ATi 4870X2(R700) PCB Debut

[G0ldBr1ck]

Must be a newer design from the other pic. theres no heat pipe on the first one.

[Omastar]

I think I just came.

[TedShred]

Would be awesome if 3870x2 H20 blox fit on these. You'd think, with the price of fullcover blox these days, that ATI and NVidia would try to allow for that to promote upgrading.

[strange|ife]

looks big. Hope the price isnt out there.

[Xoulz]

looks big. Hope the price isnt out there.

I don't think ATI is going to place the HD4870X2 too far out. $489 it will cripple Nvidia and there would be riots over this card!

[0g1]

Well, according to the cnet article, the R700 doesn't use crossfire at all to communicate between its two gpu's. "They've come up with this stellar way of doing inter-processor communications ... a new proprietary inter-processor communication technology."

The article implies that the processors communication isn't at the end of the pipeline, so it seems they would be working together in the pipeline, on the same frame, ie SFR.
Split Frame Rendering probably isn't the best way to desribe it because the GPU's are probably working on a much smaller level than half of a frame. Probably, the command processor splits commands up between GPU's. Indeed, the Command Processor has a direct link to the Hub, but it would also need a link to the other GPU. Whatever pixels are rendered by each GPU need to be sent to the other GPU to put into its own memory.

Who knows, but this 'stellar' communication link should be operating at a high frequency (core frequency) and allow for fairly high bandwidth from all the rendered pixels.

Of course the best way to do this would be to have a 115GB/s link ... maybe just connect up each GPU's memory controllers to the other GPU's memory, making it a 512 bit interface to each GPU ... but I dont think they would do this because of cost. So, it will probably be a lot less bandwidth than 115GB/s.

[martindh2005]

Anyone wants to join me on some guesses for the estimated price for release? I'll aim for $499

[Omastar]

Anyone wants to join me on some guesses for the estimated price for release? I'll aim for $499

I'll say $449.

[Rob94hawk]

When does this card come out?

[Hu1kamania]

If its the fastest card on the block by a considerable margin and considering the amount of DDR5 on that thing...I bet it will be closer to $600 or $487.20 Seriously though, I'm thinking there is going to be a shortage of these cards at release. Once things settle down though I could see this card settling at the $499 range.

[gojirasan]

The only rumors I have heard say $499, but we also had rumors of a $499 GTX280 right up to like a week before release or something. So who knows. AMD could price it at $599 if they wanted and it would still be a much better deal than the GTX280. The old ATI would definitely have priced it at something like $739 MSRP, but AMD tends to be a bit more aggressive with their pricing, especially now when they are in financial trouble. Some of you folks are truly optimists. Maybe they will just give the cards away or pay you to take one off their hands. After all AMD is making so much money they can afford to just give away their products. It's not like they have been losing millions of dollars over the past year. Too many people try to think with their emotions, believing something because they want it to be true.

[Zytek_Fan]

Excuse me while I go change my pants...

[Seraphiel]

Damn this looks like a wonderful solution for me. ATI/AMD is really looking excellent with their solutions.

I am still wondering, why many see this as a 2 x 800 SP solution, while viewing the 8800 GTX as a 128 SP and GTX 280 as a 240 SP solution. That is just flat out wrong and even GPU-Z gets these numbers wrong. If 8800 GTX is a 128 SP solution, then HD 4870 is a 160 SP solution. If HD 4870 is a 800 SP solution, then 8800 GTX is a 256 SP (and perhaps even a 384 SP) solution.

That ignored, I just love what appears to be a strong comeback by AMD/ATI for me personally. They really look strong and if I had the need for it, I would upgrade my two Ultras without question (but I don't, for now).

Looks like a R300 / R350 vs NV30 / NV35 scenario repeated, except that the prices are different.

[guess2098]

my 4870x2 doesn't boot, but windows can detect it and install the driver. (use another PCI-card to post)
it just no screen....

[Sly Fox]

my 4870x2 doesn't boot, but windows can detect it and install the driver. (use another PCI-card to post)
it just no screen....

Go on...

[Logos]

my 4870x2 doesn't boot, but windows can detect it and install the driver. (use another PCI-card to post)
it just no screen....

man I told you not to post about that other than in the XS secret section, now they'll all know we got those pre-release samples (btw mine boots fine)

[disruptfam]

how much do these use power wise!?

im hoping a 850w is enough for two

[Frank M]

my 4870x2 doesn't boot, but windows can detect it and install the driver. (use another PCI-card to post)
it just no screen....

that's too bad.
I'll give you a hand, I'll try it in my machine. Heck, maybe I'll even pay shipping.

[Helmore]

I am still wondering, why many see this as a 2 x 800 SP solution, while viewing the 8800 GTX as a 128 SP and GTX 280 as a 240 SP solution. That is just flat out wrong and even GPU-Z gets these numbers wrong. If 8800 GTX is a 128 SP solution, then HD 4870 is a 160 SP solution. If HD 4870 is a 800 SP solution, then 8800 GTX is a 256 SP (and perhaps even a 384 SP) solution.

I don't know where your numbers from or your math is seriously skewed, but that is just not right what you are saying there. It has also been know for a long long time that you just can't compare shader/pipeline (and whatever you want to call it) count between ATI and NVIDIA, if you could then the X850 cards from ATI would have been way faster than NVIDIA's 6800 cards for example and this was clearly not the case.

A GTX280 has 240 stream processors while a 4870 has 800, these SPs on the GTX280 can do 1 MADD and 1 MUL per cycle (1 MADD+1 MUL/clock cycle) while those SPs on the 4870 can do 1 MADD per cycle. So that means the 4870 can do 1600 FLOPs/clock cycle and the GTX280 can do 360 FLOPs/clock cycle, that's not to hard to see now is it? It's just that the SPs on the GTX280 run at a way higher clock and are a little more flexible because there is less dependency between the SPs, but if you would write code that is optimized specifically for each architecture then you can truly get 1200 GigaFLOPs out of the 4870 and only 933 GigaFLOPs out of the GTX280. But we by now know that these numbers mean nothing for in game results though.

[F.stop]

Anyone wants to join me on some guesses for the estimated price for release? I'll aim for $499

id say $499

[adamsleath]

id say $499

and if you want the australian price; just add 50%
no, but seriously...probly much like the 3870X2 at debut...hopefully($650aud roughly),
which would basically eliminate 9800gx2.

and if more than that i'll just put it in the gtx 280 category (ie those cards i cannot afford)

[annihilat0r]

$499 would be super fine, and if it doesn't have microstutter, it's my next card

[Piotrsama]

Anyone wants to join me on some guesses for the estimated price for release? I'll aim for $499

$549 here.

[dieqast]

$499 would be super fine, and if it doesn't have microstutter, it's my next card

*WIEWOOWIEWOOWIEWOO* Deputy Perkam, To Protect And Serve Xtreme Overclocking Cyberspace

[Shintai]

I don't know where your numbers from or your math is seriously skewed, but that is just not right what you are saying there. It has also been know for a long long time that you just can't compare shader/pipeline (and whatever you want to call it) count between ATI and NVIDIA, if you could then the X850 cards from ATI would have been way faster than NVIDIA's 6800 cards for example and this was clearly not the case.

A GTX280 has 240 stream processors while a 4870 has 800, these SPs on the GTX280 can do 1 MADD and 1 MUL per cycle (1 MADD+1 MUL/clock cycle) while those SPs on the 4870 can do 1 MADD per cycle. So that means the 4870 can do 1600 FLOPs/clock cycle and the GTX280 can do 360 FLOPs/clock cycle, that's not to hard to see now is it? It's just that the SPs on the GTX280 run at a way higher clock and are a little more flexible because there is less dependency between the SPs, but if you would write code that is optimized specifically for each architecture then you can truly get 1200 GigaFLOPs out of the 4870 and only 933 GigaFLOPs out of the GTX280. But we by now know that these numbers mean nothing for in game results though.

His math is actually right. Its yours thats wrong, unless you believe marketing.
Even your math is wrong, unless you count a MUL for half or something.

Its simply a simple shader vs complex shader discussion. And marketing couldnt stand being "less".

It would be like claiming a Kentsfield/Yorkfield is a 16 "something" CPU and Bloomsfield 32. And a Phenom X4 for 12. You would instantly debunk that, but for some reason you eat the same logic if applied to GPUs.