R480/430 roadmappery

[kemosabe]

http://www.hkepc.com/bbs/viewthread.php?tid=229551

 

[hovz]

ahahaha, a 500 point increase in 3dmark03, this is definitly worth releasing

 

[Ostsol]

So was ATI's estimate 3DMark03 score 13000?

 

[jvd]

ahahaha, a 500 point increase in 3dmark03, this is definitly worth releasing

eh the big deal is 3dmark 2005 and what increase that will bring

 

[Killer-Kris]

ahahaha, a 500 point increase in 3dmark03, this is definitly worth releasing

Well from the hints Dave has been dropping, I think that a performance upgrade was never really in the plans. Just more or less making the chip cheaper for them to produce, therefore make them more money.

 

[hovz]

ahahaha, a 500 point increase in 3dmark03, this is definitly worth releasing

Well from the hints Dave has been dropping, I think that a performance upgrade was never really in the plans. Just more or less making the chip cheaper for them to produce, therefore make them more money.

regardless, its just funny how they throw out a 3dmark estimate and its only 500 points higher. why even bother?

 

[hovz]

ahahaha, a 500 point increase in 3dmark03, this is definitly worth releasing

eh the big deal is 3dmark 2005 and what increase that will bring

yes because theres even a single developer programming a game that will push enough polygons to actually require any decent ammount of vertex power

 

[jvd]

ahahaha, a 500 point increase in 3dmark03, this is definitly worth releasing

eh the big deal is 3dmark 2005 and what increase that will bring

yes because theres even a single developer programming a game that will push enough polygons to actually require any decent ammount of vertex power

unreal engine 3

I'm sure the end of 2005 /2006 will bring many games that will push hardware like 2005 does

 

[ChrisRay]

Err doesnt Unreal 3 engine only require like 400,000 polygons? No where near what 3dmark05 uses.

 

[chavvdarrr]

but why the 0.11 part is running at lower clocks?! (obviuosly)
Providing 2 products at same price (R480Pro and R430XT), qith diffrent number of pipes (12 vs 16), will only add more mess if that is possible ?!

 

[Tim]

Err doesnt Unreal 3 engine only require like 400,000 polygons? No where near what 3dmark05 uses.

But 3dmark05 is not vertex shader limited anyway - so what is the problem?

 

[ChrisRay]

Err doesnt Unreal 3 engine only require like 400,000 polygons? No where near what 3dmark05 uses.

But 3dmark05 is not vertex shader limited anyway - so what is the problem?

I think you missed the point or the following of the conversation, Hovz said there was no game that would push vertex shaders like 3dmark05,

As JVD replied using Unreal 3.0 as point that it will push hardware like 3dmark05, And it wont even be anywhere near the vertex load of 3dmark05.

 

[Tim]

As JVD replied using Unreal 3.0 as point that it will push hardware like 3dmark05, And it wont even be anywhere near the vertex load of 3dmark05.

When 3Dmark performance is not limited by vertex shaders it is no problem that poly higher than in UE3 - it can very well still be indicative of the performance.

 

[ChrisRay]

As JVD replied using Unreal 3.0 as point that it will push hardware like 3dmark05, And it wont even be anywhere near the vertex load of 3dmark05.

When 3Dmark performance is not limited by vertex shaders it is no problem that poly higher than in UE3 - it can very well still be indicative of the performance.

It could very well. But thats still not the point I was making. And pretty much irrelevent since I was merely pointing out Unreal 3.0 will not push vertex shaders like 3dmark05.

And I disagree with your assertion on a whole. But this isnt the place for that discussion.

 

[phenix]

Dual Slot cooling for XT? :? Hope tis not true.

 

[wireframe]

Dual Slot cooling for XT? :? Hope tis not true.

What is the big deal with dual slot cooling for video cards? It seems inevitable for high performance graphics. If anything, mainboard designs and specifications should begin catering more for this, especially for the performance segment.

I understand that this may not be desirable for the SFF segment, but if you want performance you have to pay the price. "Dual slot" cooling should be catered for more thoroughly and efficiently so that bigger HSFs are not thrown at the problem, but more efficient and effective ones.

Look at the current design to see the problems. People are expecting the world from their GPUs, often having more transistors than their CPUs, not to mention boards with onboard memory capacities of 256MB and ready to grow. That is like an entire system board not many years ago. I think it deserves a little bit more than a fan sitting on a thin piece of metal to dissipate heat and only a few centimeters, at best, to the next slot. It's ridiculous.

I say go "triple slot," but do it right. Give it adequate cooling through a ducted system so we don't have to pretend a fan sitting in a sauna blowing air on the most complex component in the system is going to keep it cool and happy. Unfortunately, I don't think even the proposed/upcoming BTX spec caters for this. Just because the video card used to sit in a PCI slot, then tweaked to AGP, and now PCIe does not mean it is like the other cards in the system. A single slot, close proximity NIC and audio card I can understand, but pretending that a video card should be happy with the same is just asking for trouble.

The way it is now the board designers, and hence the GPU designers, are fighting a pointless and losing battle against the closest slot. They cliam victories over each other when they manage to squeeze into the spec and current designs. Instead they should be fighting to redesign the PC mainboard to get adequate space and features. Give it more clearance. People can't remove their video cards easily once they are seated because the release is blocked by southbridge HSFs for goodness sakes. It is clear the current design is not working.

EDIT: Of course I realize that we are working with current specs and, therefore, the cards need to cater for that. I am not saying "go triple" with the mainboards we have now. I am saying change the spec for the future because people are moaning about dual slot cooling, something we already have, works, but is clumsy and reveals the limitations of the current design. We know that "dual slot" works as there are several boards like that already. Wanting top graphics in SFF is going to be a losing battle until some changes are made.

Speaking of changes, how about a standard dual PCIex16 interface allowing video adapters to take on more of a box shape that can be used to "wind tunnel". This gives more power...and...I am dreaming... heh.

 

[phenix]

Dual Slot cooling for XT? :? Hope tis not true.

What is the big deal with dual slot cooling for video cards?
.
.
.
I understand that this may not be desirable for the SFF segment, but if you want performance you have to pay the price. "

They (SFFs) are becoming more and more widespread. I dont think high end performance should be exclusively for the big cases.

Actually that's something 'I don't want' rather than 'I don't think', which is not that important is it?

 

[wireframe]

They (SFFs) are becoming more and more widespread. I dont think high end performance should be exclusively for the big cases.

Actually that's something 'I don't want' rather than 'I don't think', which is not that important is it?

Sure SFFs are becoming more widespread. They have not really been available in quantity for very long compared to the larger formats. However, you have to look at the situation and be realistic. If video card manufacturers are fighting against constraints on the full ATX form factor, they are not getting a friendly pat on the back from SFFs.

I think it's obvious something is 'wrong' when a new format, PCIe, arrives and it immediately requires an external connector for power. Granted, this is not a problem in and of itself, it actually makes a lot of sense to always use an external power source from the transmission bus socket, but it's clear that designers are running with a wall touching the tips of their noses.

Look at it without wanting (but, please keep thinking ). Video card demand is such that "give me something equivalent to the entire system but fit it into a PCI slot." Not only is this requirement tight, but this subsystem must exist within the more spacious one influencing it, case temperatures being affected by all components and then the video card is asked to use this already warm air to cool itself.

It really is crazy what is demanded of the video card industry. They should be given a big hand for even contemplating satisfying the demand. Geforce 6800 Ultra: 50 million (~33%) more transistors than a Pentium 4 Prescott/1MB L2 cache in the GPU and 256MB RAM running at 1100MHz. This and more on one board. A board that must exist in an environment that is far from optimal. Now look at the cooler you are accepting for any CPU, or the Prescott specifically. Quite large.

I'm sure greater performance will be available to SFFs without the SFFs being redesigned, but ultimately the system on the whole needs an overhaul. All I am saying is that I would understand if high performance graphics will not limit itself to sitting pretty in a single slot, occupying no space, consuming no power. A technological leap is required.

I am a bit long and I may have come across as overly defensive about the "dual slot" comment. It is not intended as anything but a comment on this "cheering for single-slot". I thought this was a good time to bring out some reality because I find it ludicrous that ATI fans (this pun is defeinitely not inteded) seem to find fault in the GF 6800 Ultra because it uses a cooler that blocks out the adjacent PCI slot. Now high end ATI users look likely to face the same and will have to shut up about it. It is a reality.

 

[skazz]

Given SLI implementations and dual PCI-express slots from NVidia, I think there are going to be problems getting a triple slot cooling accepted

 

[KimB]

unreal engine 3

I'm sure the end of 2005 /2006 will bring many games that will push hardware like 2005 does

No game by the end of 2005/2006 will push games like 3DMark2005 does. It's vastly, vastly too vertex-limited. Futuremark once again has gotten things off. And they're bound to always do so since the driving forces behind designing a benchmark are entirely different from those behind games. This is why 3DMark always has been and always will be pretty much useless.