NVIDIA to Acquire ULi Electronics, a Leading Developer of Core Logic Technology

[Gstanfor]

Originally posted by: BFG10K

The point is simple. Part of the SM2/3 spec is vertex looping feature. On nVidia GPU, that looping is entirely hardwared accelerated, on ATi GPU's it is not hardware accelerated since the GPu must rely on the CPU to unroll the loop for it.

Looping (AKA dynamic branching) is an SM 3.0 feature which the R4xx series didn't support. Bashing ATi in that instance for using the CPU is as stupid as bashing nVidia's TNT2 for using the CPU for T&L calculations.

nVidia's dual core driver that you are so keen to bash (don't forget ATi also has a dual core driver in the works...)

Nobody's bashing anything, I'm just pointing out your double standards.

it does not perform parts of the vertex processing that should be hardware accelerated on the GPU (but are missing in the case of ATi's GPU's).

You mean like how the TNT2 didn't do T&L and hence nVidia got the CPU to do it for them? :roll:

As per usual, you are dead wrong.

Vertex looping is a SM2.0 feature. nV3x and upwards supported it R3xx through R4xx did not (I'm not certain if R5xx does or not - I'm betting it doesn't however).

And, yes, you are bashing, and you, not me, were the one to bring up vertex looping. You might like to get a clue what you are crapping on about the next time you feel the urge to open your mouth and shove your entire leg down it in public...

 

[Gstanfor]

Originally posted by: BFG10K

The point is simple. Part of the SM2/3 spec is vertex looping feature. On nVidia GPU, that looping is entirely hardwared accelerated, on ATi GPU's it is not hardware accelerated since the GPu must rely on the CPU to unroll the loop for it.

Looping (AKA dynamic branching) is an SM 3.0 feature which the R4xx series didn't support. Bashing ATi in that instance for using the CPU is as stupid as bashing nVidia's TNT2 for using the CPU for T&L calculations.

First of all, looping and dynamic branching are NOT the same thing. Vertex looping was introduced with SM2.0, nVidia supported it on the GPU with nV3x. ATi did not support it. Dynamic branching is Sm3.0 based and affects the pixel shaders, not the vertex shaders (every shader model subdivides into vertex shader operations and pixel shader operations FYI.

nVidia's dual core driver that you are so keen to bash (don't forget ATi also has a dual core driver in the works...)

Nobody's bashing anything, I'm just pointing out your double standards.

You not I brought up vertex looping and nVidia's dual-core drivers in this thread, and you were wrong both times. I don't have a double standard on the issue.

it does not perform parts of the vertex processing that should be hardware accelerated on the GPU (but are missing in the case of ATi's GPU's).

You mean like how the TNT2 didn't do T&L and hence nVidia got the CPU to do it for them? :roll:

When the TNT sereies was state of the art, no consumer oriented graphics chip performed T&L or vertex processing on chip. It was triangle setup only back in those days outside of professionally oriented graphics products that were not gaming oriented or priced. Your comparison is ridiculous!

 

[Matt2]

Originally posted by: munky

Originally posted by: Rollo

Originally posted by: 5150Joker
LOL I can't wait to see the spin nVidia fans put on this one - especially the ones that were saying ULi is trash. Now they'll do a 180 and praise it since nVidia is buying them out.

Teh Devil!

This nVidia fan still isn't totally sold on ULI(ALI), but is LOL at how smart of a business move this is.

As noted by me in other threads, it's obvious the ULI Southbridge is a big improvement over ATIs crappy one, so at very least nVidia gets to choose whether or not any Crossfire motherboards will have ULI Southbridges.

Merry Christmas ATI, here's your coal.

Or better yet, Ati can make crossfire compatible with any dual pcie board, and let Nv do all the hard work figuring out how to make their proprietary crap even more proprietary.

Had to stop reading here. I hope that was a joke munky.

I think you have your "proprietary crap" mixed up. Crossfire is the biggest joke since the GeForceFX

 

[BFG10K]

First of all, looping and dynamic branching are NOT the same thing.

Dynamic branching is used to control loops given static looping is quite limited. In that respect they can be classed as the same thing.

Vertex looping was introduced with SM2.0, nVidia supported it on the GPU with nV3x.

Except the NV3x exceeded the VS 2.0 spec by supporting VS 2.0a. Given VS 2.0 only called for static looping, I'd like to see your evidence that the R4xx didn't support that feature.

When the TNT sereies was state of the art, no consumer oriented graphics chip performed T&L or vertex processing on chip.

The NV25 didn't support SM 1.4, unlike the consumer oriented graphics chip 8500. When are you going to start slamming the NV25 for offloading this work to the CPU (or worse, not doing it at all)?

 

[Gstanfor]

Originally posted by: BFG10K

First of all, looping and dynamic branching are NOT the same thing.

Dynamic branching is used to control loops given static looping is quite limited. In that respect they can be classed as the same thing.

:disgust: they are not the same thing at all. They are not even close. They don't even affect the same portion of the 3D graphics pipeline. And *YOU* specifically brought vertex processing into this thread. What does a pixel shading function have to do with vertex processing??? or are you just like Creig - dodging and weaving?

Vertex looping was introduced with SM2.0, nVidia supported it on the GPU with nV3x.

Except the NV3x exceeded the VS 2.0 spec by supporting VS 2.0a. Given VS 2.0 only called for static looping, I'd like to see your evidence that the R4xx didn't support that feature.

Actually I never said anything at all about vertex processing until you crawled out from under your bridge and I certainly never claimed ATi didn't support static looping. That, my scaly skinned cretin is all you.

When the TNT sereies was state of the art, no consumer oriented graphics chip performed T&L or vertex processing on chip.

The NV25 didn't support SM 1.4, unlike the consumer oriented graphics chip 8500. When are you going to start slamming the NV25 for offloading this work to the CPU (or worse, not doing it at all)?

WTF has PS V1.4 (there is no SM1.4 ...) got to do with anything in this thread? Who uses it other than ATi and futuremark (you know good old Bribemark03 the "DX9 benchmark" that had only 2 DX9 shaders in it - everything else was PS V1.4 or lower...)

 

[nRollo]

Originally posted by: Gstanfor

Originally posted by: BFG10K

First of all, looping and dynamic branching are NOT the same thing.

Dynamic branching is used to control loops given static looping is quite limited. In that respect they can be classed as the same thing.

:disgust: they are not the same thing at all. They are not even close. They don't even affect the same portion of the 3D graphics pipeline. And *YOU* specifically brought vertex processing into this thread. What does a pixel shading function have to do with vertex processing??? or are you just like Creig - dodging and weaving?

Vertex looping was introduced with SM2.0, nVidia supported it on the GPU with nV3x.

Except the NV3x exceeded the VS 2.0 spec by supporting VS 2.0a. Given VS 2.0 only called for static looping, I'd like to see your evidence that the R4xx didn't support that feature.

Actually I never said anything at all about vertex processing until you crawled out from under your bridge and I certainly never claimed ATi didn't support static looping. That, my scaly skinned cretin is all you.

When the TNT sereies was state of the art, no consumer oriented graphics chip performed T&L or vertex processing on chip.

The NV25 didn't support SM 1.4, unlike the consumer oriented graphics chip 8500. When are you going to start slamming the NV25 for offloading this work to the CPU (or worse, not doing it at all)?

WTF has PS V1.4 (there is no SM1.4 ...) got to do with anything in this thread? Who uses it other than ATi and futuremark (you know good old Bribemark03 the "DX9 benchmark" that had only 2 DX9 shaders in it - everything else was PS V1.4 or lower...)

I hope you never owned a 5800, or ever stated they weren't as bad as people say Gstanfor. :beer:

 

[Gstanfor]

No, never the nV30 (5800), just nV35 (5900XT). I was, and still am quite pleased with that.

 

[nRollo]

Originally posted by: Gstanfor
No, never the nV30 (5800), just nV35 (5900XT). I was, and still am quite pleased with that.

:shocked:

Uh oh. NOW you've done it.

There's a few here who will now incorporate that you like an FX card into every single thread you post in, because the AMA long ago declared that enjoyment of a FX card is grounds for commitment to a mental institution, and negates anything else you have to say.

Didn't you see "Shady Days"? Didn't you listen to Gabe?!?!?!? WTH is wrong with you?!?! Jump on the bandwagon of wildly yelling "The FX sucked" and all will respect you again.

I'm sorry, that's the way it is.

:laugh:

 

[Gstanfor]

It's already common knowledge that I liked my GF-FX, and I could give a flying fv<ck what the fanatics think of me...

 

[nRollo]

Originally posted by: Gstanfor
It's already common knowledge that I liked my GF-FX, and I could give a flying fv<ck what the fanatics think of me...

Heh we think alike then. :beer:

 

[dug777]

Ewww, this is a public forum Gstanfor and Rollo, get a room

Seriously, get back on topic too

 

[Gstanfor]

Originally posted by: dug777
Ewww, this is a public forum Gstanfor and Rollo, get a room

Seriously, get back on topic too

What are you singling Rollo an myself out for? The thread *was* bang on topic until Creig, BFG10K & their band of merry fanatics came in and hijacked it...

 

[dug777]

Originally posted by: Gstanfor

Originally posted by: dug777
Ewww, this is a public forum Gstanfor and Rollo, get a room

Seriously, get back on topic too

What are you singling Rollo an myself out for? The thread *was* bang on topic until Creig, BFG10K & their band of merry fanatics came in and hijacked it...

i haven't read it since i posted in it a few days ago, revisted and there you two were at the bottom of teh last page

I was mainly having a gentle dig at the little FX-Series reunion tho

 

[Gstanfor]

Originally posted by: dug777

Originally posted by: Gstanfor

Originally posted by: dug777
Ewww, this is a public forum Gstanfor and Rollo, get a room

Seriously, get back on topic too

What are you singling Rollo an myself out for? The thread *was* bang on topic until Creig, BFG10K & their band of merry fanatics came in and hijacked it...

i haven't read it since i posted in it a few days ago, revisted and there you two were at the bottom of teh last page

I was mainly having a gentle dig at the little FX-Series reunion tho

Don't mock what you havn't experienced firsthand Dug777. That one of the main reasons I can't stand fanATics...

 

[dug777]

Originally posted by: Gstanfor

Originally posted by: dug777

Originally posted by: Gstanfor

Originally posted by: dug777
Ewww, this is a public forum Gstanfor and Rollo, get a room

Seriously, get back on topic too

What are you singling Rollo an myself out for? The thread *was* bang on topic until Creig, BFG10K & their band of merry fanatics came in and hijacked it...

i haven't read it since i posted in it a few days ago, revisted and there you two were at the bottom of teh last page

I was mainly having a gentle dig at the little FX-Series reunion tho

Don't mock what you havn't experienced firsthand Dug777. That one of the main reasons I can't stand fanATics...

Hey mate no need to be a jerk about it

FYI my best mate had a fx5200 until recently, and it was our main gaming rig for a long time while i had a gf4mx440...i'd say that counts as firsthand experience

 

[Gstanfor]

You can' t truly judge the GF-FX by the 5200. I have never reccomended that to anyone at all interested in gaming. In fact, IMO only the 5900 series were good for gaming out of the FX series.

I feel sorry for anyone who considers themselves a gamer and has to use entry level graphics cards from any maker - and that isn't a putdown aimed at you, nor were my previous remarks for that matter.

 

[dug777]

Ya rly, i got the shock of my life when i went from farcry on my mates fx5200 to farcry on my 9800 pro :Q

 

[Acanthus]

I dont know why anyones suprised by this, Nv has been known to do this in the past (3dfx). They are probably going to acquire all the IP and talent and use it for future projects, i dont think they are buying ULi just to screw with ATis southbridge, the aquisition doesnt even take place until 2007.

 

[Lord Banshee]

probally to battle intel more than Ati.. Intel has all the money when it comes to chipsets and IGP. So maybe Nvidia is trying to get as much resources as possible and go hard core against Intel...

 

[Gstanfor]

I think its to broaden the IP base as much as possible, while paving the way for a solid budget lineup. Screwing ATi out of alternative componentry is merely a pleasant side-effect (I';m quite sure ATi will be able to purchase any southbridge they want from nVidia in the future -- for a price, just like I'm sure nVidia would have sold them the PCI-e <--> AGP bridge chip if they had asked).

Personally I think ULi's current offerings are pretty good, and I intend purchasing the Asrock board with both AGP & PCI-e on it myself very soon. It means I can move to the AMD64 platform at my own pace without being forced to upgrade my graphics to PCI-e at the same time.

 

[BFG10K]

They don't even affect the same portion of the 3D graphics pipeline. And *YOU* specifically brought vertex processing into this thread.

Actually your brought it in with "vertex looping" and I'm still waiting for your evidence to back the claim.

Even if we follow the assumption that ATi is unrolling the static branching on the CPU the actual unrolled instructions will still be executing on the vertex shader which means it's a joined CPU + GPU operation, one which you praised nVidia's dual core drivers for doing in this thread. I don't recall you following similar logic when you slammed ATi in the past.

What does a pixel shading function have to do with vertex processing???

This is a joke right? You do know that vextex shaders implement both static and dynamic flow control just like pixels shaders do, right?

I'm not talking about the pixel shaders and never was, I was responding to your "vertex looping" comment. Since this comment is so vague and could mean just about anything I assumed you meant vertex flow control and since you were talking about the VS 2.0 spec I also assumed you were talking about static flow control.

If you want discuss the finer points of shader language I suggest you start using the correct terminology or at least clarify what it is that you mean.

Actually I never said anything at all about vertex processing until you crawled out from under your bridge and I certainly never claimed ATi didn't support static looping.

So just what the hell does "vertex looping" mean then? I've never seen it mentioned in the technical papers so I assumed you had resorted to using a simpleton description of vertex shader terminology.

WTF has PS V1.4 (there is no SM1.4 ...) got to do with anything in this thread?

You slam ATi for not supporting hardware features but don't do the same for nVidia. Each time I present a scenario which would require you to bash nVidia you dismiss it with useless rhetoric. This is typical fanboyism.

And I notice the AEG spokesman has arrived in this thread, no doubt trying to earn a Christmas bonus for trumpeting the virtues of the failed FX line.

 

[Keysplayr]

Originally posted by: BFG10K

And I notice the AEG spokesman has arrived in this thread, no doubt trying to earn a Christmas bonus for trumpeting the virtues of the failed FX line.

You can't possibly sink any lower. But I'm sure you'll try. :roll:

::: Pats 5900XT on the back :::

 

[nRollo]

Originally posted by: BFG10K
And I notice the AEG spokesman has arrived in this thread, no doubt trying to earn a Christmas bonus for trumpeting the virtues of the failed FX line.

My my, baseless inflammatory accusations from our friend BFG.

QFP and PMd to the mods. Must be trying to get the thread locked because you're losing your argument with GStanfor.

 

[BFG10K]

Must be trying to get the thread locked because you're losing your argument with GStanfor

I see, and what exactly are you trying to do in this thread Rollo?

Specifically what relevance does the FX line have with ULi Electronics?

 

[BFG10K]

My my, baseless inflammatory accusations from our friend BFG.

Notice how I never used a name but I find it cute you automatically assumed I was referring to you.