Extremely low Smartbench scores - myTouch 4G Q&A, Help & Troubleshooting

Why are my Smartbench scores so low? 941 productivity. It's lower than stock according to the app. Funny thing is, I'm scoring around 2000 in Quadrant. Running Ice with faux's kernal 0.7.7 oc'd to 1.4. Any thoughts?
Sent from my HTC Glacier using XDA App

TheNarrowGate said:
Why are my Smartbench scores so low? 941 productivity. It's lower than stock according to the app. Funny thing is, I'm scoring around 2000 in Quadrant. Running Ice with faux's kernal 0.7.7 oc'd to 1.4. Any thoughts?
Sent from my HTC Glacier using XDA App
Click to expand...
Click to collapse
Quadrant scores are inaccurate.
and 1pt in SmartBench might not equal 1pt Quadrant.

I figured the points aren't 1 for 1, but the score is very low. Quick check of the benchmark thread showed a few others with the same problem, but some are getting great scores. Any suggestions for improving mine?
Sent from my HTC Glacier using XDA App

TheNarrowGate said:
I figured the points aren't 1 for 1, but the score is very low. Quick check of the benchmark thread showed a few others with the same problem, but some are getting great scores. Any suggestions for improving mine?
Sent from my HTC Glacier using XDA App
Click to expand...
Click to collapse
I really do not go by quadrant scores.
They are inaccurate and no real benefit.
The only reason i use Quadrant is to test a ROM flash to make sure i didn't forget to wipe the Handset.
Stuff like open Applications, Widgets, Launchers, Themes background data all have factors melted into Quadrant Scores.

Related

Low benchmark score

I have ice glacier 1.1.6 and then the version 0.3.0b BFS-SLQB flash as well, my quadrant score is only around 1700, is this normal or is it something else going on?
Please post all question in the Q&A section. I moved it for you.
danielvictor said:
I have ice glacier 1.1.6 and then the version 0.3.0b BFS-SLQB flash as well, my quadrant score is only around 1700, is this normal or is it something else going on?
Click to expand...
Click to collapse
Is your phone OverClocked?
1700 isn't all that bad for a for a 1Ghrz score.
it also depends on what launcher, what widgets and other applications you have up and running at the time of the quadrant.
Yeah I found out about overclock, I still get low score, is there any guide to run it at the top speed?
Sent from my HTC Glacier using XDA App
danielvictor said:
Yeah I found out about overclock, I still get low score, is there any guide to run it at the top speed?
Sent from my HTC Glacier using XDA App
Click to expand...
Click to collapse
No really, no preset guides on how how to get "top end" Benchmarks. Just make sure you do not have alot of widgets, applications and such running. and if you didn't wipe your handset via recovery image you get low scores.
danielvictor said:
I have ice glacier 1.1.6 and then the version 0.3.0b BFS-SLQB flash as well, my quadrant score is only around 1700, is this normal or is it something else going on?
Click to expand...
Click to collapse
I've been getting ~1200 on my MT4G stock. I've been trying to figure out why, and I've come to the conclusion that my CPU changes frequencies often. I'll see it at min freq at random times. You may want to check and see what your CPU is running at when you do a benchmark.

Smartbench 2010 vs Smartbench 2011

Has anyone downloaded 2011.......... and compared the scores......epic now for everyday uses scores significantly lower on 2011 I am running synicates newest rom twilight kernal 1.2...setcpu i've tried on demand, conservative, and performance........from what I have concluded new stress test has dropped out scores.
Sent from my SPH-D700 using XDA App
Syndicate *****
Sent from my SPH-D700 using XDA App
lballer69 said:
Has anyone downloaded 2011.......... and compared the scores......epic now for everyday uses scores significantly lower on 2011 I am running synicates newest rom twilight kernal 1.2...setcpu i've tried on demand, conservative, and performance........from what I have concluded new stress test has dropped out scores.
Sent from my SPH-D700 using XDA App
Click to expand...
Click to collapse
The lower score is because Smartbench 2011 replaces the File I/O tests with additional CPU tests. A large percentage of the 2010 score was File I/O.
Quote: "Some of the test suits have been revamped from Smartbench 2010. File IO tests have proven to be somewhat inconsistent on some manufacturers' phones. Instead, another test has been introduced that will further stress the CPU."
mattallica76 said:
The lower score is because Smartbench 2011 replaces the File I/O tests with additional CPU tests. A large percentage of the 2010 score was File I/O.
Quote: "Some of the test suits have been revamped from Smartbench 2010. File IO tests have proven to be somewhat inconsistent on some manufacturers' phones. Instead, another test has been introduced that will further stress the CPU."
Click to expand...
Click to collapse
This AND the fact that the new baseline score is the G2, not the Nexus One as in 2010.
intx said:
This AND the fact that the new baseline score is the G2, not the Nexus One as in 2010.
Click to expand...
Click to collapse
Yes, thanks, I forgot to mention that.
Ty for the info was wondering y was there a huge difference...
Sent from my SPH-D700 using XDA App
How did.the.Motorola xoom score so high on that test? Wow
Sent From My Evo Killer!
musclehead84 said:
How did.the.Motorola xoom score so high on that test? Wow
Sent From My Evo Killer!
Click to expand...
Click to collapse
Dual Core at 1.5 on Android 3.0 will do that
smartbench also does some test desinged for dual+ cores so thats why our scores drop
Everything here is absolutely correct. One of the reason why I used the year number as part of the app name is so that I can change the baseline and test apps. Once these are changed, previous scores are no longer comparable, hence it really deserves a new name. As you may have guessed, Smartbench 2012 will again feature more refined set of tests and the baseline will probably be based on higher specced phone once again. By doing this, I can keep the high-end scores in check as well.
Would like to tell u the app is greatly appreciated.............keep up the gud work
Sent from my SPH-D700 using XDA App
lballer69 said:
Would like to tell u the app is greatly appreciated.............keep up the gud work
Sent from my SPH-D700 using XDA App
Click to expand...
Click to collapse
Thanks!

Linpack score of 16mflops

What's the average score for a evo on cm7 .
Sent from my PC36100 using XDA App
parmend said:
What's the average score for a evo on cm7 .
Sent from my PC36100 using XDA App
Click to expand...
Click to collapse
Just tested at 41.8
Yeah, I would say avg is anywhere from 35 to 42, depending on clock speed. What happens if you run it again right after a couple times?
One time I got and 8, and I was like, WTF? Ran it again and got a 38.
Thanks, I just ran it and it still 16-18. Does the kernal matter. I'm on timat 3.3.7
Sent from my PC36100 using XDA App
yeah on cm7 I got 40 give or take.
stock sense is like 32 for me and overclocked to 1.2ghz i get 38-39.
bfs kernels were hit or mis on my EVO. Kinda clunky on mine.
I'f you're getting low linpack scores try switching from smartass to conservative(just to score higher). Also I get an average of 41 with my phone clocked at 1190.
Sent from my PC36100 using XDA Premium App
Specs in sig, I just got 43.261, average is 42-44.
How do you check?
Arnezie said:
How do you check?
Click to expand...
Click to collapse
Download linpack from the market
mattykinsx said:
Download linpack from the market
Click to expand...
Click to collapse
Thanks
32.059
Stock rooted rom
25.59 Destroyed v2 how do i improve that?
39.6
Mik 4.61
Netarchty-toastmod 4.3.1
Setcpu CPU 1190MHZ Very Stable
Quadrant vs linpack? Which one is more important?
Sent from my PC36100 using XDA App
parmend said:
Quadrant vs linpack? Which one is more important?
Sent from my PC36100 using XDA App
Click to expand...
Click to collapse
I would say definitely linpack. Quadrant is not consistent.
you can get bad scores if you have to many services running. i get 33-35 stock non rooted hw 004. got up to 44+ on my old hw 002(got the restart when overheated problem) when 2.2 first came out, completely stock but since then they have added alot of services that slow it down
if the media scanner is running. i get about 22. so check that in the services if you are getting bad scores
Fricksta said:
25.59 Destroyed v2 how do i improve that?
Click to expand...
Click to collapse
Need more details. What kernel, what CPU settings, etc. I was getting 38-40 on d2.
Just got 40.75 on cm7 with dark horse theme and sz 2.0.0 ondemand 1190
The Linpack app also does not pull enough from CPU to ramp the CPU to max. Set your minimum CPU to the same as your max, run linpack then lower your minimum CPU speed, Interactive is best governor to use for this
all these tests are for is braggin rights...these scores will not be like, oh check it out, my browser loads .3 seconds quicker bc i got 3 more mflops...get out of hee, find a rom you like, and use it, most of them are very good. There is not one rom on this site that blows away the competetion, they are all very very very similiar in speed, nuff said

Usa v6 speed mod

dsexton702 said:
These are some tweaks created by dsexton702 that offer you a generous performance boost in I/O and giving you the newest fugu mod 3.1. This is a fast growing tweak and needs to be shared with everyone who can benefit from it
download the U.S.A.Sv6
OP: http://forum.xda-developers.com/showthread.php?t=1236779
Click to expand...
Click to collapse
It does work on our phones as I'm using right know...just thought I pass on the good work of these devs...
Sent from my PG06100 using xda premium
What is it like V6SuperCharger?
Sent from my PG06100 using XDA App
It speeds up and opyimuzes androids Input and output....its just version6 of the mod.....thy have a new version but I haven't tried yet..I've used this on my HTC hd2 and mytouch 4g and know my shift .....
Sent from my PG06100 using xda premium
Idk if I posted in right section if not can mod move this please to right one....
Sent from my PG06100 using xda premium
v6 supercharger is useless IMO. Tweak your data setting in ##DATA# with a sense rom and use a good kernel like super scary stable.
VICODAN said:
v6 supercharger is useless IMO. Tweak your data setting in ##DATA# with a sense rom and use a good kernel like super scary stable.
Click to expand...
Click to collapse
You see I did tweak my ##DATA# setting and use Scarykernel on the latest CM7 nightlies. There's a noticable improvement using V6Supercharger over without on my device.
Sent from my PG06100 using XDA App
Same here I have found a noticable speed also in quadrent score also....
Sent from my PG06100 using xda premium
bigfdaddy said:
Same here I have found a noticable speed also in quadrent score also....
Sent from my PG06100 using xda premium
Click to expand...
Click to collapse
What's your antutu score with Scary kernel on toasty at 1900mhz?
Here the score Wich includes v6 mod,scary govener,1900mhz over clock,and SD booster...
Sent from my PG06100 using xda premium
bigfdaddy said:
Here the score Wich includes v6 mod,scary govener,1900mhz over clock,and SD booster...
Sent from my PG06100 using xda premium
Click to expand...
Click to collapse
Proof that this **** is useless:
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
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 gov for benchmarking and run benchmarks on a fresh boot and close as many apps as possible.
BTW:
http://forum.xda-developers.com/showthread.php?t=1296614
---------- Post added at 01:51 AM ---------- Previous post was at 12:54 AM ----------
Benchmark scores do not prove actual real life performance...
They are just numbers ppl like to boast about and doesn't mean your phone is any faster/more responsive with a higher score.
And. Any scores wether antutu/quadrant/or whatever can be faked.
Sent from my Supreme Shift on Wires!
notsointeresting said:
Benchmark scores do not prove actual real life performance...
They are just numbers ppl like to boast about and doesn't mean your phone is any faster/more responsive with a higher score.
And. Any scores wether antutu/quadrant/or whatever can be faked.
Sent from my Supreme Shift on Wires!
Click to expand...
Click to collapse
I wouldn't be wasting my breath and posting fake benchmark scores if I didn't believe it and try this mod.
The only difference I noticed was LPP loading a little bit faster on startup.
VICODAN said:
I wouldn't be wasting my breath and posting fake benchmark scores if I didn't believe it and try this mod.
The only difference I noticed was LPP loading a little bit faster on startup.
Click to expand...
Click to collapse
Well every little bit counts doesn't it?.....ty for trying it tho and posting ya score....
Sent from my PG06100 using xda premium
bigfdaddy said:
Well every little bit counts doesn't it?.....ty for trying it tho and posting ya score....
Sent from my PG06100 using xda premium
Click to expand...
Click to collapse
Np.. just saying everytime I flash a nightly I have to do this.. it's annoying as hell where as I just flash ScaryStable once with each nightly flash and I'm good to go and I notice a HUGE difference in performance.
For what it's worth, the V6 Supercharger (currently running the update9 beta) provides an optimal RAM manager which eliminated my habit of installing AutoKiller w/ every ROM. Since running the Supercharger script, I've consistently ran w/ 175-180MB in free RAM w/ no app FC's. That alone is worth using this mod in my book.
Ty 4 trying.....tying to expand this to more people and let them know about it.I.think its worth all the hype and attention I can give it.
Sent from my PG06100 using xda premium

Cm7 cpu quadrant

How come the cpu seems so slow in quadrant standard tests.... Heck, my desire z (800mhz single core) blew through the cpu component of the test faster...This result is typical of all Cm7 roms I've tried. Below my result is... Surprise! Atrix 4g, with superior cpu rating, but inferior i/o. I find this all kind of counter intuitive...
Any insights?
Sent from my MB860 using xda app-developers app
=JKT= said:
How come the cpu seems so slow in quadrant standard tests.... Heck, my desire z (800mhz single core) blew through the cpu component of the test faster...This result is typical of all Cm7 roms I've tried. Below my result is... Surprise! Atrix 4g, with superior cpu rating, but inferior i/o. I find this all kind of counter intuitive...
Any insights?
Sent from my MB860 using xda app-developers app
Click to expand...
Click to collapse
um... your I/O bar is the biggest.
HMZX said:
um... your I/O bar is the biggest.
Click to expand...
Click to collapse
I think he meant inferior CPU and superior i/o.
HMZX said:
um... your I/O bar is the biggest.
Click to expand...
Click to collapse
=JKT= said:
How come the cpu seems so slow in quadrant standard tests.... Heck, my desire z (800mhz single core) blew through the cpu component of the test faster...This result is typical of all Cm7 roms I've tried. Below my result is... Surprise! Atrix 4g, with superior cpu rating, but inferior i/o. I find this all kind of counter intuitive...
Any insights?
Sent from my MB860 using xda app-developers app
Click to expand...
Click to collapse
have you underclocked?
i got double your cpu score on 1.45ghz
Its because cm7 doesn't initialize both cousin in quadrant. Try it on cm9 or cm10 CPU doubles.
Managed to get 4.5k quadrant score at 1.45 on the older builds. 4K with 1.3ghz
Sent from Moto Atrix 4g on Neutrino 2.91
I guess I wasn't clear enough lol
My phone was at stock 1ghz. This was to compare objectively to the bar below the "your device" bar, which was entitled "atrix 4g" which had, yes, superior cpu and inferior i/o scores.
Of course you will get better scores overclocked to 1.45 ghz. Try to understand what I'm asking about. I'll ask more clearly.
Why does a custom rom appear to get a worse cpu score than stock at same clock rate? Shouldn't it be equal or greater?
squee666 said:
Its because cm7 doesn't initialize both cousin in quadrant. Try it on cm9 or cm10 CPU doubles.
Click to expand...
Click to collapse
Is that true? System panel gives independent readouts of the load on each core and they seem to both be working. But if they don't both work in quadrant that would explain it all =) (except why quadrant lol)
Sent from my MB860 using xda app-developers app
Score I got at 1.3ghz faux kernel on am older version of cm9. Yea quadrant reads that there are 2 cores. Just does not use them properly.
Sent from Moto Atrix 4g on Neutrino 2.91
Thanks guys. That does seem to be the issue. Here is a score @ 1gHz with mokee os. Looks pretty obvious that both cores are working here:
Sent from my MB860 using xda app-developers app
Lol its not obvious at all the green bar is I/o mokee is cm7 based you should try a cm9 or cm10 based rom to see CPU difference.
Sent from Moto Atrix 4g on Neutrino 2.91
squee666 said:
Lol its not obvious at all the green bar is I/o mokee is cm7 based you should try a cm9 or cm10 based rom to see CPU difference.
Sent from Moto Atrix 4g on Neutrino 2.91
Click to expand...
Click to collapse
Look at my blue (ie cpu) bar in this image compared to the image in my first post. I believe it is, in fact, quite obvious. Lol... At least that something is significantly improved cpu-wise...
Sent from my MB860 using xda app-developers app

Categories

Resources