Log in

View Full Version : Is Windows Mobile 2003 Really Faster?


Jason Dunn
08-20-2003, 05:05 PM
I had the rare opportunity to run <a href="http://www.spbsoftwarehouse.com">Spb Benchmark</a> on a single iPAQ 3650 running three generations of Microsoft's Mobile operating systems (the Pocket PC 2003 image for the iPAQ 3650 came from Microsoft - it a special testing-only build that won't be released to the public). This gave me the metrics I needed to make this definitely statement: Windows Mobile 2003, the first Pocket PC operating system built atop Windows CE .NET 4.2, is the fastest all-around operating system we've seen so far on these little devices we enjoy using so much. I present to you a series of benchmarks that show you exactly how much faster it really is! The Pocket PC 2002 and 2003 were performed on the same device, while the Pocket PC 2000 tests were performed on a different device, but of the same model.<br /><br />First, let's look at how the Windows Mobile 2003 iPAQ compares with the baseline index, which is based on a Pocket PC 2000 iPAQ 3650. Take a look:<br /><br /><img src="http://www.pocketpcthoughts.com/images/web/2003/ipaq2003-benchmark-01.gif" /><br /><br />• Overall, Windows Mobile 2003 is 15% faster<br />• The CPU performance of the StrongARM 206 Mhz is 25% faster under Windows Mobile 2003 than under Pocket PC 2000<br />• The file system is slower (0.3%), but only slightly so in the overall score (the new OS is much slower in some key tests you'll see later on in this article<br />• The graphics subsystem of Windows Mobile 2003 is vastly improved over previous versions - a whopping 40% faster.<br />• ActiveSync has the most dramatic improvements - an incredible 103% boost in speed over Pocket PC 2000.<br /><br />Read on for the comparisons between Pocket PC 2000, 2002 and 2003 - I think you'll be surprised at how much of a speed drop we endured with Pocket PC 2002 in some tests, and what a dramatic difference there is in moving to the 2003 OS. Clearly, the resources required to move to the new Windows CE .Net 4.2 core were worth it, even if it meant that we didn't get that many new application-specific features in Windows Mobile 2003.<br /><!><br />The screen shots below do not represent every possible test that Spb Benchmark is capable of performing, only the ones I felt were more relevant and showed the most dramatic performance differences. The indexes and averages are based on the entire suit of Spb Benchmark tests.<br /><br /><img src="http://www.pocketpcthoughts.com/images/web/2003/ipaq2003-benchmark-02.gif" /><br /><i>Figure 2: We lost speed with Pocket PC 2002, but gained it back (and then some) with Pocket PC 2003</i><br /><br /><img src="http://www.pocketpcthoughts.com/images/web/2003/ipaq2003-benchmark-03.gif" /><br /><i>Figure 3: Windows Mobile 2003 offers significant CPU enhancements for any ARM processor, without any special XScale tweaks. This is further proof to me that Microsoft made the right decision in not making special optimizations for the XScale processor, and instead focusing on improving the OS for all ARM-compatible processors.</i><br /><br /><img src="http://www.pocketpcthoughts.com/images/web/2003/ipaq2003-benchmark-04.gif" /><br /><i>Figure 4: Windows Mobile 2003 gives us greater graphics performance.</i><br /><br /><img src="http://www.pocketpcthoughts.com/images/web/2003/ipaq2003-benchmark-05.gif" /><br /><i>Figure 5: As you can see, ActiveSync performance actually got worse with Pocket PC 2002, but we see huge performance gains with Windows Mobile 2003. Testing on the Pocket PC 2002 and 2003 devices was performed using ActiveSync 3.7 on a USB 2.0 port.</i><br /><br /><img src="http://www.pocketpcthoughts.com/images/web/2003/ipaq2003-benchmark-06.gif" /><br /><i>Figure 6: The platform index shows us that while Windows Mobile 2003 doesn't offer enormous enhancements over Pocket PC 2000, it does over Pocket PC 2002, which is where most of us are coming from.</i><br /><br /><img src="http://www.pocketpcthoughts.com/images/web/2003/ipaq2003-benchmark-07.gif" /><br /><i>Figure 7: Unfortunately, the file system is where the performance losses in Windows Mobile 2003 are most noticeable - this test measures how long it takes the Pocket PC to enumerate (list) a directory with 2000 files. It's not as grim as it looks though - this test is ranked very low in the Spb Benchmark scores, because while the chart would visually indicate a huge difference, it's really the difference between Windows Mobile 2003 taking 0.06 seconds to enumerate 2000 files, and Pocket PC 2002 taking 0.015 seconds. Both are fast. ;-)</i><br /><br /><img src="http://www.pocketpcthoughts.com/images/web/2003/ipaq2003-benchmark-08.gif" /><br /><i>Figure 8: Internal database speeds were one of the few things that got faster with Pocket PC 2002, and that trend continued with 2003.</i><br /><br /><img src="http://www.pocketpcthoughts.com/images/web/2003/ipaq2003-benchmark-09.gif" /><br /><i>Figure 9: The BitBlt graphics test measures how quickly the Pocket PC can draw an image to the screen using GDI functions. Windows Mobile 2003 shows a tremendous speed boost over previous operating systems, which impacts normal graphics operations like displaying pictures, simple games like Solitaire, etc. Games use GAPI, which this test does not measure.</i><br /><br /><img src="http://www.pocketpcthoughts.com/images/web/2003/ipaq2003-benchmark-10.gif" /><br /><i>Figure 10: A Windows Mobile 2003 device is capable of opening a 240 KB Pocket Word document nearly 300% faster than a Pocket PC 2000 device.</i><br /><br /><img src="http://www.pocketpcthoughts.com/images/web/2003/ipaq2003-benchmark-11.gif" /><br /><i>Figure 11: Yes, that's right - things got slower with Windows Mobile 2003 when it comes to browsing the Web. Pocket Internet Explorer is still a little pokey when it comes to processing HTML files. One of the reasons that Pocket PC 2000 is much faster than Windows Mobile 2003 in this test is that it's a much simpler browser and supports less HTML functions than Windows Mobile 2003. I feel there's a lot of room for improvement with Pocket Internet Explorer - Windows Mobile 2003 didn't meet my expectations in this regard.</i><br /><br /><img src="http://www.pocketpcthoughts.com/images/web/2003/ipaq2003-benchmark-12.gif" /><br /><i>Figure 12: JPEG loading performance is likely improved under 2003.</i><br /><br /><img src="http://www.pocketpcthoughts.com/images/web/2003/ipaq2003-benchmark-13.gif" /><br /><i>Figure 13: When it comes to opening a folder with 2000 empty files using File Explorer, Windows Mobile 2003 offers dramatic gains over Pocket PC 2002.</i><br /><br /><img src="http://www.pocketpcthoughts.com/images/web/2003/ipaq2003-benchmark-14.gif" /><br /><i>Figure 14: Although you'd think that ZIP compression speed would be solely a function of the CPU, Windows Mobile 2003 gives a solid 128% speed boost over Pocket PC 2002. Cool. 8) </i><br /><br /><img src="http://www.pocketpcthoughts.com/images/web/2003/ipaq2003-benchmark-15.gif" /><br /><i>Figure 15: ActiveSync speeds saw the greatest boost - I wonder if this means they finally got rid of the USB over serial driver that hobbled previous ActiveSync speeds? The speed increase in moving data from the desktop to the Pocket PC is so dramatic (139% faster) that this is the first generation of Pocket PC where I don't think people need to have an external memory card reader to load music onto their device - it's finally fast enough to do directly from desktop to Pocket PC. Well done Microsoft!</i><br /><br /><img src="http://www.pocketpcthoughts.com/images/web/2003/ipaq2003-benchmark-16.gif" /><br /><i>Figure 16: Although not as dramatic as the speed difference above, the speeds at moving data from the Pocket PC to the desktop PC are also enhanced.</i><br /><br /><img src="http://www.pocketpcthoughts.com/images/web/2003/ipaq2003-benchmark-17.gif" /><br /><i>Figure 17: This is the overall index showing how the three iPAQ 3650's each compare to a Dell Axim X5 running Pocket PC 2002, and a new iPAQ 2215 running Windows Mobile 2003. Incredibly, even the blazing iPAQ 2215 still trails the original iPAQ 3650 slightly. What kind of nitro did Compaq put in that device? I find it somewhat sad that, over three years after the fist iPAQ came out, we can't completely trump it with brand new devices.</i><br /><br /><img src="http://www.pocketpcthoughts.com/images/web/2003/ipaq2003-benchmark-18.gif" /><br /><i>Figure 18: I continue to have my doubts about how good of a CPU the XScale processor is. The performance dominance of the XScale PXA255 version is obvious when compared to the XScale PXA250 (Dell Axim), but when you consider that the iPAQ 3650 with a score of 1251 is running at 206 Mhz, the score of 1784 for the 400 Mhz XScale PXA255 becomes much less impressive. I hope Intel's next-generation XScale processor offers more impressive performance. I wish I had the Windows Mobile 2003 upgrade for my Dell Axim X5 so I could see exactly how much difference Windows Mobile 2003 makes on a PXA250. Was the PXA250 being held back by the Pocket PC 2002 operating system, or is it the other way around?</i><br /><br /><img src="http://www.pocketpcthoughts.com/images/web/2003/ipaq2003-benchmark-19.gif" /><br /><i>Figure 19: As readers of my iPAQ 2215 review will know, the 2215 benchmarked strangely slow in some graphics tests. This chart shows how poorly the 2215 benchmarks against the iPAQ 3650, but it also shows how much the OS alone improves the graphics performance, making the difference in performance of the 2003 iPAQ 3650 (1399) and the iPAQ 2215 (567) even more aappalling- a massive 247% difference! iPAQ 2215 owners take heart - it's been speculated that the Spb Benchmark tests don't tap into the MediaQ chip, so these results shouldn't be taken as a true measure of your device performance. Hopefully future versions of Spb Benchmark will take advantage of the MediaQ chip.</i><br /><br /><img src="http://www.pocketpcthoughts.com/images/web/2003/ipaq2003-benchmark-20.gif" /><br /><i>Figure 20: Further reiteration of what a big difference the operating system makes when it comes to ActiveSync speeds. The iPAQ 2215 pulls ahead here (as it should), and the speed of the Axim running 2002 shows that the more modern hardware can account for as much as a 52% speed boost, even with the same OS (2002 Axim vs. 2002 iPAQ).</i><br /><br />A big thanks to <a href="http://www.spbsoftwarehouse.com">Spb Software House</a> for making such a powerful tool for us reviewers to use!

The Half-Ling
08-20-2003, 05:39 PM
iPAQ 3650 (2003, 205 Mhz)

A buuhhhh?


Didn't relize they upgraded those models hehe...or maybe thats just a nice feature of spb benchmark utility :-p

Jake-

Steven Cedrone
08-20-2003, 05:53 PM
iPAQ 3650 (2003, 205 Mhz)

A buuhhhh?


Didn't relize they upgraded those models hehe...or maybe thats just a nice feature of spb benchmark utility

Read it again... :wink:

Pocket PC 2003 image for the iPAQ 3650 came from Microsoft - it a special testing-only build that won't be released to the public

Steve

peterawest
08-20-2003, 05:56 PM
the Pocket PC 2003 image for the iPAQ 3650 came from Microsoft - it a special testing-only build that won't be released to the public

Too bad. I sure would enjoy upgrading my iPAQ 3630. :?

At least I have 2002 running on it. It may not be as fast, but I still like the look of the OS.

Thanks for the nice article.

easylife
08-20-2003, 06:19 PM
Nice benchmarks, Jason. I would have liked to get some benchmarks on my e310 in comparison to the rest of the list... :( Anyways, one O/T feature request for SPB Benchmark: The devices have the same colors chart to chart. This way, you can tell at a glance who came out on top, on bottom, etc. without having to read the captions every time. (Yes, I really am that lazy that I don't want to read the captions... :wink: ) Anyways, great benchmarks! :)

EDIT: Hmmm... was it just me or did this thread disappear for a little while... :wink:

disconnected
08-20-2003, 07:34 PM
PIE is my biggest disappointment in PPC 2003.

My iPAQ 3630 (PPC 2000) with Bluekite, a Supplynet cable, and a Sprint phone (PRE-1xrtt) seemed nearly as fast as my 5555 does now at home with WiFi and cable modem. And it was much, much faster than the 5555 is with a Supplynet cable and a Sprint 1xrtt phone. Of course Sprint's new method of disconnecting and reconnecting (supposedly transparently) between pages doesn't help matters.

whydidnt
08-20-2003, 08:31 PM
Great article Jason, thanks for taking all the time to compare these devices. This sure proves what all lot of has have been thinking about performance improvements with 2003.

I guess the one thing that seems to hit me in the face is how much of a fraud the 400 Mhz X-Scale CPU is. It seems we would all be cruising along in the left lane if Intel had just decided to continue developing/improving the original StrongArm CPU's rather than forcing a new technology down our throats.

This bears out when you compare the lower-clocked 1940 with the Samsung chip to the higher clocked 2215 as well.

AND -- what the heck did they do to the graphics system on my beloved 2215? How could HP gone that far back in this area. It seems all we hear about is the multimedia capabilities of the device and how it is meant as a "consumer" device. Someone dropped the ball on the graphics side of things.

Whydidnt

guinness
08-20-2003, 09:35 PM
How does the 1945 compare? If the StrongARM did so well against the XScale, how does the Samsung chip fare?

Jason Dunn
08-20-2003, 10:00 PM
Nice benchmarks, Jason. I would have liked to get some benchmarks on my e310 in comparison to the rest of the list...

You can go to the Spb Benchmark comparison page (http://www.softspb.com/products/benchmark/compare.asp) and compare the e310 we tested with other Pocket PCs in the list. Or you can download Spb Benchmark and test your e310 yourself. :wink:

srs
08-20-2003, 10:35 PM
[office space]yeah, I'm going to need you to go ahead and send me that wm2003 update for my ipaq [/office space]

seriously, though, HP should really consider releasing an update for the 3600 and 3700 series, as this proves theres really no technological reason behind not offering updates, and especially since this update may even improve the speed of these devices.

edit: and a word about the graphs.... I found them hard to read and compare because the order (and color!!) of the devices changes between each graph. The order and color should stay fixed for each device to make comparing graphs easier.

MonkeyGrass
08-20-2003, 10:49 PM
:roll: Sure would be nice to see those performance enhancements on my e740... :mecry:

Although, getting an improved, speedier PIE was really the only thing I wanted/needed. If that isn't the case, then I don't feel so bad.

... Or at least, I won't until I can lose this Trashiba for the 4xxx iPaq when it comes out! :mrgreen:

petvas
08-20-2003, 11:07 PM
After using WM2003 there is no going back. Microsoft did a great job optimizing this OS for all ARM Based (including XScale) processors. The improved reliability, the new kernel and the speed improvements make it a must...
And of course there is the added bonus of a muchly improved Connection Manager and Pocket Internet Explorer which is much more usable now than before...

ctmagnus
08-20-2003, 11:09 PM
the Pocket PC 2003 image for the iPAQ 3650 came from Microsoft - it a special testing-only build that won't be released to the public

Too bad. I sure would enjoy upgrading my iPAQ 3630. :?

Yeah, even if it's not civered unde Care Pack warranties. I have to do something with my 3670. It's WM2003 or Linux (http://handhelds.org).

Jason Dunn
08-20-2003, 11:17 PM
How does the 1945 compare? If the StrongARM did so well against the XScale, how does the Samsung chip fare?

My next set of tests is to compare the 2215, 1940, and 5555. Hopefully next week...

petvas
08-20-2003, 11:24 PM
How does the 1945 compare? If the StrongARM did so well against the XScale, how does the Samsung chip fare?

My next set of tests is to compare the 2215, 1940, and 5555. Hopefully next week...
That would be great!!! Could you possible add the 3970 also (with WM2003)? I hope it would be relatively easy to find one.... 8)

Jason Dunn
08-20-2003, 11:56 PM
Could you possible add the 3970 also (with WM2003)? I hope it would be relatively easy to find one.... 8)

I don't have a 3970, so that would be hard. :|

JTWise
08-21-2003, 01:07 AM
I too am very disappointed that PIE speeds have not improved. What good does it do to add a bunch of features to something that is so PAINFULLY slow :twak:

These benchmarks are a really good indicator as to why Dell is so slow to release the (yet again) delayed version of the OS upgrade. :mecry: We Axim users are keeping our fingers crossed that the guys at Dell will be able to get it right with the upgrade.

mv
08-21-2003, 01:32 AM
I too am very disappointed that PIE speeds have not improved. What good does it do to add a bunch of features to something that is so PAINFULLY slow :twak:
.

well, at least avantgo is a LOT faster. And I noticed that PIE is not a lot faster, but faster anyway. And it can shows more pages, too. I think that wm2003 is a great improvement, but all changes are under the hood.

jage
08-21-2003, 01:44 AM
I wonder if the CPU index has some OS function calls in it. Because, if it does, it's not a CPU speed indicator anymore. I've had my doubts about it since the first time I saw the results.

Ah, the methods are documented.



From Spb benchmark web page

CPU index

Indicates overall CPU and memory speed. It is the average speed of the following tests taken with the following weights:

Test Factor
Compress 1 MB file using ZIP 30
Decompress 1024x768 JPEG file 30
CPU test: Whetstones MWIPS 15
Copy 1MB using memcpy() 25


Well well... I hope that ZIP test and JPEG test have no I/O! Why? Because any changes in buffering and VM code in the OS will invalidate the results! They're also the most heavily weighted scores...

edit: The same also applies to memcpy test, unless the pages were precommitted - OS VM code changes might affect the test results.

mhowie
08-21-2003, 03:47 AM
That would be great!!! Could you possible add the 3970 also (with WM2003)? I hope it would be relatively easy to find one.... 8)


This is what I am looking for as well... a comparison of the 3970 w/ WM2003 and the 2215!!!

Thanks,

QYV
08-21-2003, 04:40 AM
Jason, thanks very much for the extensive testing/reporting! I have been waiting for this article since the launch of WM 2003 to see whether or not I could expect some nice speed gains with an Axim upgrade to 2003. While it's not apples to apples of course, your numbers give me a lot of hope - if Dell ever manages to ship the darn thing. :evil:

One thing to note about the much-maligned Xscale - in the "fixed" 255 form, it does seem to offer substantially more battery life than a 206MHz StrongARM. To me, performance that's almost identical with a big boost in battery life is a decent reason to upgrade. At the same time, of course, I suspect Intel could have done better, although we'll need comparisons of the new non-Intel ARM chips to support that claim.

quickstrike
08-21-2003, 05:53 AM
I've read all about the good things about Windows Mobile 2003, but I have a concern. I was trying out games such as Turjah II, Motocross Stunt Racer, and Bust'em and have noticed slowdowns and slightly choppy sound on my iPAQ 1940. I went to CompUSA and tried playing those same games off my SD card and found that the Pocket PC 2002 devices played them all fine, but the ones with Windows Mobile 2003 had the same problem. Has anyone tried these games also on their WM2K3 device? My guess is that this may have some relation to the slowdown that Dell fixed.

Abba Zabba
08-21-2003, 06:52 AM
I guess the one thing that seems to hit me in the face is how much of a fraud the 400 Mhz X-Scale CPU is. It seems we would all be cruising along in the left lane if Intel had just decided to continue developing/improving the original StrongArm CPU's rather than forcing a new technology down our throats.


I thought that the new processor gives improved battery life. What would be a good test is to test the battery life of the 545x upgraded to PPC 2k3 compared to that of the 555x. I thought Jason that you were doing some sort of article about that a few months back? What ever happened to that :?:

Marc Zimmermann
08-21-2003, 06:52 AM
I guess the one thing that seems to hit me in the face is how much of a fraud the 400 Mhz X-Scale CPU is. It seems we would all be cruising along in the left lane if Intel had just decided to continue developing/improving the original StrongArm CPU's rather than forcing a new technology down our throats.
You should keep in mind that the XScale is not only about performance, it is also about power consumption.

ErectionJackson
08-21-2003, 08:37 AM
Great work Jason, shows me that it is indeed a good move to upgrade my ancient (I bought it one year ago damnit! :twisted: ) 3870 :) But what I'm interested in is the stability of the 2003 OS compared to 2002. Do you notice less crashes and lock ups? Not that my 3870 crashes often, but since it's a big complaint from a lot of PPC users I was wondering if this has been improved as well

ChuckyRose
08-21-2003, 09:19 AM
Well, I guess I don't feel so bad about sticking with my old iPaq 3600 series iPaq with the first Pocket PC OS (2000) I would have lost a lot a speed it looks like with the upgrade. Maybe this will convince me to get a new PDA. Hmm. So many choices!

petvas
08-21-2003, 09:42 AM
That would be great!!! Could you possible add the 3970 also (with WM2003)? I hope it would be relatively easy to find one.... 8)
This is what I am looking for as well... a comparison of the 3970 w/ WM2003 and the 2215!!!

Thanks,
Then you should read This (http://www.gpspassion.com/fr/articles.asp?id=60)

petvas
08-21-2003, 09:44 AM
Great work Jason, shows me that it is indeed a good move to upgrade my ancient (I bought it one year ago damnit! :twisted: ) 3870 :) But what I'm interested in is the stability of the 2003 OS compared to 2002. Do you notice less crashes and lock ups? Not that my 3870 crashes often, but since it's a big complaint from a lot of PPC users I was wondering if this has been improved as well

Windows Mobile 2003 is much more stable and reliable than Pocket PC 2002. I barely have to restart anymore using my H5555. I reboot only after switching batteries... 8)

qmrq
08-21-2003, 09:57 AM
I too am very disappointed that PIE speeds have not improved. What good does it do to add a bunch of features to something that is so PAINFULLY slow :twak:

If you are trying to 'sync' pages from desktop for later viewing, you should check out iSilo (http://www.isilo.com/). Much faster than IE.

Chris Leckness
08-21-2003, 01:05 PM
Jason, thanks very much for the extensive testing/reporting! I have been waiting for this article since the launch of WM 2003 to see whether or not I could expect some nice speed gains with an Axim upgrade to 2003. While it's not apples to apples of course, your numbers give me a lot of hope - if Dell ever manages to ship the darn thing.

Sorry to chime in, I normally only read here, but I wanted to offer up these results I have had with the Axims.

http://www.aximsite.com/reviews/results.html

Enjoy.

david291
08-21-2003, 01:51 PM
It's nice to see 2003 so much faster in all those areas. There's one very important benchmark missing, however. Windows Mobile 2003 is significantly slower when it comes to reestablishing wireless connectivity after a power on. This capability is critical to our business application. My own benchmarks indicate the following troubling results in the amount of time it takes for wireless to reconnect when you turn on your unit (these results vary by about 10% depending on wireless card and driver):

PPC 2000: 1.2 seconds
PPC 2002: 3.4 seconds
PPC 2003: 12 seconds!

My technical sources tell me it is the fault of the otherwise useful new WZC (Wireless Zero Config) feature in Windows Mobile 2003. And from what I've been able to find out so far, there's no way to bypass all that extra logic going on at power on.

Jason Dunn
08-21-2003, 02:23 PM
One thing to note about the much-maligned Xscale - in the "fixed" 255 form, it does seem to offer substantially more battery life than a 206MHz StrongARM.

Do you think it's the CPU, or the evolution of better screen and battery technology? Let's see...the only real way to proove that would be to test a 3850 vs. a 3950, right? I think those units were identical other than that CPU and screen. It would be interesting to do some testing on this. I can't remember if, on those battery tests I published a few months back, I had any tests of a StrongARM vs. an XScale iPAQ with the screens turned off.

I'll say this much: regardless of whether or not the iPAQ 3650 is faster than more modern units, the screen and battery improvements are worth the performance loss. But I think we should have all THREE. :mrgreen:

Jason Dunn
08-21-2003, 02:25 PM
This is what I am looking for as well... a comparison of the 3970 w/ WM2003 and the 2215!!!

I've seen that another site has a comparison of a 3970 running 2002 and 2003, so you'll be able to compare those results with the 2215 results I've published previously. This 3970 review isn't translated to English yet, but when it is, I'll post on it here.

Jason Dunn
08-21-2003, 02:27 PM
I thought Jason that you were doing some sort of article about that a few months back? What ever happened to that :?:

Not me - I don't have any of the iPAQ 2003 upgrades yet.

Jason Dunn
08-21-2003, 02:30 PM
It's nice to see 2003 so much faster in all those areas. There's one very important benchmark missing, however. Windows Mobile 2003 is significantly slower when it comes to reestablishing wireless connectivity after a power on.

Very interesting! That's not something that Spb Benchmark can test, nor is it something that I'd even think of testing - I suppose because I don't use WiFi on my Pocket PCs very often. Thanks for the info! 12+ seconds is an eternity... :|

whydidnt
08-21-2003, 03:30 PM
I'll say this much: regardless of whether or not the iPAQ 3650 is faster than more modern units, the screen and battery improvements are worth the performance loss. But I think we should have all THREE. :mrgreen:

Amen to that statement. :werenotworthy:

txcas
08-21-2003, 03:31 PM
It's nice to see 2003 so much faster in all those areas. There's one very important benchmark missing, however. Windows Mobile 2003 is significantly slower when it comes to reestablishing wireless connectivity after a power on.

Very interesting! That's not something that Spb Benchmark can test, nor is it something that I'd even think of testing - I suppose because I don't use WiFi on my Pocket PCs very often. Thanks for the info! 12+ seconds is an eternity... :|

Twelve seconds does not sound right. My 3970 with WM2003 and a Dell CF Wifi card takes between 5-6 seconds. Anyway, I am sure it took me more that 12 seconds every time I had to change my networking settings from home to office or vice versa. Now it happens on the fly and I love that new feature.

daveshih
08-21-2003, 03:34 PM
I generally don't buy first-gen products. I wait until version 2 and usually by that time most of the roughness has been ironed out.

So when I saw PPC 2000, I waited. Then came ppc 2002, it was looking great -- by the way, where was SPB benchmark then? I swear if I'd seen the benchmark that Jason put out today, I wouldn't have bought my Jornada 567 -- a ppc2002 device).

I then considered the "brand" factor -- only buy something that a lot of people buys, even if that particular product isn't the top of the pack compared to competitors.

But then I saw Jornada 56x, which has everything IPaq at that time was lacking, and it's a 2nd-gen product (after the 54x series) in my mind. And HP was a good, reputable brand. For a while, those Jornada 56x were selling like hot cakes. Plus the fact that Dell was selling them with some good discount.

So, long story short, I bought a Jornada 567.

Now this.

But I probably shouldn't complain too much, cause even if I bought a IPaq then, I still wouldn't get a WM2003 upgrade now.

:2gunfire:

Thank you for tolerating my rant....

david291
08-21-2003, 04:02 PM
It's nice to see 2003 so much faster in all those areas. There's one very important benchmark missing, however. Windows Mobile 2003 is significantly slower when it comes to reestablishing wireless connectivity after a power on.
Very interesting! That's not something that Spb Benchmark can test, nor is it something that I'd even think of testing - I suppose because I don't use WiFi on my Pocket PCs very often. Thanks for the info! 12+ seconds is an eternity... :|
Twelve seconds does not sound right. My 3970 with WM2003 and a Dell CF Wifi card takes between 5-6 seconds....
How are you measuring that 5-6 seconds? Are you just looking at the wireless connectivity icon at the top? If so, you're right, it will indicate connectivity after about 6 seconds. But actual throughput is somewhere between nil and minimal until the 12 second point, depending on hardware. On a Toshiba e750, there is zero throughput until 12-15 seconds have passed. On an Axim with socket card, I can get a small packet or two through at the 6 second point, then wireless stutters (hangs, pauses, whatever you want to call it) until about the 12 second point before it is finally fully available. Your experience with your 3970 sounds similar to the Axim.

Jason Dunn
08-21-2003, 08:22 PM
So when I saw PPC 2000, I waited. Then came ppc 2002, it was looking great -- by the way, where was SPB benchmark then? I swear if I'd seen the benchmark that Jason put out today, I wouldn't have bought my Jornada 567 -- a ppc2002 device).

Pocket PC 2002 was launched in October 2001.
Spb Software House released their first application in January 2002.

Don't second-guess yourself - the Jornada 567 was, and is, an excellent device.

ctmagnus
08-21-2003, 10:05 PM
How are you measuring that 5-6 seconds? Are you just looking at the wireless connectivity icon at the top? If so, you're right, it will indicate connectivity after about 6 seconds. But actual throughput is somewhere between nil and minimal until the 12 second point, depending on hardware.

I can access the Internet ~ 5 seconds after I turn the wireless on, on my iPaq 5550.

carrigaline
08-22-2003, 12:48 AM
Hi,

I wonder if someone could tell me if the 3835 would have similar improvements to those that Jason found on the 3600 series with WM2003?

I recently bought a 2215 and can play my SNES ROMS on that unit no problem - but I could never play any on my 3835 - it would just not work right. I used the most recommended emulator but nada. The machine just seemed too slow.

:?: Do you think these improvements brought by WM2003 will allow my 3835 to run these ROMS like the 2215?

petvas
08-22-2003, 12:57 AM
Hi,

I wonder if someone could tell me if the 3835 would have similar improvements to those that Jason found on the 3600 series with WM2003?

I recently bought a 2215 and can play my SNES ROMS on that unit no problem - but I could never play any on my 3835 - it would just not work right. I used the most recommended emulator but nada. The machine just seemed too slow.

:?: Do you think these improvements brought by WM2003 will allow my 3835 to run these ROMS like the 2215?
Definitely, WM2003 would bring major speed improvements to your device... Of course you cannot install the H2215 ROM to the 38xx device...
You will have to wait for an upgrade to be made available by HP...
This makes me very angry as they do have the RTM bits since April and they simply do nothing about it... :evil:

mhowie
08-22-2003, 02:47 AM
That would be great!!! Could you possible add the 3970 also (with WM2003)? I hope it would be relatively easy to find one.... 8)
This is what I am looking for as well... a comparison of the 3970 w/ WM2003 and the 2215!!!

Thanks,
Then you should read This (http://www.gpspassion.com/fr/articles.asp?id=60)

Very good... thanks!

Len M.
08-22-2003, 04:27 PM
Jason,

Could you extend the benchmark comparison to cover storage card specs too, perhaps with a CF card Expansion Pack and both a 1.0 GB IBM/Hitachi Microdrive and a fast 1.0 GB flash memory card?

It would be helpful to see the impact of the different operating systems on that too.


Len Moskowitz
Core Sound

carrigaline
08-22-2003, 05:00 PM
Hi,

I wonder if someone could tell me if the 3835 would have similar improvements to those that Jason found on the 3600 series with WM2003?

I recently bought a 2215 and can play my SNES ROMS on that unit no problem - but I could never play any on my 3835 - it would just not work right. I used the most recommended emulator but nada. The machine just seemed too slow.

:?: Do you think these improvements brought by WM2003 will allow my 3835 to run these ROMS like the 2215?
Definitely, WM2003 would bring major speed improvements to your device... Of course you cannot install the H2215 ROM to the 38xx device...
You will have to wait for an upgrade to be made available by HP...
This makes me very angry as they do have the RTM bits since April and they simply do nothing about it... :evil:


Petvas,

Thanks a lot for your help on that one. I will buy the upgrade from HP - I think someone was saying it will be available in 2-3 weeks so thats not too bad.

How certain can we be about these SNES ROMS working on a 38xx with WM2003? Don't mean to split hairs on this - just want to make sure I am not wasting $30 which could be put towards a new ipaq. Is there anything that can be established from those benchmarking tests Jason has done that would indicate this objectively?

Thanks again Petvas and hope Greece is getting some relief from the heatwave in Europe!

Before I close - well done Jason for such a fine site - will be subscribing when you get the visa payment method up and running.

Jason Dunn
08-22-2003, 05:03 PM
Could you extend the benchmark comparison to cover storage card specs too, perhaps with a CF card Expansion Pack and both a 1.0 GB IBM/Hitachi Microdrive and a fast 1.0 GB flash memory card?

Unfortunately, no, I can't - I no longer have a Pocket PC 2000 3650. I can try to see the difference between 2002 and 2003 when I get my Axim upgrade though....

GO-TRIBE
08-22-2003, 06:23 PM
We must have that MS images of 2003 for the old iPaqs!!!!

0X

bothari
10-12-2003, 05:02 PM
We must have that MS images of 2003 for the old iPaqs!!!!

0X

Yes, who must we beg/bribe/"sacrifice to some dark gods" to get our dirty little mittens on it? My 3630-with-memplug-dualcf-and-256-cf is just begging me to get it....

and yes, i *would* be willing to pay for it AND sign a waiver of responsability in case it kills my device to get it... anyone from hp/ms reading this?