Axim Slow performer

I have noticed some peculier behaviour of my new Axim X5 400 mhz, just wanna inform everyone of my findings. The thing I'm having problem with is that this PDA has Incredably SLOW graphic performance for its processor speed. B4 you start flaming me, plz hear me out.

for the BMQ DRAW benchmark, my brand new Axim X4 400 mhz (set to run at 400 of course) only scores 156, this is really extremely slow. For comparison. the Ipaq 3870 with 206 strongarm processor scores 188 points while the Toshiba E550g which also uses XSCALE scores 240 at 200 mhz and trashes the Axim by over 120% with a score of 384 in 400 mhz mode. with the SAME processor, I dun't see why this is the case

(figures of non-axim from: http://www.happycorner.com.hk/sp008.htm )

This brings me to the next point that when I alter the speed of my Axim via the control tab in settings from 400mhz to 200 mhz, the result of the DRAW benchmark does NOT change! why is that?? for every other PDA benchmarked with BMQ I find on the net, faster Processor speed == faster draw performance.

of course, take note that this only applies to GRAPHIC benchmarks. I have noticed the same atrocious performance in the 3D bench section of the GAPI benchmark, my friend's Mitac Mio 338 (with Xscale 200 mhz so it has nothing to do with architecture) scores 56 on the 3D mark while the axim X5 at 400 mhz lags behind at a paltry 36. of course, what I've mentioned above also applies in this benchmark, with no discernable difference in the score when the Axim is lowered to 200 mhz.

Of course, this DOES impact real world performance, as far as I know, the Axim seem to be the only 400 mhz xscale machine incapable of playing back high quality Divx without dropping tonnes of frames. For example, when I played back the matrix reloaded trailer ( latest 0.8 pocket Divx player with 12mb of full cache enabled ) I get 32 dropped frames, this is simply unacceptable compared the the 300 mhz Xscale E330, which drops only 8 frames playing back the same video file. For another PDA with the same kind of processor as the Axim, the 3970 manages to drop ZERO frames.

(figures for non axim from: http://www.pdabuyersguide.com/viewsonic_V35.htm)

I've verified that all benchmark which I ran are the SAME VERSION as the ones which I compare to. ( yes, I CAN read chinese ), I also ran every benchmark FRESH after a soft reset each time. my PDA is empty so memory is NOT the problem and I ran the divx test off my ram so it should actually score better than when I run it off an SD card like what they did on the above site.

I'm puzzled by all these and I hope that Dell can provide a patch or some sorta of program to fix this issue of slow graphic performance, seeing dell's adroit handling of the D-pad Issue, I'm sure you'll provide a fast remedy. With faster graphics, the Axim will truly become a fantastic PDA. Thanx for reading

[Emonair]

I have had no problems with my Axim performance. I haven't done a benchmark test. It seems to me that you are comparing apples to oranges. You can't take Benchmark test A with pdas a-c and compare it to Benchmark test B with pdas d-f. They are probably different types of benchmark tests that test useing different algorythmes. Your findings would be more credible if we did the same benchmark test for several different pdas, axim included. Even then that would just be applicable to your benchmark test, not how it runs Age of Empires or Streaming video or whatever. I'm a little bias too, so it would take a piano falling on my head type of test result to get anything negative through my thick skull about my baby...

[Mike C]

I just played the matrix thing and it dropped 0, w/ 24fps.

I copied the matrix file to SD card in power save mode, and it took 38.5 sec.
I copied the matrix file to SD card in normal mode and it took 33.9 sec.

RESULT: The speed change does have an effect.

SUGGESTION: Set your "Drop frame tolerance" to high.

[LeapinLew]

You would think the performance benchmark would improve or get worse depending on the Mhz it was running at.

interesting.

[shaneincottondale]

Word of advice on benchmarks: play with them, read the numbers, then move on. Benchmarks are so biased one way or another that you will never get anything reliable. I've run 3DMark2001 on all mine and my friends systems....my score is 8453 but I can't tell any difference whatsoever when using a pc with a 10000 score. "Wow, a benchmark says my Axim has an arbitrary number that I dont have a real world reference to, woe is me." The only thing I've found benchmarks are good for is bragging rights between friends. We've all sat around and not one person can tell a difference playing the same game or using the same app on a pc ranging from a 5000 to a 10000 score. Do you really think your eye can tell the difference if you get 3 or 4 fewer frames per second? Doubt it. If you get all involved in numbers, you will never be satisfied with anything because no matter what you buy there will always be something better. I'm not flaming you, just opening your eyes a bit about turning useless data into a religion. :P

my 2 cents
shane

[LeapinLew]

I disagree. Benchmarks are what we use to determine performance. The Onion 3dMark test you refer to isn't an overall benchmarking tool - it mainly uses your video card to determine it's ranking. Put a Geforce 4 PCI card in a Pentium 75 and get a score in the 5-6,000 range. Try to play a recent 3d based game and you'll see the difference.

When I bought my Axim I looked at Benchmarks - well, all the benchmarks I could find. I just wanted to make sure it was in league with other devices.

I run speedtests against my T1 at work to make sure we are getting what we are paying for. I would say thats a benchmark. I have other T1 speeds to compare it to - so I think benchmarking a processor to make sure your getting everything your supposed to isn't out of line and asking questions when the benchmarks fall way short isn't a bad idea - in fact it's quite consumer concious.

That being said - I still would like to know why a drop in the Mhz doesn't affect the score of the benchmark. Makes me wonder if we are running at 200mhz the whole time or if the X-scale has some issues to work out still.

Curious!

[shaneincottondale]

ill bet you use 97 octane gas...

:P


shane

Emonair:
>>>>>>>>>>>>.Y our findings would be more credible if we did the same benchmark test for several different pdas<<<<<<<<<<,,

thats what I did, I compared the results of the SAME benchmark on a few PDAs, in this case, the DRAW benchmark of BMQ and 3D bench of GAPI


Mike C:
>>>>>>>I just played the matrix thing and it dropped 0, w/ 24fps.<<<<<<<<

umm... could you please tell me what player do you use? and what settings? thanx

>>>>>>>RESULT: The speed change does have an effect<<<<<<<<<<<

yes, speed change DOES have an effect on MEMORY performance, I'm saying that it does not have an effect on just the GRAPHIC peformance


shaneincottondale:
>>>>>>play with them, read the numbers, then move on<<<<<<<<<<

sure, but these are REAL LIFE benchmarks.. you gotta realize that all games for PPC are programmed with GAPI.. so naturally the GAPI bench will tell you how your games will perform

[Mike C]

could you please tell me what player do you use? and what settings? thanx

Pocket MVP, 12 meg buffer, DFT all the way to high. The file is: matrix2-teaser.avi 2711kb
Vid playing in window'd view (about 120 x 240) Tried full screen, but it was only showing me half of the scene, as if it was set to 200% zoom. Don't know how to change that.

Info from the pop-up that appears after the clip ends.

I used VOBenchmark (on my 300MHz Axim) and noticed that the Axim was superior in the CPU and Filled tests. However, it performed at the same level as the UR-There running a 131MHz MIPS under the Stretch Bit test. Weird. The Memory Move test was my biggest concern. As soon as I saw the score (.38) I was nonplussed. I had noticed that it moved files a little slower than I thought it should and the test confirmed it. td3d, since you can read Chinese, could you tell me which link you used to get the BMQ Draw Benchmark utility? Thank you.

[SkyVoice]

Snooze. Yawn. My Pocket PC is faster than your Pocket PC. Whoopee!!! Most folks have far more proceesing power than they need anyway.

The water is real deep too!

Ciao,

Ron