2.0 test without FSAA (The one of egypt) and see the results of each one of us.
Dlownload glbenchmark here: http://www.glbenchmark.com/download.jsp
Please specify model, rom, kernel, launcher, running apps, free mem before doing the test... Like every bit of info you can give, so we could find the best settings!
Thanks to all ! I'll run the test and post in few minutes!
Ok, here is the template 1.0 to use ;P just copy and paste and modify!
OK, got:
2438 Frames (21.6 Fps) with no fsaa.
My settings:
MODEL: European, I9000 SGS 16gb Vodafone
SPEED (GHZ): 1.0
LAUNCHER: Go Launcher
KERNEL: Lastest speed mod.
ROM: Darky's 8.0
LAGFIXES: On, System converted, on fast mount.
PROGRAMS INSTALLED: Very few, just wiped it.
FREE RAM: I forgot!
OTHER INFO: Wifi on, and overclock @ 1.2 makes worse FPS.
Lagfix will have no effect on glbench
Sent from my GT-I9000 using XDA App
Bench marks do not say how responsive or laggy the phone is and unless you are obsessive with numbers are a waste of time .
jje
yes, but games performance is equal to FPS, so more FPS=More optimized system.
Just because that reason i proposed to use glbenchmark, as it is an only graphical test.
This thread is to showcase before and after speed increase/decrease that you may get from using either one of the Kernels that have this implemented..
I request you do the following..
1: State your Rom
2: State your Kernel
3: State your before speed
4: program used to verify speed
This thread is not a Q&A thread..
It is also not a thread about which is the best Kernel to use..
It is a thread that the developers can refer to to get an idea whether it is worth while to implement "Veno" and or any other Wireless coding to help improve input & output of our devices and the wireless chips/programming that supports them..
I will quote ricardopvz.. if he does not mind.. (Ricardo.. PM me if you have a problem with this)
Wifi is definitely better with test16. Where I had 36Mbps, now I have 54Mbps.
Page loading may be a little faster, but needs some more testing.
Click to expand...
Click to collapse
http://forum.xda-developers.com/showpost.php?p=25136849&postcount=721
dgcruzing said:
I request you do the following..
1: State your Rom
2: State your Kernel
3: State your before speed
4: program used to verify speed
Click to expand...
Click to collapse
Using the same application for testing internet speed and benchmark for result consistency... like quadrant & speedtest.net..
Here is the result after veno:
Megatron v 1.0.4 @ 1600mhz
Guevor test16 - veno
Speedtest.net
download - 1986kbps
upload - 372kbps
Quadrant score - 4069
will post shortly after flashing other kernels...
AOKP Build 33
1.504mhz - 216mhz - Interactive
Test16 kernel
SIO (M.S) read ahead 2048kb
SD read ahead 1024kb
Dalvik byte code false build.prop tweaks etc
wifi bluetooth etc untouched
Before veno and after veno no difference at all in fact in most speed tests veno got slightly less but then seemed a little quicker in other respects eg browsing at times so the tests did not help me make a definative decision.
I really don't know what to think on this but I would like to see more improvements like this and see our kernels and all round speed improve
Sorry in advance for any spelling mistakes AOKP still has a borked autocorrect and my spelling lacks in every respect lol
Sent from my Transformer TF101 using XDA Premium HD app
1 - ARHD 3.2.0
2 - TastyMehIcs .8 (before) 1.0 (After)
3- Before 8Mbps After 12Mbps
4- Speedtest.net app.
The real improvement seemed to be in responsiveness, link clicks aren't lost asmuch
Megatron v 1.0.4 @ 1200mhz
Guevor test15 & test16
Speedtest.net
Measurements:
Time Download Upload Ping Server
test15
1:38PM 3347kbps 1615kbps 47ms Veendam
1:39PM 3328kbps 1560kbps 51ms Veendam
1:41PM 3183kbps 1595kbps 20ms Mechelen
test16
1:44PM 3146kbps 1566kbps 17ms Mechelen
1:45PM 3092kbps 1639kbps 19ms Mechelen
1:56PM 3285kbps 1629kbps 17ms Mechelen
back to test15
2:08PM 6367kbps 1628kbps 22ms Mechelen
2:09PM 4434kbps 1592kbps 19ms Mechelen
2:14PM 3030kbps 1330kbps 19ms Mechelen
As you can see time and server can have an influence on the results, so you can't just do one measurement before and one after to make a conclusion.
Yes, I am having this thought too..
Was going though test programs yesterday..
At this stage "ping" can only be the true medium you can test this.
Or using a samba drive attached to your WiFi network and download upload the same file and different kernel settings using a monitor program to get average speeds..
As using results from multi pads connected to the internet brings too many variables into the data..
i.e..users downstream internet speed once it leaves the local network..
What we really want is "local" WiFi to pad increases.. as its the tweaking of this that will allow for incremental increases (hopefully) of the user experience in the downstream leg of the connection.
sent from yet another MikG HTC Evo
I did a factory reset (latest AOSP w/ Motley kernel) and installed nothing but Androbench (didn't even add my Google account) and got these Androbench results:
Squential Read: 24.3 MB/s
Sequential Write: 1.41 MB/s
Random Read: 7.81 MB/s
Random Write 0.23 MB/s
SQLite Insert: 18.92 TPS / 15.85 sec
SQLite Update: 23.45 TPS / 12.79 sec
SQLite Delete: 22.67 TPS / 13.23 sec
Browser: 2830 msec
Market: 891.25 msec
Camera: 1404 msec
Camcorder: 1323.75 msec
Pretty abysmal overall, random write especially falls off the charts altogether. This is really sad since my N7 used to run great but recently has degraded significantly with lots of lag and ANRs.
Is it safe to say I need to RMA this device due to bad storage?
Have you tried using a different ROM? Try one like Paradigm or Paranoid Android and see if there's a difference. When I ran quadrant on some ROMs I got around 2,000 consistently for I/O but when I ran it on some others (eg. Paradigm Android, BAKED Blackbean) I scored 7,000-8,000 consistently (I also used _motley's kernel).
android1234567 said:
Have you tried using a different ROM? Try one like Paradigm or Paranoid Android and see if there's a difference. When I ran quadrant on some ROMs I got around 2,000 consistently for I/O but when I ran it on some others (eg. Paradigm Android, BAKED Blackbean) I scored 7,000-8,000 consistently (I also used _motley's kernel).
Click to expand...
Click to collapse
Thanks for the suggestion. I get ~1,200 for I/O in Quadrant normally.
I installed BAMF Paradigm and my scores got much, much worse (about 1/3rd as fast across the board). I tried installing Trinity kernel on top of Paradigm and it's much better (roughly 2x standard). Unfortunately my random write is the same, which is absurdly, brokenly low.
However my Quadrant score is now ~7000... I just don't know what to believe any more :crying:
What I/O scheduler are you using?
how much space do you have left in your storage?
Using the default scheduler in any given case. And I wiped my SD Card so there's plenty of room left.
I just un-rooted and installed the factory image - the I/O performormance has largely jumped back to normal now, so I'm at a complete loss as to what's going on I'm going to take it and run with stock for a while I think.
Thanks so much for the help all!