Follow Slashdot stories on Twitter

 



Forgot your password?
typodupeerror
×
Handhelds Hardware

Explaining Disappointing XScale Performance In Pocket PCs 133

JYD writes: "I found this new article on a Pocket PC web site where Microsoft talks about why XScale Pocket PCs aren't as fast as people thought they would be. Is it the OS? The CPU not supporting ARM4 properly? I wonder if the Linux port would run faster on 400 Mhz ... or did Intel screw up the CPU?"
This discussion has been archived. No new comments can be posted.

Explaining Disappointing XScale Performance In Pocket PCs

Comments Filter:
  • You think thats slow (Score:3, Interesting)

    by brejc8 ( 223089 ) on Saturday June 22, 2002 @01:14PM (#3749571) Homepage Journal
    My group has been working on a syhthesizable secure G3 card CPU and it will probably be the slowest ARM ever made.

    The CPU will be fully delay insensitive and asynchronous to stop power and clock glitch attacks.

    We are currently looking at 4 Mhz on 0.18 process.
  • by MrBandersnatch ( 544818 ) on Saturday June 22, 2002 @01:23PM (#3749606)
    a review I read showed a 400Mhz XScale performing at 50%-75% the speed of a 206MHz Strongarm chip. I would be really interested in some none OS specific tests that showed whether or not the XScale offers any performance benifit whatsoever - I know that it is supposed to scale to 1Ghz and has better battery life than the 206Mhz Arms but if it NEEDS to run at 800MHz just to perform at the same level as its older sibling then it is a waste of space.
    • Typical. found it [ourpocketviews.net]

      I was off in my above assessment BTW. Snippit from the French :-

      A small video test under Pocket TV as a proof: Ipaq @ 206: 23 Fps Xscale @ 400: 19 Fps! Xscale @ 200: 14 Fps!!
      • Simple arithmetic: if it was CPU-bound, halving the clockspeed should roughly halve the FPS. Suspect the graphics chip. BTW, having it beaten by an iPaq in a graphics benchmark sucks rocks. A friend of mine got a 50-fold speed improvement in iPaq graphics by rewriting the GDI driver. Either the gfx driver is broken or there's something badly wrong with the gfx hardware.
    • by Chanc_Gorkon ( 94133 ) <[gorkon] [at] [gmail.com]> on Saturday June 22, 2002 @01:43PM (#3749663)
      Well I found it and the performance is NOT 50-75 percent slower then an iPaq. From the numbers on pocketnow.com [pocketnow.com], the Toshiba e740 is actually ahead in most categories with exception of graphics. There's the real kicker. I don't think it's the Xscale so much as it's the ATI imageon graphics chip in it. This is also a new chip, and as the benchmarks prove, it's driver has a problem or so it would seem. I actually heard that it's kind of operating in a emulation mode of sorts (kind of like standard SVGA on a desktop). ATI should provide driver code to Toshiba and it can then be fixed in a flash. I have a e740 and love it so much. The Xscale is a nice chip and will indeed improve in peformance as it's flashed up, but in my book, the other features are worth more. The wireless works well, the dual slots are a godsend (WE DON'T NEED NO STEEEKIN SLED! ;) ) and the price is GREAT for what you get. All in all, I would buy another one or an updated one (like the Toshiba e550 coming out soon!). One thing I am looking for is the availability of the 3000 mah high cap battery. The standard is fine for day to day use, but when you use the wireless alot you hear a giant sukcing sound coming from the battery. The other accesory I would look for is the 99 buck adapter that goes on the bottom. You add that and you can attach a USB keyboard and also drive a SVGA monitor or a Projector with it and have your handheld run your Power Point stuff on the road.
    • I have also seen poor performance by XScale compared to: 1- StrongARM 2- ARM920T ( Samsung ) 3- ARM7TDMI ( Samsung, ATMEL, etc ) It would appear that in implementing the XScale design they have broken rules with standard ARM design and made their product perform very poorly with software written for other ARM processors. Sounds like Itanium running Win32 code.
      • The Xscale has a 10 stage pipeline, compared to the 3 stage pipeline in an ARM7TDMI, and 5 stage pipelines in the StrongARM and ARM920T The main problem is the much greater load result delay on the Xscale pipeline (3 cycles(?), compared to 1 cycle on the ARM920) which means that the instruction ordering needs to be significantly different to generate optimal Xscale code. A short code segment should clarify this:
        # cycles ARM7TDMI ARM920T Xscale(guess)
        LDR r0, label1 3 1 1
        LDR r1, label2 3 1 1
        ... interlocks 0 1 3
        ADD r2,r0,r1 1 1 1
        total 7 4 6
        On this trivial piece of code the Xscale is 50% slower at the same clock cycle than the ARM920T. However, this effect would not make a 400MHz Xscale slower than a 206MHz StrongARM by itself.
  • The tool from MS has marketspeak in place of information except for "We decided to support V4, we didn't bother to retarget our compilers to V5 for our traitorous former buddies at Intel, and if XScale's V4 compat is weak, that's their problem." It is, in fact, but that doesn't make MSFT's laziness any less lame.

    On the other hand, Intel often gives little thought to enhancing performance of old code on new processors. If memory serves me right, Intel's Pentium Pro ran 16-bit code embarrassingly slowly.

    -jhp

    • On the other hand, Intel often gives little thought to enhancing performance of old code on new processors. If memory serves me right, Intel's Pentium Pro ran 16-bit code embarrassingly slowly.

      No, Intel gave a lot of thought to that. It takes several years to develop a complex CPU like the pentium family. They just thought that Micro$oft would have a 32-bit operating system out by the time the PPro was released. Oops! Windows 95 wasn't completely 32-bit despite all the "32-bit" marketing and hype. And it was so new that everyone was still running a lot of 16-bit windows 3.x software. Of course, performance was much better if you were running a Real Operating System ;)

      • No, Intel gave a lot of thought to that. It takes several years to develop a complex CPU like the pentium family.
        It only takes a year or two to develop a relatively simple CPU like ARM or MIPS. RISC designs tend to be far more straightforward and simple. Many computer engineering students implement the MIPS architecture as an exercise. See the Hennessey book (Computer Architecture: A Quantitative Approach, 2nd ed. [barnesandnoble.com]) to get an idea of just how simple a processor like ARM should be.

        Besides, the much-vaunted new feature of the PPro was the CISC->RISC translator, and it shouldn't take much to rejig that to handle 16-bit mode more effectively if the market (asses that they are) demands it.

        -jhp

        • RISC designs tend to be far more straightforward and simple. Many computer engineering students implement the MIPS architecture as an exercise.

          yep. Been there, done that. Well, almost. Computer Architecture was probably my favourite class in Uni. We didn't implement it ourselves, but that was about 90% of the class lecture and notes. Starting with simple logic gates, we went through how to build registers, latches, ALU's, register files, all the way to pipelining. Fascinating stuff if you can stick through it all and have a great lecturer! It really gives you an appreciation of how the stuff works.

          Besides, the much-vaunted new feature of the PPro was the CISC->RISC translator, and it shouldn't take much to rejig that to handle 16-bit mode more effectively if the market (asses that they are) demands it.

          I didn't know that. All I knew was that Intel was betting on M$ migrating everyone over to 32-bit software by the time it was released to market. Considering their close deals in the past, I'm sure this was based on information that M$ had given them.

          To end this post on a non-anti-MS note, the CISC-RISC converter is software upgradeable. Recent Linux kernels provide a /dev/microcode device so that you can feed it a file (presumably) supplied by Intel. See http://www.urbanmyth.org/microcode/ [urbanmyth.org] for more information.

  • Amulet cores (Score:5, Interesting)

    by brejc8 ( 223089 ) on Saturday June 22, 2002 @01:25PM (#3749614) Homepage Journal
    The Amulet group [man.ac.uk] has been working for year to make a low power yet high speed asynchronous ARM processors.
    The Amulet 3 [man.ac.uk] runs at 120 MHz and consumes very little power. Most of all its asynchronous so when you dont have mych processing to do it just sits there consuming "no" power.

    They take a hell of a beating and still run. I connected one to a hamster wheel and you can see it here [man.ac.uk] running despite the power fluctuating madly.

    The only reason it only goes at 120MHz is because the memory isnt fast enough.

    Its a little strange that only three ARM production lisences were given out. One to intel one to motorola and one to Amulet group.
    • The Amulet 3 runs at 120 MHz and consumes very little power. Most of all its asynchronous so when you dont have mych processing to do it just sits there consuming "no" power.

      When you say 120 MHz do you mean that it has the equivalent performance of a 120 MHz ARM? I'd thought that an asynchronous chip didn't have a clock speed as such.

    • Um, I thought the whole point of asynch chips was to eliminate the need for a power hungry clock. If this chip is truly asynch, how can you say it runs at 120MHz?
  • Intel's Xscale architecture -should- have impressive performance. However, it seems that circumstances have conspired to keep it from showing its potential. In order to retain binary backwards compatibility, Microsoft has kept their software compiler rather basic, ensuring that it will work for multiple architectures. This lack of optimization also means that any architectural improvements Xscale has over Arm V4 or whatever it's called will not mean anything. Hopefully the history of Xscale will work like netburst architecture's history, where about 1 year after inception, software that makes use of its architecture efficiently (like SSE2 with the P4) will start to appear.
  • Stranding Users... (Score:5, Interesting)

    by Anonymous Coward on Saturday June 22, 2002 @01:28PM (#3749625)
    From the article: "We're not prepared to strand an installed base of over 2 million iPAQ users."

    Umm... right, that's why my PocketPC 2000 Cassiopiea E115 is now as useful as a doorstop as it has a MIPS chip in it.

    When I got my PocketPC, MS touted that 'software matters' - even in their publicity. Suddenly, they ditch all the SH3 and MIPS users and just support ARM in PocketPC 2002. Not only that, but applications like Terminal Services and Messenger they won't release for the older machines. I see a lot of people saying that this is becasue PocketPC 2002 is based on CE.NET - that's not correct. PocketPC 2002 is just another revamp of PocketPC 2000, which are both based on CE 3.0. So when it all boils down, it's just Microsoft playing marketing tricks. Net result of their decision - my £450 PDA became obsolete in 18 months.

    I now own a Palm.

    • by Cardbox ( 165383 )
      And that's why, as a software developer, I daren't target PocketPC. When I ported my application from PalmOS 3 to PalmOS 5 (the ARM one), I had to change ONE LINE of code (and the program remained backwards-compatible).

      If you ask the users, the current installed base of PocketPC systems is as follows:
      PocketPC 2002 - 1%
      PocketPC 2000 - 0.5%
      PocketPC "I don't know what version" - 98.5%

      We can target either of the first two quite easily, but the last operating system in the list has no programs that are compatible with it.
    • From the article: "We're not prepared to strand an installed base of over 2 million iPAQ users."

      Umm... right, that's why my PocketPC 2000 Cassiopiea E115 is now as useful as a doorstop as it has a MIPS chip in it.

      Sorry, there were only 1,999,999 users of that specific system, so it was below our threshold. :)
    • but I haven't any points. I'll just throw out the Compaq Aero series as another MIPS arch that was very quickly made useless. Not to say it was ever very useful. We bought five but the sync process was so shitty that we never used them.
    • We can use MIPS portables here.
  • by Qrlx ( 258924 ) on Saturday June 22, 2002 @01:28PM (#3749627) Homepage Journal
    Pocket PCs aren't as fast as people thought they would be. Is it the OS?

    It could be the OS, which is the obvious answer since it's a Microsoft OS, and this is Slashdot. But I don't know. I've never tried running anything other than PocketPC OS on the iPaq, and probably never will. (It's a work thing.)

    How did Microsoft become so popular? It was DOS, wasn't it? The program that ran on any x86 computer. Well, Microsoft should take a page from their previous success and allow a little more flexibility in PocketPC design. The main gripe that I and everyone else has about these gizmos is that they're locked into a 240 by 320 by 16-bit color display. That's lame, especially if one of the highlights of PocketPC is how easy it is to port your Win32 app. If you have to redesign all the screens to fit in a tiny-ass space, it's easy on the coders but hell on the systems analysts.

    It looks to me like Palm have a much more open approach, they are using the same tactic that established Microsoft's dominance with DOS back in the 80s. You can get that new Sony Clie' with TWICE the screen real estate (as in pixels) of ANY PocketPC available. Kind of a no-brainer if you ask me.

    Off to the solstice parade!
    • I believe the spec for PocketPC calls for a 320x240 screen. However, you can get WinCE machines that run at any resolution, far beyond any silly PalmOS device.

      Sure, you can get a Sony Clie with a 480x320 screen. But why would you want to, you'd have to put up with that sad excuse of an OS. Makes a pretty kickass (and damned expensive) oraganizer, I'm sure. What good is a 480x320 screen if it's about the same size (in cm by cm) as the other options *and* there's no real handwriting recognition?

      Personally, I still carry around a 5 year old Newton 2100u most of the time. I have an iPAQ as well, for development mostly. It has a 480x320 (lower DPI than the Clie, which means more space to write!) screen and a 162 MHz StrongARM. And a real OS, with the facilities to develop first-class apps while never touching a desktop. Seems like a no-brainer to me too. But we're on Slashdot, and it doesn't run Linux, so I'll just kick back and wait for the on-slaught of "h3y j00 st00p1d mac lover fsck u && UR pee-pee-pda!!!1"

      Have fun at your parade! :)
      • by aussersterne ( 212916 ) on Saturday June 22, 2002 @02:40PM (#3749850) Homepage
        No onslaught here.

        The Newton 2100 kicks ass. I used Palm and Windows CE before finally trying out a Newton 2x00 series. The Newton made me swoon.

        It's the best damn computing device out there, PC, PDA, or otherwise. I used to do my e-mail, my diary-keeping, my word processing, etc. on my PC in Linux, but now I even write my books and do 90% of my e-mailing on my Newton 2100 directly over ethernet. I read news on it, make travel plans on it, I have my household inventory on it (in Notion)... and I read BBC World News and Slashdot on it in Newt's Cape.

        The PC only gets touched every few days. The Palm and CE devices are long gone. I only regret that Apple killed the Newton, so there won't be a color version. :(
      • > Sure, you can get a Sony Clie with a 480x320 screen. But why would you want to, you'd have to put up with that sad excuse of an OS. Makes a pretty kickass (and damned expensive) oraganizer, I'm sure. What good is a 480x320 screen if it's about the same size (in cm by cm) as the other options *and* there's no real handwriting recognition?

        Um, maybe for viewing extremely clear and vibrant JPG's, watching widescreen movies, reading eBooks that don't strain the eyes, etc, etc. The Clie NR70 is just lovely, IMHO.
        • Again, that's what you use may use a PDA for. I don't really need a PDA for looking at clear and vibrant JPGs. Porn in the toilet is of the utmost quality, I'm sure. But it's not my bag. Never have eye strain with the Newton or my iPAQ, the black on green is of high enough contrast for me. I'm sure the screen is nice, but it doesn't provide anything for me that isn't good enough elsewhere.

          They keyboard is sure badass, though. A remarkable piece of industrial design.
    • Screw you man! I don't want them any bigger. Sure a nice big screen would be nice, but in something that's supposed to fit in your pocket, then the current specs work. The only way I would want a bigger screen would be if it was virtual (project on your wall, eye, whatever....). Now I WOULD like a 15 inch wireless web pad for at home. But I guess that'll wait for the mira. Also, that clie with twice the real estate is a heck of a lot bigger too. I like the 240x320, although I would not mind more pixels per inch. That would make things much sharper and clearer.
  • by brooks_talley ( 86840 ) <brooksNO@SPAMfrnk.com> on Saturday June 22, 2002 @01:32PM (#3749632) Journal

    Q: What could possibly have gone wrong?

    A: While we acknowledge that some peoples' perception is of something having gone wrong, we believe that any wrongness is unavoidable.

    Q: Well, some analysts say it's intel's fault

    A: We have implemented what we could implement, and don't believe there is any implementable implementation that would implement significant gains.

    Q: Analysts also say it will be 2004 before the issue is fixed

    A: It is too early to talk about 2004. That said, we are committed to delivering a good product.

    Q: This is really bad news for the Pocket PC platform

    A: Yes, it is. However, fortunately the issue is so small that this really isn't bad news for the Pocket PC platform.

    Cheers
    -b

    • I'm the VP of Marketing for a large Internet company whose name I cannot disclose in a public forum. I'd like to offer you a director's position in our marketing department. Name your price. Can you start on Monday?
  • They've probably used aggressive power saving on the chip to save every electron but at the expense of performance.

    Thats not such a bad thing, most of these things run address books and sync to email. The battery is the real problem with them, not the fact it can't encode video streams!

    Sure they'll get a few complaints, but nothing like the slating they've been getting for the battery life problem.

  • SUWANJINDAR (of Microsoft): "Our software remains the same. This is the same Pocket PC 2002 software that performs fabulously across other ARM processors (StrongARM 1110, OMAP710, etc).

    Well, that statement clearly deserves a +5 Funny.

  • THOUGHTS: Early reports based on those who own the Toshiba e740 Pocket PC 2002 device are telling us that XScale at 400 MHz performs slower than a StrongARM at 206 MHz on some tasks. This came as a surprise to many people.
    SUWANJINDAR: "We are aware that PXA250 (XScale)-based devices are not demonstrating the huge performance gains that were anticipated. That said, Pocket PCs continue to offer the best performance and the richest functionality vs. other handhelds on the market today."
    THOUGHTS: Some of those same analysts have said it will be 2004 until there's an OS that can use the XScale CPU properly. Is that an accurate estimate?
    SUWANJINDAR: "It's too early to talk about the next version of our software. That said, we're committed to delivering best-in-class functionality and performance while providing a foundation that enables our developer community to continue to innovate and build successful businesses on our platform.
    I find this sort of talk simply insulting. Couldn't he even find some irrelevant facts to spin in support of his cause? How stupid do they think we are?
    • that' depends on who you mean by "How stupid do they think we are?"

      Are you talking about your average consumer, or are you talking about the average slashdot reader?

      (Afer allhow many slashdotters would buy this product anyways?)
  • MS admits in the linked article that the OS is not "optimized". It fails to use the new ARM instruction set, and worse, does not seem to use the power-management capabilities of the XScale. Supposedly the Xscale uses half the power of the StrongARM, but battery tests on the new PPCs do not show this savings. This fix will be a while coming, as the next version of the OS [pdabuzz.com] does not appear to be optimized either.

    Interestingly, Asus in their upcoming Xscale PPC is coming up with workarounds, such as on the fly automatic clock and voltage throttling [anandtech.com]. So while the Xscale supports capabilites that MS is not using, the vendors are not waiting for next year for MS to get their act together.

    Hopefully the vendors will also figure out a way to speed up the terrible benchmarks [pocketnow.com] of the Xscale PPCs.
    • Terrible? It beats the iPaq in every area except for memory moves, and graphics. Memory moves are not that much slower either. Are we quibbling about mere microseconds?? I believe the problem could be the OS but may be more likely to be with the ATI chip. Rumor has it the driver has not truely been released yet.
      • Considering that its running at 400MHz compared to the 206 of the Ipaq, then yes, those benchmarks are terrible.

        Bonus *off* for replies.
        • For a category that has NEW silicon running it! Driver problems can create lots of issues. The imageon 100 graphics processor they are using in the Toshiba is brand new. My point is are we expecting too much out of a device that oculd have been rushed? I mean how fast do you need your contact list to come up?? As long as it plays MP3's fine, it works good for me. Also, this is nothing like a x86 chip. I t has all kinds of optimizations for power concerns and is also a RISC processor I believe. It's too soon to create any verdict on the Xscale. There's only ONE model out now (in the US....Japan has the really cool new Genio model that won't be here for a while). Also, HP iPaqs are due out soon as well. Palm is also going to Xscale too. The Xscale is a god processor, it's just too soon to issue sucks/doesn't suck verdict on it right now. Personaly, I like my e740 very much and owuld NEVER step DOWN to an iPaq. Besides I hate them sleds anyway (why can't they include a CF sled instead of the usless basic sled).
          • *** For a category that has NEW silicon running it!***

            Um, yes, it is new silicon. That doesn't make the benchmarks any better. It's new silicon with terrible benchmarks. What is your point? Saying it will get better later doesn't make it good now.

            And while you only care about MPEGs, some people care about performance and battery life. Some people run apps that use a bit more processing power than "contact lists", even if you don't.

            Bonus *off* for replys.
            • Re:It's the OS (Score:2, Interesting)

              by xswl0931 ( 562013 )
              I believe the point is that the benchmarks were not *terrible*. Terrible implies that the new processor is = the old 206 Mhz one, which certainly isn't the case. The problem is that people expect a 400Mhz processor to give about a 2x performance increase over the old 206Mhz one, but for a bunch of geeks, it's amazing how easily you people forget that 1Mhz != 1Mhz when you're comparing different processors (even if they both use the same instruction set), such as AMD vs. Intel. Speaking of Intel, remember that Intel realized that common folk are easily sold on a high Mhz number rather than instructions processed per clock cycle, so Intel obviously designed their processors (Pentium family and StrongArm family) to scale up to high Mhz.
              • THANK YOU! Exactly my point. I do believe it SHOULD be better, but there are SO many new things in the Toshiba e740 that it just was not possible to get everything right software wise. The Xscale, the imageon, the integrated wireless all make for one really really small and very complex device. The firmware is BOUND to have problems. This much should be a given. All things said, I have yet to see the people's problems with things like wireless (been pretty flwaless here), and video. There are annoying things like the buttons taking a vacation after using the wireless sometimes, a weird block in the messenger program command bar, weirdness with the notes app when storing notes on the SD card instead of memory, the ANNOYING "feature" of enabling the radio when doing a soft reset (I WANT TO LEAVE IT OFF IF IT WAS OFF DAMMIT!) and others. I have only had it for a week so it's hard to say if that's all the bugs, but those are the ones I notice. Personally, for a desktop, I would rather have a 450 MHz SPARC or RS/6000 instead of what I have. At least both Sun and IBM run more QA stuff then the PC manufacturers do! That 450 Mhz risc machine would run rings around any PC I could get (with exception of a Itanium, properly configured...). RISC machines are proof that doubling MHz will not necessarily double performance.
    • Without a compiler that has optimizations for the XScale, you will still get poor performance. So all the tweaks in the world to your existing code base will be for nought without a corresponding change in the compiler which is targeted for ARM7/9 cores and has only basic support for XScale.
  • Itroducing the Toshiba e740 Pocket PC 2002, now with a Genuine Intel XScale PR 166+* processor!

    *Actual clock speed 400 mhz

  • MS : "Moving to ARM V5 would break upgrade compatibility."

    Translation : "We can't or won't write portable code."

    There's absolutely no technical reason they can't take advantage of the V5 enhancements while still retaining support for ARM V4 and a common code base. This must have been a business decision, but I can't fathom the thought processes which led to it.

    • Comment removed based on user account deletion
      • Did you ever consider MS's other responses at face value, ie, they did not recompile for V5 because it would render incompatible all their partners and clients applications

        Yes I did, and they don't hold any water. I suggest you read your response above and consider why it doesn't make any sense (hint - the OS is not an application). Like I said, this has to have been a business decision.

        Having said that, your point about MS not liking multiplatform support is spot on. MS has some very competent programmers - that's not at issue. The problem is at the corporate decision making level.

        As to solving all their problems with a tarted up p-code system, well, if that's going to work at any sort of acceptable speed then they'll *really* need to optimise for the processor.
      • the next version of Pocket PC - that will surely have a .NET runtime in it. That will make all the problems go away, won't it? Just compile the apps to IL, target, and distribute.

        Let's hope your skepticism is justified. Because if it isn't, Linux as a platform will be in very serious trouble.

        Linux has no answer to cross-platform code, the one exception being Gnome with Mono. If that remains the only effort, and continues to attract hype and developer support, one day soon we'll wake up and find that the single viable open source platform to write to is under the technical direction of Microsoft.

        However did this happen?
  • I might add..... (Score:4, Interesting)

    by Chanc_Gorkon ( 94133 ) <[gorkon] [at] [gmail.com]> on Saturday June 22, 2002 @02:03PM (#3749729)
    This complaint was also based on the FIRST Xscale pda to EVER be released. Sure there's GOING to be problems. The iPaq started off with similar issues, but you don't hear anyone talking about it now do ya? There's alot of reasons that add up to create the total performance picture. Maybe Toshiba used cheaper internal ram? Maybe they need more memory for video (I think it has like 256 K maybe?? I don't know but I know it has dedicated video ram). The point is the performance on ONE Xscale based PocketPC does not make a prediction on how the others will perform. Also as these are flashable, we can expect even the Toshiba to get better performance as flash updates are made available.
  • This fits perfectly in Intel's "Megahertz sells" paradigm.
    Just push clockspeed up at any cost - who cares about performance? It's already running windows - so what can you expect?!
    • Re: (Score:3, Informative)

      Comment removed based on user account deletion
    • Just push clockspeed up at any cost - who cares about performance?

      What? It's true that clock speed alone isn't a valid measure, but it is certainly a very important part of the equation. Intel's "Megahertz sells" paradigm is turning out to be a pretty effective strategy. They have been able to jack up their clock speeds almost at will (2.533 GHz on the P4 now with a clear path to 3.3 GHz by the end of the year). On the other hand, AMD has been struggling all year to speed up their processors. The result is that Intel is leaving AMD in the dust [tomshardware.com]. As toms hardware puts it, "the Athlon design is already a bit outdated and is now reaching its limits."

      The consensus is that performance problems with the new XScale platform are because of poor software - not because of flaws in the hardware.
  • Comment removed based on user account deletion
  • When shopping for a PDA I don't remember speed meaning anything but "cost to much"

    People buy PDAs for cheap lap tops or simple organisers. Nither needs speed.

    Pocket PCs can get faster and faster while Palm Os PDAs outsell them.

    The Palm Os devices are cheaper and use less power.
    This is becouse they are slower.

    It's not speed... it's memory....
    Handspring Visors have memory cartrages and the Palm m500 use media cards so while Power PC devices play with added speed and don't get it Palm os devices get added memory.

    Thies things are just portable databanks they aren't for processing information just storing it.

    Want to play MP3s? Slap on an MP3 player... a sound chip that has an mp3 incoder built in and some added ram.

    Want to do presentations? Slap on a presentaion device.

    Go on the Internet? Snap on a wireless... (Unless it's built in)

    Play Quake? compile data?

    Hotsync with desktop...

    I'm looking for a keyboard and a wireless for my Visor (the i705 can't handle telnet) so I can use a shell account from my PDA...
    I'm not going to have any real computting power on a PDA. Thats not what a PDA is for.
  • The Intel PXA250 has only 32K/32K of cache, which means that any real application will experience an extremely high cache miss rate. The memory bus is 16 or 32 bits, and has a maximum clock rate of 100 MHz. So, if you're running the maximum width bus at its maximum speed, you're likely to see an instruction dispatch rate of about 50~100 million ops/second. That's slow, and there's really nothing to be done short of adding much more cache.
    • I haven't worked on X-Scale cpus, but I've certainly worked a ton on StrongARMs. A StrongARM SA-1100 has 16KB code and 8KB data cache. It's half-line 'dirty flag' write-back, 32 bytes per line, 32-way associative, and round robin evicted. The cache is very rarely the bottleneck - low MHz is more often the problem.

      It's usually pretty hard to thrash a code cache to the point of it being the bottleneck. You pretty much have to deliberately write code to do that. For reference, an Athlon has a 64KB code cache, and that's running at a far higher speed than both of these ARM processors. Your figure of 50-100 million ops/second assumes an unreasonable 100% instruction cache miss rate. You'd have to have a program totally devoid of loops to achieve that. At a still unreasonable hit rate of 95% you'd still get 96% ((95*400+5*100)/40000) of full performance.

      My guess why the real world performance is so bad is probably Microsoft's lack of optimization specific to the processor. There's a few trade-offs Intel have made to get the clock higher, including:

      • The latency of the barrel shifter is now 1 clock, instead of free. The compiler will no doubt assume the latter, as will tons of hand coded assembler.
      • The multiplier has higher latency (I think), but you can pipeline it along side loads, stores and arithmetic (nice). It won't have been compiled for this.
      • Loads/stores have higher latency, but you can pipeline them along side other stuff. The load/store multiple instruction has been "obsoleted" such that it's slower than separate instructions. That's pretty major for a compiler to ignore.

      I certainly wouldn't expect to be able to take code targetted for StrongARM and see all of the performance increase that 200->400 would indicate. I can imagine hand coding assembler to work around the latencies and getting near to 100% performance. It's not that hard for a modern compiler to work this out either - a current day x86 is far more difficult to target than an X-Scale.

  • by zealot ( 14660 )
    Before blaming Intel for going with an Arm 5 core with "slow" (slow being relative, as the benchmarks vary) Arm 4 emulation, remember that all they did was produce a CPU for the embedded market that can run on batteries. The MS Pocket PC market is just one market for these processors. They want them used in cell phones and powering all kinds of devices, just like the StrongArm did.

    Obviously, they felt that the majority of their customers would want an Arm5 based device. Wait a few months, and you might see some pretty impressive cell phones or linux based devices that use Arm5.

    The complaint against Intel is only legitimate if their Arm5 scores are terrible. Otherwise it is the fault of the device maker for using a chip that doesn't perform well for the task at hand, or MS for not optimising.
  • by jeffmock ( 188913 ) on Saturday June 22, 2002 @07:05PM (#3750803)
    It's important to differentiate between architecture optimizations
    and CPU specific optimizations. The ARMv5 instruction set is a
    relatively minor architectural tweak to the ARMv4 instruction set.
    The names give you the impression that it's some grand change between
    v4 and v5, if a technical guy did the naming it would be ARMv4 and
    ARMv4.01. ARM is playing some games with architecture naming
    to protect their business position with patents in a silly way.

    ARMv5 adds a couple of new instructions over v4, an instruction to count
    leading zeros in a register (which a compiler would likely never
    use), and a better method of switching between the ARM instruction
    set and the 16-bit Thumb instruction set. The later isn't
    relevant for PocketPC since Thumb mode isn't supported. I think
    v5 might having a new debugging hook as well.

    The new XScale parts are ARMv5te, the T is for the 16-bit Thumb
    instruction set, which no one seems to care about. The "E" adds
    some DSP oriented instructions that are pretty interesting for
    media codecs and such. They are the MMX equivalent for the ARM
    world. They likely won't improve performance of the general
    purpose aspects of the platform.

    I think it's a red herring to chase Microsoft for not optimizing for
    the ARMv5, the changes are really small and I don't see any
    performance impact, certainly not if you have to maintain another
    version for all of the strongARM based products.

    Now, as far as CPU specific optimizations for the PXA250 (XScale)
    implementation of the ARM architecture. IMHO Intel chased
    MHz and left behind a lot of good sense about system performance.
    The high order bit is bus performance as others have already
    pointed out.

    In addition to the bus performance, Intel made many tradeoffs
    to optimize for clock speed: The 7-stage pipe has a 4-clock penalty
    for a mis-predicted branch. This is compared to the circuit
    design heroics in the strongARM that implements "all branches
    are 2-cycles". The Xscale approach is much more complicated, it
    probably doesn't perform any better, but you get a high clock speed.

    Intel adds clock cycles to all load/store-multiple instructions
    in Xscale. This is a pretty big deal in ARM since they are
    used in the entry and exit of most C functions, in memcpy(),
    and any time you are moving chunks bigger than a register.

    The load-use penalty is bigger in Xscale. This is a pretty big
    deal in ARM. The ARM instruction set is pretty compact. It is a
    RISC processor, but the combination of shifting operations
    combined with ALU operations makes it possible for a good compiler
    to generate reasonably compact code. As a result, it's harder
    for a compiler to put instructions between a load and instructions
    that use the destination of the load. This is another trade-off
    in Xscale that allows a higher clock speed but hurts performance
    otherwise.

    I go on too long, but the DEC designed strongARM used in the SA1100
    is a tour-de-force of clean implementation and balanced system
    performance. It's amazing that core was designed in 1993 (I think,
    someone please correct me) and is still the leader for handheld
    apps. The Intel guys went after clock speed at the expense of
    everything else in Xscale and it will probably never optimize well
    for a platform like PocketPC.

    jeff

    • Somebody, please mod this up because jeff is right damn it!!!

      I've worked with both the SA-11x0 (StrongARM) and the PXA250 "Cotulla" (Xscale) CPUs and everything jeff says is pretty much on the money (except the CLZ instruction is far from useless, it's *awesome* for fixed-point logarithms, dude).

      Also, the DSP coprocessor in the X-scale is about as useful as tits on a bull for codecs with 16-bit data streams. You spend so many clocks marshalling data around to get it in and out of the thing that it's *much* more efficient to use the MAC instructions native to ARM v4 on normal registers! Even the Intel engineers who put together their IPP's [intel.com] have avoided the DSP coprocessor since it provides no real advantage.

      It's pretty clear to me the v4/v5 thing is a red herring. Let's face it DEC was much better at putting out a general purspose ARM-based CPU than Intel.

      • Well, comparing the Alpha (designed in the 80's) to the Itanium, it seems DEC was much better at designing all kinds of CPUs than Intel.
      • I likewise have worked with both the StrongARM and the Cotulla and Sabinal flavors of the XScale. The performance problems are real and in fact it was very difficult to get Intel to own up to it. Only after we had clear data showing what was going on that they came forward and revealed to us that there was a problem. I really fealt that Intel was not being very open about these issues. There is nothing particularly innovative with the V5 architecture and Intel will try to hype it for as much as it's worth but as the previous poster noted, it's more a variant of V4 with deeper pipeline and the extended DSP functions. Overall, the ld/str multiple performance is pathetic without changes to your standard C library routines.
  • I love this comment:

    We are aware that PXA250 (XScale)-based devices are not demonstrating the huge performance gains that were anticipated. That said, Pocket PCs continue to offer the best performance and the richest functionality vs. other handhelds on the market today.

    Translation: We know your new car only goes 40mph instead of the 65mph you old car did, but it beats a bicycle, doesn't it? (credits to Jim S for that one).

    Even better:

    I think the market expectation of what performance on a 400 MHz processor vs. 206 MHz processor has been unreasonable.

    Not at all. The process is almost twice as fast, I don't think it is utterly unreasonable to expect the product to be at least one and a half times faster.

    But my question is, how is the battery life on one of these things? If it really is the 12-16 hours instead of the 8 currently then the XScale is still a worthwhile bet.

Ocean: A body of water occupying about two-thirds of a world made for man -- who has no gills. -- Ambrose Bierce

Working...