Intel Please Read!

[cideo]

We now know that Microsoft has zero plans to optimize for XScale. The question then becomes what kind of raw speed increases do we need in an XScale chip to allow performance gains in existing programs written for the StrongArm? When will OEMs have those chips to release in PPCs or can we see those gains by overclocking the 300MHz chip?

[Cruzz]

Originally posted by Gamma Ray
We now know that Microsoft has zero plans to optimize for XScale. The question then becomes what kind of raw speed increases do we need in an XScale chip to allow performance gains in existing programs written for the StrongArm? When will OEMs have those chips to release in PPCs or can we see those gains by overclocking the 300MHz chip?

Well, the way I see it, no matter how high you raise the clockspeed, the speed still won't increase too much above the Strongarm as the software still deals with it nearly the same way as it does with the 206MHz processor. As a slight positive, this might also mean that the 300MHz Xscale is nearly the equal of the 400MHz model.

[r0b456]

The thing to keep in mind are that Microsoft has stated time and time again that they are committed to ONE standard. If Intel's chip does not abide by that standard or offers Poor Performance to that standard, then stop buying it. The thing is, Intel does not sell their processor to You, average Joe-PDA. They sell their xScale chip to Compaq/HP, Toshiba, Dell, Viewsonic, etc. You make your statement to Intel AND The OEM by not buying xScale based PDA's. Tell the OEM's this is your plan. The OEM's will stop buying xScale chips from Intel, and Intel will put effort behind developing a REAL improvement to our current processors. Instead of releasing Arm 5 chips and then say "We're Intel. We say the world must now adopt OUR technology. We Say So.", Intel should be forced to follow the standard that is currently in place and rather than just develop what they WANT, have them develop faster Arm 4 processors.

This whole mess is NO different than if Intel suddenly released the Pentium 5 (Or the Pentium Pentium) and it performed WORSE than the pentium 4 -- even at TWICE the CPU speed. Intel then comes out and says that it's not Their fault -- the processor works fine so long as you use it with an operating system that DOES NOT YET EXIST. It's not even in development. No. Instead, Intel expects everyone who ran out and bought these processors to WAIT for Microsoft to catch up and develop a WHOLE NEW OPERATING SYSTEM simply because Intel didn't feel like following the same Standard as everyone else.

That's what we have. And I am TIRED of Intel. IMHO, Intel started sliding downhill the day they slowed down production on Processors and started making little plastic USB Microscopes. I am perfectly willing to wait to adopt Texas Instruments or even an AMD entrance to the Arm market. At this point, I'm willing to wait for a Curusoe based Pocket PC before I blindly commit to Another Intel based unit. Between Intel losing ground to AMD, this whole PXA250 fiasco, the lawsuit with Intergraph and the Itanium chip, and several other Stupid moves by Intel over the last several years, I am ready to wash my hands of them. I will continue to use my Intel based Pocket PC. But I won't be happy about it. I'm so frustrated with the whole mess, I'm going to send e-mails to Dell, HP, and Toshiba and tell them how unacceptable the xScale situation is and that I will refuse to upgrade my PDA to any of their upcoming Pocket PC's until they stop using xScale processors. I have been very interested in the upcoming 5400 series iPAQ's and the Viewsonic v35 units. But whatever slight upgrades these units have, I will continue to use my 3835 iPAQ and it's ARM 4 processor. And if every Pocket PC on the market is xScale based when my 3835 dies and needs to be replaced, I'm going Clie. I love the Pocket PC, and I'll be throwing away hundreds of dollars in accessories, but the point of a PDA is to be Usefull. I refuse to drop $600 to UPGRADE to a SLOWER machine. The thought is not only ridiculous, it's insulting.

[notx]

Originally posted by r0b456
...This whole mess is NO different than if Intel suddenly released the Pentium 5 (Or the Pentium Pentium) and it performed WORSE than the pentium 4 -- even at TWICE the CPU speed.

Hmmm.. Ever hear of the Pentium Pro? Ran normal Windows like crap. P100 could beat a PPro 200. Really.

It however was increibly effective at working 32Bit Multi-Threading jobs. Quad PPro200's can sometimes beat Dual P2 400s. A single PPro 200 could sometimes beat a P2 400 (been there, done that).

I would bet that the xScale sereis is going to be a similar event.

[scata]

Whats the point of shouting Intel down, all they have done is released a new chip with better power managment, and with a newer ARM revision.

It's MS that is at fault here, I've been programming for well over 15 years now, and never have I written a program, and said "thats it I'm finished now, and I'm not going to to update it run on newer hardware"

Unless >20% of Win CE was developed in ARM V4 assembly, it would not be a big job to optomise it for xscale, I've not actualy checked this, but I bet Intel have released a compiler for use with their xscale chips (or if they haven't they will soon)

All intel are doing is looking to the future, like MS *SHOULD* be doing, lets face it, how many NEW PPC's are using the old chip

The only problem I see with MS optomising for xscale, is the fact that they will be locking the code into a manufacturer specific processor, and it would force PPC OEM's to use that chip.

Scat.

[r0b456]

Ok, Hypothetical:

Cyrix releases a new RISC processor ideal for PDA's and other small devices. Low heat, Low power requirements, zippy speed. It's 15% faster than the ARM -- so long as the software is designed to take advantage of it, and has 30% better battery life. Microsoft, Seeing into the Future, sees this and completely rewrites a whole new operating system for the new processor. Rather than continue to support and improve the Existing Pocket PC operating system for the StrongARM and xScale devices already on the market and being used, they instead throw all of their attention to supporting the NEW technology. Essentially telling all existing Pocket PC owners to Get with the Times and Upgrade their Pocket PC.

Now you've got separate groups of Pocket PC owners. You've got the ARM-based users, and the Cyrix users. Were now at a time where you've got Millions of Pocket PC users harping developers for a _____ version of their program because they currently only support the _____ CPU. You've got dozens of posts on these specific boards Daily by confused Consumers who do not know what kind of CPU they have -- all they know is that they have a Compaq Pocket PC and they want to play Tetris.

Now, as an xScale user -- someone who purchased a TOP OF THE LINE device, paying Top Dollar, less than 1 year ago -- how would you feel knowing that Microsoft is no longer interested in supporting your PDA? Microsoft is now Looking to the Future. So what do you do? Do you just shrug your shoulders and upgrade -- even though the PDA you had already did everything you needed it to do, except the New PDA has more Support and 30% better battery life?

It's easy to Sit there and say you can make Microsoft's decisions for them. It's REAL easy -- YOU do not have to live with the consequences. You do not have to feel the backlash from the MILLIONS of angry customers. To some people it may seem like a simple solution -- just support multiple processors. However, this violates one of the Cardinal Rules that Microsoft established for the Pocket PC 2002 operating system. This had been made clear to EVERY OEM. And despite Microsoft's clear message, the OEM's still chose to put ARM 5 CPU's into devices that were Optimized for ARM 4. It should not be Microsoft's responsibility to fix the problems created by this action.

[scata]

There is a simple answer to all this, chip makers need to develop compilers that can take the MS source code and target the build for their own processor, that way MS keeps it's one codebase and the optomization is done via the chip makers compilers. Much in the same way as using intels compiler to target the P4.

This kind of development technique has been going on for years, I've used it since the PPro days.

This would also open up the chip market to the PPC as well, how about a cruso/athlon/p4 powered PPC?

[r0b456]

Which is exactly why I hold Intel more responsible than anyone else. As a Consumer, I should hold the OEM responsible -- they are selling me a PDA for more money because it is "Faster" when it really is not. In turn, the OEM then gets angry at Intel for selling them processors that are SUPPOSED to be faster but really are not. How Microsoft got dragged into carrying the responsibility, I'll not really figure out. You cannot really fault Microsoft for doing exactly what they Said they would do -- support 1 standard. And we cannot, Should Not, force Microsoft to go back on that promise because of an error made by the OEM's and Intel. Of course, this is entirely IMHO, so everyone is free to disagree with me.

1.) Intel sold PROPRIETARY processors that did NOT have 100% backwards compatability, let alone actually being an Improvement.

2.) The OEM's installed these processors into devices and sold them to consumers, and charged MORE for them, even though they had no performance improvement.

3.) Intel has Failed to release Compilers or software updates to the public or to Microsoft to Fix the problem. Not to achieve true xScale performance, but to AT LEAST perform as well as the StrongARM standard.

And so someone is sitting around saying that when you add all that up, it is Microsoft who is responsible.


So here is what we are left with -- Microsoft will not support the xScale specifically. They have no plans now or in the future to release device specific optimizations. This means that MS bundled apps in Pocket PC 2002 and CE.Net like Media Player, Word, Excel, etc and the OS it's self will not be modified to take advantage or Compensate for the xScale processor. Instead, Microsoft will be using their R&D resources for improving the Existing, ARM 4 based code.

Several third party developers that are facing the biggest xScale related issues are already releasing xScale-specific versions of their software. Multimedia applications such as Pocket TV, and high-end games such as those from Jimmysoftware.com are now being re-released with xScale optimizations. This extra effort and support does not come cheap, and as such some software that was once Free, such as Pocket TV, must now be Purchased in order to receive these optimizations. So not only has this created a burden for the end user, but now Pocket PC software developers are now feeling pressure from consumers to start offering xScale-optimized software. Only to be criticized if they Charge for their extra efforts. It has also created specifically what Microsoft did Not want -- a world of Pocket PC software divided into several groups based on what kind of CPU that you have.

Unless there was something going on behind the scenes, it seems rather foolish to blame Microsoft for the mistakes of others. I'm not an expert in all of this, simply an observer. But I can say with some degree of certainty that if I were Microsoft right now, I'd be pretty peeved at Intel AND the OEM's. I would fine all of the OEM's for breach of contract for not abiding by the requirements of the Pocket PC 2k2 License (which regulates what Type of hardware can be used). I would then either demand Intel resolve the issues with the xScale processor OR the OEMs move over to strictly Texas Instruments produced ARM chips.

It's truly a Shame. The Millions of dollars that Intel spent on developing a half-assed xScale chip (Intel has admitted it has issues other than the ones we have touched upon) and releasing it Unfinished to make a quick buck, could have instead been used to develop faster ARM processors, or a True StrongARM successor.

[scata]

I think that Microsoft is partly responsible, by not being more flexible with their approach to this problem, lets face it xscale is now becoming the de-facto standard processor for PPCs.

It's no use burying your head in the sand, and hoping a problem will go away, you have to look at the market, and respond, not stick rigidly to your original idea of the way the world should work.

Don?t get me wrong, I?m not a Microsoft basher, I use their products everyday, I think they have an excellent product portfolio, but they need to keep Win CE moving forward and able to perform optimally on the de-facto standard hardware.

From what I?ve seen on these boards, I?d say a good 40% of people who buy a PPC are using it for entertainment (Video/Music/Games) as well a basic PIM functions. By not moving Win CE forward there is a chance that it will loose market share as these people loose faith in the product and look elsewhere for a device to fill these needs.

Some people would love this to happen, but I personally think it would be a disaster, as I think Win CE is driving the whole market segment forward. (look at the new Sony, Palm, and Sharp devices)

[r0b456]

Scata,

You failed to consider my Cyrix analogy. Nearly a dozen times you mention Moving Forward, and this being Microsoft's responsibility. This is all good and well -- I'm all for moving Forward. However, Intel does not OWN the Pocket PC. Nor does any OEM. When people refer to Pocket PC's, in General, do they say "The Intel Pocket PC" or "The HP Pocket PC"? Everyone knows it is the Microsoft Pocket PC. And yet the Pocket PC is being taken away from Microsoft by bad decisions made on the parts of Intel and the OEM's.

Yes, moving forward is good. Yes, the Pocket PC's need to hold on to their Multimedia superiority. But they Cannot do so at the cost of abandoning every Existing Pocket PC. I find it hard to believe that Microsoft Never planned on moving forward. On improving upon the Technology in both Software and Hardware. And I am Certain that such a plan would do so in a manor that would not instantly Screw the 99.9% Existing Pocket PC population. You calmly hold your position right now, saying that xScale has already been decided upon as the Future, so Microsoft might as well Give In. Yet you fail to truly put yourself into the possition that I stated in my hypothetical. More than likely, you own an xScale device, so you are seeing it from an xScale point of view. What IF Microsoft came out Tomorrow stating that they Would support a new hardware standard, and they Would move forward -- but not to xScale. Instead, all Pocket PC's from here on out would support a whole new Standard. All existing Pocket PC's running ARM or xScale processors are no longer supported. It's crazy, right? Not only are you Abandoning your existing supporters, but you're also saying that all of the existing Pocket PC software that you have worked so hard to accumulate is now Worthless. Even though the new OS will have built in support for the new Processor, your individual third party apps will need to be replaced by new versions which Also carry support for the new processor.

It's not a Hypothetical that is THAT hard to believe -- that's exactly what you're going to see with xScale. Say Microsoft gives in and re-releases Pocket PC 2k2 with xScale optimizations. xScale is now the standard. The OS is Optimized for it. 99% of your Pocket PC apps, however, are STILL ARM 4 based. We already know that xScale does NOT offer 100% backwards compatability with ARM 4. Having an optimized OS would not change everything. Some apps may still function OK. But others may still run Slow. Or Too Fast. Or Not at all. This decision would severely damage the Pocket PC community. This is NOT how Microsoft wanted to roll out new technology. It is NOT how they envisioned "Moving Forward". You do not accomplish anything by blindly running full speed ahead. You do so intelligently, and by making sure that you do not leave your followers too far behind.

If you STILL believe that this is a Microsoft issue, then consider this -- Toshiba uses a PROPRIETARY video chip in their e740. It was to supposedly make it "Faster". However, the drivers for this chip Suck. Many, Many programs, especially Multimedia, are Worse for having it. People are STILL waiting for Toshiba to release a software fix for this hardware, who in turn is waiting for ATI to release a better driver. In your opinion, is Microsoft required to step in and release a new version of Pocket PC built around this ATI chip, because of poor planning and development on the part of both ATI and Toshiba? Your argument is that Microsoft is Obligated to change because xScale is now the Standard. Ok. So now HP has released all of their new iPAQs with the ATI chip, as well as Toshiba, and you also have devices from Viewsonic and Dell which also use it. Why? No one really knows. Supposedly the new chip uses less power than the previous graphics controller, so it allowed the OEMs to use smaller, cheaper batteries and still get the same battery life. So now you all have Slower, less Efficient PDA's because ATI sold the OEM's some Piss-Poor video controllers, and the OEM's, being geniuses, Sold them to you -- and charged you More for it because it's "Moving Forward".

So by some logic that I still fail to grasp, Microsoft must now step in and clean up the mess created by the OEM's and by ATI. They must now take it upon themselves to completely change the Pocket PC OS so that the basic system it's self can run acceptably with the ATI chip. And even after that's said and done, and the non-ATI Pocket PC owners are completely segregated, the New Pocket PC owners must STILL replace roughly Half of their software like Games and Media Players because OS upgrades alone cannot change the way these 3rd party programs use the new PROPRIETARY hardware.

So what does it accomplish? Very little. All you're basically doing is taking support AWAY from the millions of Existing Pocket PC owners. You're not instantly improving a widespread problem. It's not like Microsoft is going to go into the source code and add a line which will instantly allow people to get 60 frames per second in the FREE version of Pocket TV, or make device specific games like Turjah work on the xScale hardware. Those programs (the ones that have the most trouble with the hardware) will STILL need to be re-written around the xScale hardware. The only thing that will change is that the built in apps like Media Player, and the OS it's self will take advantage of the xScale's Speed. This can be good. But good enough to alienate so many loyal users? Like I said, it's easy to sit there and say Microsoft should do this, or Microsoft should do that. But it's a completely different story when it's Your baccon on the line.


IMHO