X1800xl

[Unknown Soldier]

I'm rather disappointed in the X1800XL. I would have expected it to at least beat the 7800GT.

So seeing as it doesn . .here's my question. Will X1800XL buys be able to do a BIOS update to get it to run X1800XT speeds?

The R420 Pro was slower than the GT 6800 when it was released but buyers(of the VIVO ver.) were able to do the BIOS upgrade to get it to R520XT PE speeds and got good value for money.

Can we expect much of the same for the X1800XL? It's pretty much a pointless card if it can't.

US

 

[wireframe]

Perhaps this should be in the XL thread or one of the numerous X1000 threads, but here goes.

First of all, there is nothing to unlock, so your only hope is overclocking.

Second, ATi is using an older revision of the chip, one with a soft ground problem that did not provide sufficient yields for the XT, for early XL cards. This will most likely mean that the XLs are very poor overclockiers. Once this old stock is sold out (quantity unknown), the situation will likely improve.

I am not sure about the memory used and if that has some overclocking potential. I believe the PCB is the same as the XT, but it may be populated differently enough not to be able to power an XT level GPU. Notably, the XL seems to consume half the power of the XT at full sprint, so take that for what its worth when it comes to the necessary power regulation of the board.

EDIT: Sorry, I made the mistake of comparing a 512MB XT to a 256MB XL in power consumption. Both cards use 8 memory chips, but the 512 ones are twice as dense. Not sure what the difference in power consumptions between these memory type is. Maybe we could guess a half/half for the GPU and memory for the increased power consumption so that the XT GPU requires roughly 50% more juice than the XL.

 

[Unknown Soldier]

Thx m8

 

[ANova]

The XL currently uses 1.4ns memory so in that respect it has alot of overclocking headroom from the default 500 MHz; this may change in the future though, especially if they lower the price. The core is an unknown at this point, however I wouldn't expect it to go higher then 550 MHz on average.

 

[Cartoon Corpse]

i dunno why anyone will buy an x1800xl. maybe if they gave away a free puppy with it?

any reason anyone have to buy an xl (instead of wait for xt if you're a diehard ati fan)?

and the 1300 sounds no better than my 9700pro!

 

[Joe DeFuria]

and the 1300 sounds no better than my 9700pro!

If the 1300 is as good as your 8 pipe, 256 bit SM 2.0 only 9700 pro, gamers in general should rejoice.

 

[Geo]

If the 1300 is as good as your 8 pipe, 256 bit SM 2.0 only 9700 pro, gamers in general should rejoice.

Good point, but I think I'll save my best "play that funky music, whiteboy" moves for when IGP is that good. . .

 

[swaaye]

9700 forever Amazing how it's still not considered a low end card!

 

[Joe DeFuria]

I think people continually underestimate the "difficulty" in driving the cost of these things down while maintaining all functionality and performance. Keep in mind that every generation, the "highest-end" parts keep on getting more and more expensive to produce than the previous high-end parts. (This doesn't look to change with the R580 rumors) The low end parts don't get any more expensive to compensate.

 

[Geo]

When I hear ATI squealing about margins on IGP, I think I get it. They are to be commended that apparently backing off on the performance is not the way they are going to address it.

But, see, you're being reasonable, and I'm just looking at what the need is. The gap between Intel IGP and tippy-top is getting alarming, don't you think?

 

[IbaneZ]

The XL is too weak. ATI is all about brute force these days. 500/500? Please. 625/750? Now you're talking. :smile:

 

[Geo]

The XL is too weak. ATI is all about brute force these days. 500/500? Please. 625/750? Now you're talking. :smile:

Gotta wonder just how prohibitive the cost would have been for a wee-bit faster memory. The ratios would suggest 1200 would be the right call for XL memory. And, alas, so do the benchmarks.

We're all a bit handicapped until the OC tool boys catch up with this part.

 

[Joe DeFuria]

The XL is too weak. ATI is all about brute force these days. 500/500? Please. 625/750? Now you're talking. :smile:

I think ATI's bigger problem with their line up may be with the 1600/RV530 part. I feel that 4 ROPs and 4 Texture units just don't quite have enough oopmh for the target market. IMO ATI will either need to seriously ramp those clock speeds, or replace the chip with something like a 8-1-3-2 part.

Even if ATI went with something like 4-2-3-2...with the RV530, that would probably go a long way to getting some extra performance with games currently on the market.

 

[Subtlesnake]

I think that ATI have backed themselves into a corner a bit with the XL naming scheme. If they do want to launch a faster part to take the low-high end back, what will they call it?

 

[IbaneZ]

The ratios would suggest 1200 would be the right call for XL memory. And, alas, so do the benchmarks.

Agreed. 1200 would have been a sweet spot.

I will buy an XL though. (I've actually already pre-ordered one) The 1.4ns memory should fly!

I hope it's not the 512 MB RAM on the XT that makes it rip.

 

[OpenGL guy]

I think ATI's bigger problem with their line up may be with the 1600/RV530 part. I feel that 4 ROPs and 4 Texture units just don't quite have enough oopmh for the target market. IMO ATI will either need to seriously ramp those clock speeds, or replace the chip with something like a 8-1-3-2 part.

So you're saying the chip needs to be twice as large? Sure doesn't look that way from the benchmarks. When I look at new games and the shaders they use, the ALU/TEX instruction ratio is usually greater than 4:1. Texturing power is less important on new games.

 

[Joe DeFuria]

So you're saying the chip needs to be twice as large?

Is 4-2-3-2 twice as large? (Or even 8-1-3-2?)

But of course, I would love to see a 8-1-3-2 at the same price. Not saying it's possible, but that's what I'd like to see.

Texturing power is less important on new games.

I agree, on many NEWER games, texturing power is less important. However, these things are going to sell based on a combination of benchmarks with newer games and older ones. I just think ATI might be a generation "too early" in de-emphasizing the importance of raw pixel / texel rate.

Why is the R520 16-1-1-1 and not 8-1-3-2, if universally, it's more important to have higher ALU to pixel/texel rate for "new games?"

 

[OpenGL guy]

Is 4-2-3-2 twice as large? (Or even 8-1-3-2?)

Think about it. 24 pixel shader ops per clock, 16 depth operations per clock, it would be a competitor for the 1800! What good does that do ATI?

But of course, I would love to see a 8-1-3-2 at the same price. Not saying it's possible, but that's what I'd like to see.

Well if you're going to fantasize, why would you fantasize about midrange products?

I agree, on many NEWER games, texturing power is less important. However, these things are going to sell based on a combination of benchmarks with newer games and older ones. I just think ATI might be a generation "too early" in de-emphasizing the importance of raw pixel / texel rate.

The chip is plenty fast at old texture limited games. The improvements in efficiency help make up deficits on paper.

 

[Geo]

Is 4-2-3-2 twice as large? (Or even 8-1-3-2?)

But of course, I would love to see a 8-1-3-2 at the same price. Not saying it's possible, but that's what I'd like to see.

Coming Nov. 2006 --the X1900GTO.

 

[Joe DeFuria]

Think about it. 24 pixel shader ops per clock, 16 depth operations per clock, it would be a competitor for the 1800! What good does that do ATI?

Just taking the RV530 and adding twice the texture capability (to make it 4-2-3-2) would put the RV530 in R520 league?

The chip is plenty fast at old texture limited games. The improvements in efficiency help make up deficits on paper.

Define "plenty fast." "Plenty fast" in an absolute sense, I would agree. "Plenty fast compared to the competition at the same price point...." not so sure. Especially come December when we'll actually see these parts.

Again, I think ATI went with the RIGHT choice with the 1800 at 16-1-1-1. However, based on what you're saying, ATI should have gone with 8-1-3-1. Why didn't they?