I left that bit out on purpose :tongue:

It's okay, you know some of us do research :grin:


I wouldn't be surprised if the cards were programmed natively, they could manage even higher quality.
AFAIK all the drivers use compilers to convert from the "asm" in DX to their native format. I'm guessing they have longer instruction limits than the DX spec enforces, and possibly more optimal code paths.

That's indeed very possible - by reading carmacks .plan on specific nvidia extensions etc, you get the idea that neither OpenGL and especially not DirectX suits all the GeForce cards all that well. Same goes for vs/ps2.0 performance (without cheats ;)). I guess hardware vendors should design their card around the API instead of trying to "emulate" it...


DX10 / OGL2 speced cards will really show some impressive effects.

DX9 already looks pretty damn good :P


P.S. GL hasn't stopped being updated, it just moves so slowly it just looks that way.

Mmh, I hope I didn't say it's not being updated. It's just too bad you have to rely so much on vendor extensions to do anything on GL :/ - especially since vendor extensions also require the platform-dependant "getprocaddr" type calls.
Posted on 2003-12-03 16:44:29 by f0dder
I'm guessing they have longer instruction limits than the DX spec enforces, and possibly more optimal code paths.


Well, in short, ATi R3x0 IS the DX specs :)
That is to say, ATi's plans for the R300 were the prototype for the DX9 specs, much like the GF3 was for the DX8 specs.
ATi's design follows the specs very closely, it helped to define these very specs. In fact, there is not even specific ATi extension for the shaders in OpenGL, since the ARB extension is basically ripped from DX9's ps2.0, and that is pretty much optimal for the R3x0 series anyway.
GeForce FX on the other hand, has its own extensions, and they differ quite significantly from the ARB/DX shaders.
The most important difference is that you can combine integer and float operations in the same shader (because unlike ATi, NVIDIA doesn't emulate integer operations on a float pipeline, which would make such an option useless, but instead they have a float unit bolted on to the integer pipeline). It's not useful for anything other than higher speed and lower quality rendering than what ps2.0/ARB can offer though. Not something you'd want for Maya, I guess. Well you don't want it for anything really, but NVIDIA will decide that you want it, so you can get half-decent framerates in popular games, through driver cheats.

DX10 / OGL2 speced cards will really show some impressive effects.


I would not put them in the same class... It seems that OpenGL 2.0 will be mostly what Direct3D 9 already is.
By the time DirectX 10 comes out, there will be an entire new generation of hardware out, so DirectX 10 will probably go well beyond what OpenGL 2.0 will have to offer, and we'll have the same story all over again.
Besides, who would use OpenGL 2.0? Most developers are already using Direct3D 9, and it works fine for them... Why bother changing to get the same, knowing that you'll have to change back again eventually, because you're behind on Direct3D?
I think OpenGL may well get into big trouble on the PC... well it is already, Carmack is the only one still using it.

P.S. GL hasn't stopped being updated, it just moves so slowly it just looks that way.


What's the difference? :)
Posted on 2003-12-03 16:58:49 by Bruce-li
This is programed via a DX interface.


Where did you find this bit of info, Mirno?
I could not find any info on the pages regarding whether they used OpenGL or Direct3D.
Posted on 2003-12-03 17:19:53 by Bruce-li
hmmmm,

I have not seen the numbers for the Windows/DirectX supercomputers yet. Where you can just go and buy off the shelf a SGI box with the specs I posted in without having to build a much bigger one, is it because there number are NOT there ?

For all the spin and whinging about the lack of data, when its there the response is not.

Lets hear it again for the dummies.

FAST graphics mean dedicated BIG iron running the flight simulators for the military, NASA projects, the Jet Propulsion Laboratory and tasks of this like and size. Where is the Windows box and its directX ? How much spin will you need to generate to avoid the obvious, you don't run high end graphics on Windows/directX boxes.

I would guess that the guys floating around here who actually know their way around big iron would just about die laughing at the idea that kiddies toys are up in the big league.

This is the time to put the numbers where your mouth is instead of your foot. :tongue:

f0dder,

Perhaps hutch should stick to the things he knows about: PowerBASIC, writing superfast HAND OPTIMIZED ASSEMBLY string scanning algorithms that can beat ANY HIGHLEVEL algorithms WHATSOEVER, and dabbling with philosiphy and political ramblings. And preferably on his own ego-rubbing board, and let us have peaceful and techical discussions here.

Thread carefully here, for someone who has in the past walked around with their heart on their sleeve crying victimisation, you run the risk of being told to go blow your nose for the impertinence.

If you want to start slanging off at people, they will stop treating you like a charity case.

Make sure you both have PHUN.

http://www.asmcommunity.net/board/cryptmail.php?tauntspiders=in.your.face@nomail.for.you&id=2f46ed9f24413347f14439b64bdc03fd
Posted on 2003-12-03 17:23:48 by hutch--
I have not seen the numbers for the Windows/DirectX supercomputers yet. Where you can just go and buy off the shelf a SGI box with the specs I posted in without having to build a much bigger one, is it because there number are NOT there ?


You must have missed the part where I said that the Onyx4 graphics specs divided by the number of graphics processors is about equal to ATi R3x0 specs.
Or, for dummies: A cluster of 32 PCs with R3x0 cards will have the same raw numbers as the largest Onyx4.
Or, even better... There are some special cards with 4 R3x0 cores on one board... So then you'd only need 8 PCs with one card each.
Shall we compare prices now?
Oh, and I think you never mentioned Windows/DirectX in your original supercomputer statement. I believe it was only about x86... Re-read the thread perhaps?
And I'm certain that it was not about comparing Windows/DirectX numbers against the largest and most expensive monster that SGI has to offer.

you don't run high end graphics on Windows/directX boxes.


Yes we know, you said that before, we've been trying to discuss that from the start. You said you had a list of technical limitations that explain this phenomenon? Well, my good man, let's hear them please :)

I would guess that the guys floating around here who actually know their way around big iron would just about die laughing at the idea that kiddies toys are up in the big league.


Well by all means, let them join the discussion and give you a hand in explaining why exactly you cannot build supercomputers with x86 (except that Pixar uses them for high-end graphics), and why DirectX cannot be implemented on other systems than Windows x86-based PCs (except that it's available on Itanium2 and G5).

This is the time to put the numbers where your mouth is instead of your foot.


Whose foot is in whose mouth anyway?
Last time I read the discussion, it was not about numbers... It was about technical limitations. When are you going to mention these? Never? Why not? There aren't any?
You're not going to fool us with your smoke-and-mirrors approach. We know perfectly well what you stated during the earlier parts of the 'discussion', and everyone can read back the earlier posts, so don't change the statements around in such a way that you think we can't disagree with them, because that still doesn't mean your original statements are right.
It just means that you are a sore loser who tries to put on a nice show to save his face.
Oh, and this might be a good time to remind you that you still have not answered any of my or f0dder's questions during the entire 'discussion'. That's not very gentleman-like, is it? Shall we discuss those "manners and humility" again?
Posted on 2003-12-03 17:40:45 by Bruce-li
Its still a case of put up or shut up, vapourware is not a substitute for the specification of existing hardware that on the market already.

When you can find a set of numbers for the specifications of you high end competitor running Windows/directX, feel free to post them and I will then make reference to the big stuff that is being used by military, NASA and the like.

:tongue:
http://www.asmcommunity.net/board/cryptmail.php?tauntspiders=in.your.face@nomail.for.you&id=2f46ed9f24413347f14439b64bdc03fd
Posted on 2003-12-03 18:09:43 by hutch--
Its still a case of put up or shut up, vapourware is not a substitute for the specification of existing hardware that on the market already.


Hardware on the market is not a substitute for proof that it is technically impossible to use slightly different hardware and software to reach similar results.
So I agree, it's a case of put up or shut up. List me those technical limitations of x86, DirectX, and whatever other things you have been claiming.

When you can find a set of numbers for the specifications of you high end competitor running Windows/directX, feel free to post them and I will then make reference to the big stuff that is being used by military, NASA and the like.


I really don't like the way you are trying to push me into the position you want so you can think you can win the discussion. Read back the thread, and I never said that there were Windows/DirectX machines that could beat the top of the line in any specific field. Why would I anyway? It's not relevant to the actual point of the discussion anyway. Again, re-read the thread, or at least read the statements I've put in my signature. THAT is what we are discussion, not the things that you would LIKE to discuss, because you cannot discuss the things that yoy SHOULD be discussing. I'll wager a bet on why this is... You cannot back up your statements. Which is no surprise since some of them were rather bizarre...
Speaking of bizzare statements, and the military... How about that 70s flightsimulator then?
The 90s Japanese arcade system was quite funny aswell though.
Anyway, if you think it will prove your point, why wait for us to post any numbers (which we won't, since it's not what we were discussing), and why don't you make reference to that big stuff right away? I'm especially interested in that arcade machine.
Posted on 2003-12-03 18:17:35 by Bruce-li
Well,

For a man who waded into this debate with smartarse wisecracks about who did not know what they were talking about, I now see the capitulation you have been trying to hide from your earlier posts onwards.

I posted the specifications of immediately commercially available high end graphics box so that you could show us all just how superior the Windows/directX package you have in mind actually is.

This is a debate you started with your smartarse wisecracks but at the bottom line, you will not put up the data to prove your assertions about the superiority of Windows/directX performance.

Now when it comes to really high end performance in the image manipulation area, the specification I posted are not into the top end either. This is why big iron companies like SGI build special hardware for the really big users like military, NASA and the like.

Now I suggest its about time you gave up on this nonsense that Windows/directX is a competitor to really high end hardware and software as the specifications are there for all to see.

What I have asserted from the beginning is that Windows/directX/x86 is not in the big league where SGI/OpenGL is. I have made the point that while directX is a competitor to OpenGL on a PC scaled architecture, it isn't a competitor in the big league.

Now I have not heard anyone arguing that x86 is not cheap and useful, either running directX or OpenGL but you have continued to argue more than that, you have argued that it can compete with far larger scale combinations.

Feel free to post the specifications of the Windows/directX/pick_processor_that_will_run_it that actually can outperform big iron.

I am not going to hold my breath waiting. :tongue:
http://www.asmcommunity.net/board/cryptmail.php?tauntspiders=in.your.face@nomail.for.you&id=2f46ed9f24413347f14439b64bdc03fd
Posted on 2003-12-03 18:58:28 by hutch--
For a man who waded into this debate with smartarse wisecracks about who did not know what they were talking about, I now see the capitulation you have been trying to hide from your earlier posts onwards.


Nice try, but I'm not giving up anything, the moves you are trying to make now, seem to indicate however that you are running out of options.

This is a debate you started with your smartarse wisecracks but at the bottom line, you will not put up the data to prove your assertions about the superiority of Windows/directX performance.


Nonono, not so quick my friend... and I use the term loosely... I never claimed any superior performance... Let's see your claims, and I quote:

X86 architecture is simply too old to hold anything like the high end of graphics work and there have been and are better hardware that is not strangled by the architecture design inherant in X86 hardware.


What I suggested is that using X86 hardware to design a multiport image manipulation system is at best naive, something like trying to win a formula 1 race with a T model ford. DX runs good on the T model ford of computing but to win a formula 1 race, you simply need later technology.


Care to elaborate on this?

This is why big iron companies like SGI build special hardware for the really big users like military, NASA and the like.


Yes, but you would not consider Pixar a 'really big user' for example? And why not?

What I have asserted from the beginning is that Windows/directX/x86 is not in the big league where SGI/OpenGL is. I have made the point that while directX is a competitor to OpenGL on a PC scaled architecture, it isn't a competitor in the big league.


Funny, the quotes above say something quite different. More like "it couldn't be a competitor in the big league", and you even went as far as implying that there were solid technical reasons why.

Now I have not heard anyone arguing that x86 is not cheap and useful, either running directX or OpenGL but you have continued to argue more than that, you have argued that it can compete with far larger scale combinations.


And, you don't think that Google, Pixar, or the truckload of x86 systems high up in the top500 list of supercomputers qualify?

I am not going to hold my breath waiting


I should not hold my breath waiting for you to argue in a technical fashion either.
As far as I can see, all you have seen is some SGI advert that their system is the fastest, and have just accepted this as the truth, without any understanding of the used technology in their systems, or in supercomputers in general, or with OpenGL/DirectX for that matter...
And then you just made some ignorant statements, which you are now trying hard to deny, and failing miserably.
Posted on 2003-12-03 19:16:26 by Bruce-li
Both DirectX and OpenGL are ok...
They are only usefull for realtime stuff and "previews".... realtime stuff... they can't cast a ray at anything nor do global illumination for you.
I'm an OpenGL guy myself and i don't really mind the portability since i only write windows applications. But I know DirectX is probably a better choice at some specific stuff (I've seen DirectX demos do things to the framebuffer, a la "back to writing to 0xa000 days" way faster than I was ever able to do with OpenGL). On the other hand I don't like to instanciate a dozen COM objects to get a single triangle on the screen... or let's say it the other way... OpenGL has a less pronunciated learning curve ;]

I don't think one is better than the other on windows, and it isn't fair to get way over the subject and say than an "SGI" running a GL application has killer perfomance, because instead of saying "DirectX only runs on windows" we could say... "on SGI GL is the ONLY choice".

my 2 pesos.
Posted on 2003-12-03 21:14:37 by panoramix
Lucky I don't suffer your analysis huh ?

You wanted proof, the specs are proof. I asked you to post your own and you can't. I have repeatedly said there has been, there is and most prebably will be better hardware and software than Windows and directX for high end image work and the big iron proves it. Further the programmer who wants to use OpenGL does so with the knowledge that it is an industrial standard that is used with the big iron, with or without your approval.

Its superiority over time is matched by its scalability from PCs to big iron.

I am about as interested in the flavour of the day graphics cards as I am in the current IDE controller, state of the art com ports or any other gadgetry.

You can of course have any conversation you want but when you make your original post with a smartarse comment about who does not know what they are talking about, you have ended up with this one and this is the one you will keep having until you learn to keep your foot in your mouth.

Maybe it feels good to share plaititudes with a group of like minded people who enjoy pissing in each others pocket about how profound and leading edge their Window/directX stuff is but there is a real world out there that you have missed.

Vapourware is no substitute for knowing what you are talking about and with your comments on big iron, its clear you don't. Platitudes about what you remember at a university about SGI boxes don't match the specifications and you have not been able to find any that help you.

Your argument died with the lack of Windows/directX boxes in your own supercomputer list and the subsequent assertions have left you in no better position to support you view.

Like it or lump it, corporations with high end demand for graphics don't listen to trolls pissing around with the kiddies end of the market, they buy high end hardware on the basis of their demand and they have the expertise to determine that demand.

Same old question, why do the Lockheed use SGI for their flight simulators, why do NASA use SGI for graphics work, why does the Jet Propulsion Lab use SGI for image work. Are you so naive that you think they don't know the economy of scale using cheapies ?

How many x86 chips do you need to strap together to change their minds ? :tongue: Maybe you can work out how to strap together the right number of Taiwanese terrors to take the place of graphics big iron.

Weep as you calculate how many x86 chips you need in the face of the specs of the real thing. Just to irritate you a little further, big iron is extendable as well and it has the grunt to start with.

Now as usual, have a tantrum or two, stonewall a little longer and cuddle up to your comfortable theories of how the real world should be seeing things your way and DO have PHUN. :grin:
http://www.asmcommunity.net/board/cryptmail.php?tauntspiders=in.your.face@nomail.for.you&id=2f46ed9f24413347f14439b64bdc03fd
Posted on 2003-12-04 00:57:40 by hutch--
Let's see... has hutch misunderstood all the previous posts, or is he just trying to bend stuff to his own advantage?

Which API is techically better doesn't have much to do with what hardware it's currently implemented on. DirectX is the better API, there shouldn't really be any doubt about that (unless, of course, you haven't looked at the APIs, of course). Sure, SGI has some pretty decent hardware that only runs OpenGL - doesn't really mean OpenGL is the better API though.


Your argument died with the lack of Windows/directX boxes in your own supercomputer list and the subsequent assertions have left you in no better position to support you view.

Since when did anybody claim that supercomputer had anything to do with windows/directx? Apart, of course, from yourself?


Same old question, why do the Lockheed use SGI for their flight simulators, why do NASA use SGI for graphics work, why does the Jet Propulsion Lab use SGI for image work. Are you so naive that you think they don't know the economy of scale using cheapies ?

Obviously because the high-end SGI boxes have enough fillrate to make up for poor coding ^_^ (or, trolling aside, because the SGI boxes have the fillrate necessary - the right tool for the job).

Obviously you don't use OpenGL nor DirectX for the really heavy movie rendering (as noted in a lot of previous posts, both APIs are meant for hardware acceleration and thus lack support for things needed in movie rendering, as they don't (currently?) lend themselves well to hardware). Thus, SGIs monster graphics subsystem becomes irrelevant, and any supercomputer cluster will do - obviously pixar decided that x86 has an acceptable performance/price ratio.

Perhaps I should just give up, since hutch seems to continue mixing and matching unrelated statements to his liking.
Posted on 2003-12-04 04:45:42 by f0dder
You wanted proof, the specs are proof. I asked you to post your own and you can't. I have repeatedly said there has been, there is and most prebably will be better hardware and software than Windows and directX for high end image work and the big iron proves it.


No, I asked proof for your OTHER statements, you know, about the technical limitations of x86 and DirectX.
You must be confused.
This proves nothing, other than that you cannot back up your original statements.
Perhaps it's time to admit that those original statements were perhaps a bit exaggerated?

Its superiority over time is matched by its scalability from PCs to big iron.


You seem to confuse technical superiority with what I would call 'geographics', the type of systems that it's actually implemented on.
Strictly speaking, there is no relation between the 'superiority' or 'scalability' of an API and the systems it is implemented on.
If OpenGL is superior to DirectX, and if there is a technical reason why it is superior or more scalable, then you should be able to name it, right?
What you're saying is no more than "DirectX is implemented on XBox, so it is superior to OpenGL for game consoles", or even "OpenGL is not implemented on XBox, so you can't use it for game consoles".
Is that true? Well it could be... then again, it may not be. And if it indeed was true, there'd be some technical differences between the two to indicate that, don't you think?

Vapourware is no substitute for knowing what you are talking


Indeed, I prefer technical facts. You are in fact using 'vapourware' to 'prove' your point. You have gotten no further than "There is no DirectX on SGI systems". While this is true, I don't see any technical value in this statement. It is just a simple fact, but says nothing technical about either DirectX or SGI. So if you were to assert that DirectX couldn't run on SGI systems (as you did, re-read the thread), you'd have to come up with something more technical than that to prove it.

Your argument died with the lack of Windows/directX boxes in your own supercomputer list and the subsequent assertions have left you in no better position to support you view.


Firstly, re-read the thread and see it was not about Windows/DirectX boxes... Secondly, the Windows 2000 HPC website mentions the event of a Windows-powered (with DirectX) supercomputer entering the top500 list. While I can't be bothered to figure out exactly which one it is, and if it is the only one on the list or not, it is ample proof that it is there. So you are again making the same foolish statement, which nobody is going to buy.

Like it or lump it, corporations with high end demand for graphics don't listen to trolls pissing around with the kiddies end of the market, they buy high end hardware on the basis of their demand and they have the expertise to determine that demand.


I really think it is time you start commenting on Pixar now.

To make this short:
You have constantly avoided to talk about actual technical issues with x86, other CPUs used in supercomputers, DirectX, OpenGL, Pixar, Google, Windows 2000 supercomputers, and anything else that relates directly to the statements as you have originally made them.
Are you really so stupid that you don't know what we were actually discussing?
Or do you think the readers of this forum are so stupid that they think you are actually right?
Either way it's pathetic.
Why don't you do the smart thing, and either adjust your original statements, or discuss them at a technical level and explain everyone why you can't build a supercomputer with x86, and why you couldn't use DirectX on it.
Oh, and I am still interested in hearing how you would render "Finding Nemo" in OpenGL. It would save me a lot of time :)
Can you for once, just once, actually stick to the point?
Read back f0dder's questions, they might help you to explain your view to us. That's all we want, we just want you to explain your view. We never asked you to preach the glory of extremely expensive SGI systems.
You can explain your own view, right? I mean, it is based on something technical, right? We are scientific-minded people, right? Who want to see technical reasons why something is, or isn't true, right? Who don't just believe something because someone says it, a 100 times, right?

PS: If you like 'big iron' so much, allow me to point out this HP visualization center: http://www.hp.com/workstations/risc/visualization/sv7.pdf
It has a higher fillrate per graphics processor than the SGI Onyx4 (3.2 gpix/s vs 2.4 gpix/s, oh and they aren't ashamed to mention that they use 'game' cards for it either, they use NVIDIA Quadro FX2000, you know, the cousin of that puny GeForce FX that these silly gamers use in their silly PCs), and on top of that, it doesn't have an upper limit of 32 processors. You can add an 'unlimited' amount of graphics nodes to the machine, easily beating anything that SGI has to offer.
So if you insist on your false arguments, this is a better example than the tiny, weak SGI system you've been using.
Posted on 2003-12-04 05:02:08 by Bruce-li
Damn,

I thought there may have been some rest from the loop logic coming out of our friend and his associate. When neither want to back up their assertions in the face of the specifications that I posted, ther only deduction is that the assertions were nonsense to start with.

Remember who introduced the list of supercomputers to try and prove a point that if you strap enough x86 chips together you make a giant killer. Then there is the problem that nothing in the high ends market runs Windows and directX.

Apart from the platitudes from our friend that SGI hardware was past it as he recollected from university, as a matter of fact, they perform in a class beyond Windows/directX compatible computers.

We have continued to hear from our friends that directX has a better set of APIs but for all the assertions, there are a lot of people who prefer OpenGL, especially as it scales up to far more powerful hardware.

Waffling between high end assertions and "gee its good for it price" is a failure as the shift makes the original assertions the nonsense they are.

When you have real big time stuff like military flight simulators, space research facilities in NASA and the JPL being specified with high end graphics companies like SGI, should we believe a couple of directX junkies that these people don't know what they are talking about ?

The simple answer is that they are right and you are wrong. The other factor of course is that they would be far better informed about strapping together cheapies and what you can get out of it than a couple of directX junkies.

This has never been a conversation about the innards and features of taiwanese terrors and the like, it has been about the assertions of superiority of directX over OpenGL where the evidence is that OpenGL scales from PCs to supercomputers that are sold and used for high end graphics.

Like it or lump it, OpenGL is used across a far wider range of hardware platforms than directX and it is for an obvious reason, generally high end computers don't run Windows. Yes we have heard that it has been ported to a MAC but that has to do with compatibility of games, not any intrinsic advantage it has.

Now in the absence of any specifications being posted to back up this claptrap, how many x86 chips do you have to strap together to compete with big iron. Weep as you calculate the numbers as you read the high end specifcations and stop deluding yourself.
http://www.asmcommunity.net/board/cryptmail.php?tauntspiders=in.your.face@nomail.for.you&id=2f46ed9f24413347f14439b64bdc03fd

PS: Do have PHUN. :grin:
Posted on 2003-12-04 08:32:58 by hutch--
dude... a little off topic, but have you actually seen a Lockhead or Nasa flight sim? they totally suck b*lls... ;]
Posted on 2003-12-04 09:18:57 by panoramix
I used to work for an Australian company that is a major supplier to the automotive trade, big names like Ford, Holden, Toyota etc... This large company was divided into roughly four sections - the paperchasers, the design engineers, the production engineers, and damage control (systems engineers, maintenance engineers etc).
My job was to analyse the flow of various plastics at various temperatures under various pressures within large injection moulding tools, at the time, mostly bumper bars. I managed that on a pc. Still, it never ceased to amaze me what the design guys were up to, they would pull up a model of the next Commodore or whatever, complete car down to the last detail in micron resolution, and throw it around in 3D and not even feel it. I couldn't load a model larger than a single car part at once on a single pc. Yet these guys could throw the whole car around - at the time I was quite impressed. There were probably 30 of these awsome machines in the design department, and proudly stamped on each organic-looking case was SGI.
(I bet they ran Quake pretty good too :tongue: )

In reply to the post above mine, you think theirs are bad?
check this article about Australian Defence simulator screwup
Posted on 2003-12-04 09:27:30 by Homer
Remember who introduced the list of supercomputers to try and prove a point that if you strap enough x86 chips together you make a giant killer.


Quote the exact sentence where that point was made, and by whom.

Then there is the problem that nothing in the high ends market runs Windows and directX.


This is a fact, are there technical reasons which cause this fact, and if so, can you name them?

Apart from the platitudes from our friend that SGI hardware was past it as he recollected from university, as a matter of fact, they perform in a class beyond Windows/directX compatible computers.


You seem to lack the ability to distinguish between PCs, workstations, mainframes and supercomputers.

there are a lot of people who prefer OpenGL, especially as it scales up to far more powerful hardware.


This is a fact, are there technical reasons which cause this fact, and if so, can you name them?

The simple answer is that they are right and you are wrong.


Is Pixar wrong?
You have never commented on the Pixar-case yet. This is extremely weak. Is it because Pixar does not fit into the view you try defend?
Besides, I cannot be wrong, since I merely asked you to clarify your statements. If I am wrong, then you are wrong.

Now in the absence of any specifications being posted to back up this claptrap, how many x86 chips do you have to strap together to compete with big iron. Weep as you calculate the numbers as you read the high end specifcations and stop deluding yourself.


Weep as you continously fail to give any technical arguments whatsoever, and make a complete ignorant fool out of yourself in front of everyone.
You cannot win the argument you started yourself.
Perhaps you should not speak of things you have no technical knowledge about, in future.
There are people that do, and they will know when you are wrong, and point it out to you.
And your stupid weasel-act does not impress anyone.
Argue the points, or give up.
Posted on 2003-12-04 10:00:56 by Bruce-li
I moved it to the Crusades as it looks more and more like it
Posted on 2003-12-04 15:41:06 by BogdanOntanu
Bogdan,

There is great wisdom in you approach.

Now for our friends, I have been committed to two things from my original assertion.

1. OpenGL is available on a wider range of hardware and far more powerful hardware than x86.

2. x86 hardware inherits architectural limitations from the age of its design.

The first is simply a matter of fact and so is the second. Our friends wanted evidence so I produced the specs of a current off the shelf SGI box and like it or not, x86 hardware is not in this performance class.

Our friend introduced clustering to try and cover up the obvious yet it is still not in the class of an off the shelf SGI box.

Bottom line is I stand by my original assertions that x86 is old architecture with well known limitations and OpenGL is an older, more mature software system that scales across a far wider range of hardware than the more recent directX which owes its origin to a problem to the HAL in Windows.

I am not even going to address the latest set of Taiwanese terrors because it does not matter in this argument, big iron delivers the performance where it claims and the rest drive video games for the kids.

Our friends are free to deliver the SGI killer set of specs and existing examples of x86 hardware running directX that outperform the big iron but I would not hold my breath on the outcome. :tongue:

As usual, make sure you have PHUN doing it.

Muhahahaha.
http://www.asmcommunity.net/board/cryptmail.php?tauntspiders=in.your.face@nomail.for.you&id=2f46ed9f24413347f14439b64bdc03fd
Posted on 2003-12-04 19:40:00 by hutch--
Mike: "Hey Bob, which car should I buy? The red or the blue one?"
Bob: "The red one, it's very fast. Red cars are used in Formula 1 racing. Blue cars can't be used for racing."
Bob: "Blue Formula 1 cars don't exist."
Mike: "Okay, thanks Bob... But, are you sure? I think I saw a blue racecar once."
Bob: "Blue Formula 1 cars don't exist."
...
Mike: "Hey Bob, it seems I was right. I did see a blue racecar. It's a Formula 1 car actually. Here is a picture of it."
Bob: "Blue Formula 1 cars don't exist."
...
Mike: "Wow, did you see the Formula 1 Grand Prix this weekend? It was quite exciting. The blue car finished 4th, and the red car finished 41st. What do you say about that?"
Bob: "Blue Formula 1 cars don't exist."

And banana's are yellow, not white like you said. So you know I'm right and you're wrong.
Posted on 2003-12-04 19:54:25 by Bruce-li