ATI Radeon HD 4000 Series discussion

[Helmore]

The only and also main difference I'm seeing is the bottom plate, or GPU and cooler contact area. The contact area on the cooler pictured over at VR-Zone forums has a square that's supposed to make contact with the GPU in line with the rest of the cooler and a lot smaller then on the 2900 cooler. This little square is positioned diagonally on the 2900 cooler. (I hope you guys can figure out what I mean )

[Papu]

Similar? No man, this is the same as 2900XT, in design and visuals. x1900 and 3870 are not... That (copper?) block with those heatpipes is a 2900XT exact copy.

afaik it looks like the shroud from a 2900pro? thats my xt and its not the same cooler

[AliG]

Because the GDDR5 memory is in final testing stage . I think that AMD will release only "Performance" RV770's part in May, equipped with GDDR3 memory and Flagship parts RV770XT&R700 with GDDR5 will come later at Computex time probably .

Hmm, that's possible, but you're forgetting the rv770xt IS "performance", most likely the 4850 will come out first as that's supposed to have either gddr4 or gddr3 (don't remember off the top of my head). But I highly doubt the 4870 will come out with gddr3 unless if 3rd party manufacturers decide to do that as an option same way there are 3870 gddr3 and 3850 gddr4 counter parts.

that cooler looks quite short!

also its interesting from the nordichardware link that ATi are planning on releasing a GDDR3 version before the GDDR5

Hopefully that means the 4870 will be quite short also. That's a smart move on ati's part to do so, but I'm pretty certain that's going to be the rv740 coming out with gddr3. But either way, I looking for to seeing the 4670's performance. Think about it, 240 shaders, 20 TMUs, and 256 bit memory with the possibility of gddr4. This thing will probably be at the performance of the 3850 or above, especially with af and aa enabled because it's a lot less bottlenecked than the r600 ever was. Plus it should be cheap as hell, so I'm considering to make it my next upgrade if the 4870 is out of the question (gt200 for sure, wayyy too expensive, even if the performance is there, yields will be terrible).

and the 2900 cooler looked like the x1900 cooler and 3870 , they all look very similar.

No that's not quite true. The x1900 coolers do not look like the 2900xt cooler, nor the majority of the x1950 coolers, the x1950xtx cooler however is a simpler version of the 2900xt's cooler. I just wish the 2900xtx's cooler would have been put on the 2900xt, that thing was a monster, something like 4 heatpipes and a blower attached, but it was long as hell because of it. And the 3870's cooler is also considerably different, larger better designed fan, no heatpipes, and quite a bit smaller.

However this rv770 cooler and the 2900xt cooler look very similar, but the 4870 has 2 heatpipes while 2900xt has 3, and as someone else pointed out the bases are slightly different. Most likely that's due to the r600 die being considerably larger and thus needing to be placed diagonally to conserve space.


But hey I'm not complaining, if this means low temps and silence, I'm all for it. The 3870 cooler was quiet, but wasn't great at heat removal because it was just a copper block. The 3850 cooler was kinda meh but looked kinda cool. The 8800gt's cooler just kinda...sucked...period. However the g92 gts cooler was both quiet and efficient at removing heat and that's why the g92 gts(s) are selling like hotcakes even though the 8800gt is like $10-$30 cheaper and offers almost the same performance. In other words, I think ati was taking notes on how well the g92 gts was selling compared to the 8800gt and 3870 (relative to how many cards the respective companies produced) and thus decided to do the same.

To be honest the 2 heatpipes probably aren't completely necessary, but once again, I'm not complaining, I just wonder what the new Ice Q 4 turbo cooler for the 4870 will be considering the 4870's stock cooler is better than the Ice q 3 turbo (with the exception of the UV housing)

[STaRGaZeR]

afaik it looks like the shroud from a 2900pro? thats my xt and its not the same cooler

Lately they added a third heatpipe for the 2900Pro and XT as they use the same cooler, but the original 2900XT cooler was like mine, one of the first Sapphires out there:



The red plate, the fan... but the heatpipes are bended a bit different though.

[HKPolice]

RV770 is both impressive and disapointing at the same time.

RV670 Specs:
16 TMUs
320 SP arranged in 4 SIMD arrays of 80sp each
16 ROPs
256Bit DDR3/4 Ram
666Million Transistors on 55nm

RV770 Specs:
32 TMUs
800 SPs arranged in 5 SIMD arrays of 160sp each
16 ROPs
256Bit DDR4/5 Ram
~830Million transistors on 55nm

Now the impressive part is that ATi managed to more than double the number of SPs and double the amount of TMUs and only add about 200Mill transistors. The disapointing part is that overall performance is only expected to be 50% higher than RV670. One would expect more than that especially since RV770 will finally offload AA to the ROPs.

There is no "true next gen" chip for the DX10 generation, even the GT200 is just a revamped G92 with more shaders/ROPs and TMUs. Things in the GPU world will be boring until DX11 comes.

[adamsleath]

overall performance is only expected to be 50% higher than RV670

nice for a value card tho' assuming it pans out.

[AliG]

RV770 is both impressive and disapointing at the same time.

RV670 Specs:
16 TMUs
320 SP arranged in 4 SIMD arrays of 80sp each
16 ROPs
256Bit DDR3/4 Ram
666Million Transistors on 55nm

RV770 Specs:
32 TMUs
800 SPs arranged in 5 SIMD arrays of 160sp each
16 ROPs
256Bit DDR4/5 Ram
~830Million transistors on 55nm

Now the impressive part is that ATi managed to more than double the number of SPs and double the amount of TMUs and only add about 200Mill transistors. The disapointing part is that overall performance is only expected to be 50% higher than RV670. One would expect more than that especially since RV770 will finally offload AA to the ROPs.

There is no "true next gen" chip for the DX10 generation, even the GT200 is just a revamped G92 with more shaders/ROPs and TMUs. Things in the GPU world will be boring until DX11 comes.

Thing is I don't think ati is offloading aa to the ROPs, that's probably the reason for the high shader increase and no ROP increase, and that could very well mean that there's a need for all those shaders. However I don't think that transistor count is right, there's no way that's possible to more than double the shader count with only adding 200 some transistors

[NH|Delph1]

RV770 is both impressive and disapointing at the same time.

RV670 Specs:
16 TMUs
320 SP arranged in 4 SIMD arrays of 80sp each
16 ROPs
256Bit DDR3/4 Ram
666Million Transistors on 55nm

RV770 Specs:
32 TMUs
800 SPs arranged in 5 SIMD arrays of 160sp each
16 ROPs
256Bit DDR4/5 Ram
~830Million transistors on 55nm

Now the impressive part is that ATi managed to more than double the number of SPs and double the amount of TMUs and only add about 200Mill transistors. The disapointing part is that overall performance is only expected to be 50% higher than RV670. One would expect more than that especially since RV770 will finally offload AA to the ROPs.

There is no "true next gen" chip for the DX10 generation, even the GT200 is just a revamped G92 with more shaders/ROPs and TMUs. Things in the GPU world will be boring until DX11 comes.

The first story about 50% faster was posted by Fudzilla, I think he beat me with a few hours on that one. I'm not sure what his source told him, but mine was very vague and after hearing some more I was thinking that the "50% faster" was based on the fact that RV770 was believed to have 50% more SPs, I.e. 480 all in all. I can't say that for certain though, it's just a speculation on my behalf.

I've been hearing the figure 480 over and over again, but the context has been a bit different every time. Right now it seems that we're looking at a total of 800 SPs, which means 480 added SPs. If I'm right about my first speculation, we can throw the early stories of 50% faster out the window.

But there's one more thing to it. (I wrote this is in the original "50% faster" article) If you go back and check how much faster each [true] generation than the last one has been, you won't see much higher figures than 50%. With each generation that has passed, the last generation high-end has performed like the new generation mid-range. But with the lack of a genuine high-end chip from AMD/ATI, I can surely understand that most people are terrified by this.

And then we have the fact that AMD/ATI has come up with a way for two chips to share a memory buffer, which at least makes me really eager to see how the dual-GPU R700 will perform, not to mention two cards with four GPUs and just two memory buffers.

I'm very skeptical toward whether AMD/ATI will ever make a new super GPU. It just makes more sense to spread the workload over several GPUs instead, to them. And considering how well the concept of R680 worked out, I can only imagine that AMD/ATI will keep heading down that road. The fact that the RV670 was a poor performer in AA/AF situations shouldn't cloud your opinion of the R680 concept. Just plug it in and it works. No hassle.

//Andreas

[flopper]

And then we have the fact that AMD/ATI has come up with a way for two chips to share a memory buffer, which at least makes me really eager to see how the dual-GPU R700 will perform, not to mention two cards with four GPUs and just two memory buffers.

I'm very skeptical toward whether AMD/ATI will ever make a new super GPU. It just makes more sense to spread the workload over several GPUs instead, to them.
Just plug it in and it works. No hassle.

//Andreas

what is the advantage of that solution with memory buffers?

and the cards with ati/amd seems to be a better driver support due to the program they run for a while.

however, rumours aside, we want facts

[largon]

Shared memory would eliminate the usual drawback of mirroring identical data that both GPUs use thus saving memory for other things like higher res textures and what-not.

[NH|Delph1]

what is the advantage of that solution with memory buffers?

and the cards with ati/amd seems to be a better driver support due to the program they run for a while.

however, rumours aside, we want facts

Well, in the case of CrossFire, it would double the amount of memory.

Today the data is mirrored in each buffer, which means that the 3870X2 1GB, really only uses 512MB, and two 3870X2 cards have the same data mirrored four times.

With two GPUs sharing one 1GB buffer, the entire 1GB would be used in CrossFire. Link two of these cards and they would still have just 1GB, even though the buffer is 1+1GB. But we've doubled the buffer, or halved the loss of effective memory. The data would only be mirrored once, instead of four times.

That's theory and rumor though

//Andreas

[RedBull78]

Well, in the case of CrossFire, it would double the amount of memory.

Today the data is mirrored in each buffer, which means that the 3870X2 1GB, really only uses 512MB, and two 3870X2 cards have the same data mirrored four times.

With two GPUs sharing one 1GB buffer, the entire 1GB would be used in Crossfire. Link two of these cards and they would still have just 1GB, even though the buffer is 1+1GB. But we've doubled the buffer, or halved the loss of effective memory. The data would only be mirrored once, instead of four times.

That's theory and rumor though

//Andreas

Nice. That's like eliminating the middle man. I just hope you're right. No mirrors at all and just one main memory buffer. Huge improvement in latency as well and no more slowdowns in multi-GPU solutions that was talked about a few months ago...

[Unbornchild]

No more "mini-stuttering" (i suppose) you mean. Which would be awesome.

[Marios]

I think nothing is clear yet. Just guessing.
RV770pro GDDR3 55nm soon. RV770XT 45nm GDDR5 4GHz 680SP, 36TMU in July.

[Helmore]

I think nothing is clear yet. Just guessing.
RV770pro GDDR3 55nm soon. RV770XT 45nm GDDR5 4GHz 680SP, 36TMU in July.

If there is going to be a R(V)7** chip on 45nm then it will probably not be named RV770. Oh and there wont be any 45 nm. GPU chips around July (as far as I know, but I don't have any form of news source like DilTech and BenchZowner do).

[AliG]

Where are you getting those numbers from? The rv770xt specs you named have absolutely nothing to back up what you have

[Helmore]

Where are you getting those numbers from? The rv770xt specs you named have absolutely nothing to back up what you have

I think he's just pulling some numbers out of his behind, after sitting on them for a little while, if you get what I mean .

[LordEC911]

45nm won't be ready until late Q3 and at that time TSMC's 40nm will also be ready...

[HKPolice]

I know from insider sources + techreport's own editor also agreed that the RV770 offloads AA to ROPs. The original design of the R600 was supposed to do AA in ROPs but it never worked properly so they were forced to write drivers that used shaders for AA.

Some ppl claim that this wasn't a big performance hit and it was the TMUs limiting the RV670 more than anything. That I'm not so sure about.

[LordEC911]

I know from insider sources + techreport's own editor also agreed that the RV770 offloads AA to ROPs. The original design of the R600 was supposed to do AA in ROPs but it never worked properly so they were forced to write drivers that used shaders for AA.

I doubt that...
The original DX10 spec made it mandatory for software AA.

[HKPolice]

I doubt that...
The original DX10 spec made it mandatory for software AA.

Doubt all you want, doesn't make it true.

I have sources who work @ ATi

DX10 spec is a joke. This is why there isn't a single game that runs faster on DX10 vs DX9 and multi-GPU setups take a huge hit in DX10 as well. This is also why both ATi and nV are holding back on developing any 'true next gen' DX10 chips. Especially with the US economy the way it is, they're playing it safe until DX11.

[Papu]

DX10 spec is a joke. This is why there isn't a single game that runs faster on DX10 vs DX9 and multi-GPU setups take a huge hit in DX10 as well.

Assassins Cread Low dx10>dx9

[Zytek_Fan]

Doubt all you want, doesn't make it true.

I have sources who work @ ATi

DX10 spec is a joke. This is why there isn't a single game that runs faster on DX10 vs DX9 and multi-GPU setups take a huge hit in DX10 as well. This is also why both ATi and nV are holding back on developing any 'true next gen' DX10 chips. Especially with the US economy the way it is, they're playing it safe until DX11.

The GPU companies could also be at fault for DX10's issues.

[KoHaN69]

AA on ROPs would rock so hard, it would roll!

shared mem + AA rops + 2 gpus = epic dual-gpu singlecard win

[hstuehmeyer2000]

Assassins Cread Low dx10>dx9

yeap