AMD RV770 refresh -> RV790

Will the AMD RV790 Architecture be Something Different?

This article provides a nice bit of perspective:

http://www.pcper.com/article.php?aid=675

During development of the R600, the decision was made to stop going for “top spot” with a large, complex die which would be expensive to produce and only exist as a $400+ product.
[...]
Going back to the RV670 (HD 3800 series), we see that it was essentially a shrunk R600. It contained the same number of stream units, RBEs, and texture units as the older R600, but it was a slightly optimized design that was produced one full node shrink below what the R600 was. The R600 was produced on a high speed 80 nm process from TSMC, while the RV670 was produced on TSMC’s then brand-new 55 nm process.
[...]
Now we are approaching the release of the RV790 and AMD has of course been very silent about it, and the only rumors we have heard about it up until this week are that it would be merely a speed optimized RV770 chip that would consume about the same amount of power and produce the same amount of heat as the older HD 4870 cards. Looking back at AMD’s design history, we can guess that this is probably so far from the truth that it was intentionally planted by AMD to potentially mislead NVIDIA into thinking that they were going to wait for 40 nm before introducing a new high end part based on a new or improved architecture.
Sure, I own a signature that's looking about 10% likely, but I think his conclusions are badly argued as he doesn't weigh the alternatives in any depth and is misguided on some significant points (e.g. the reason for the double-precision throughput of AMD's architecture).

Jawed
 
Sure, I own a signature that's looking about 10% likely, but I think his conclusions are badly argued as he doesn't weigh the alternatives in any depth and is misguided on some significant points (e.g. the reason for the double-precision throughput of AMD's architecture).

Jawed


With the supposed release less than a month away now. wouldn't those things be known about now? Even if the board partners speak of, just that why would we be expecting anything different?

But then, why would ati take over 9 months for what's basically a slightly altered and optimized chip?
a 20% upclocked RV770 would suggest to me that RV790 isn't what it was supposed to be.

Right now I'm leaning more towards this "RV770+" (which equals a GTX285 by all means) than a RV790 that would actually win.
 
New info about HD 4890 ( i have sample without drivers) - 800 SP cores, 850 MHz GPU, 3600 MHz MEMs, TDP 220W, 6pin + 8pin ..., similar cooler like HD 4870 ...
http://www.xtremesystems.org/forums/showthread.php?t=217833&page=9

So it seems nearly to be confirmed, that 4890 is just a high-power clock-up of RV770.

Lets see what the AIBs could squeeze out it for their OC-SKUs.
In one point the standard 4890 allready seems to beat the GTX 285 - power consumption. :LOL:

there was an edit by OBR:
But price will be very cool for that performance ... BUT YOU ALL will be surprised, what will release NV that same DAY! 6.4.2009 against HD 4890 .... hehehe
 
Take it with a grain of salt :) OBR told us, that he has a sample. Than he wrote, that RV790 is slower than GTX260-216, because he has it at home and he can compare it. Than he told us, that he has no compatible driver, so he can't test it (how could he find out, that the board is slower than GTX260-216?) - this question was never answered or explained. He didn't show any single picture of the card, any single screenshot, so... :rolleyes:
 
how can it be slower when the 4870-1GB is just as fast? Heck, how does he know anything if he can't get the card up?
 
AMD's official documents sent last month to board partners mentionned 800 SPs for RV790. Not many more info in there as these docs are used to give partners high level marketing advices to sell the product.
 
AMD's official documents sent last month to board partners mentionned 800 SPs for RV790. Not many more info in there as these docs are used to give partners high level marketing advices to sell the product.

Basically, that's what the slides from last November said too. "same process as RV770, significant increase in SCP and clock speeds"
 
http://www.xtremesystems.org/forums/showthread.php?t=217833&page=9

So it seems nearly to be confirmed, that 4890 is just a high-power clock-up of RV770.

Lets see what the AIBs could squeeze out it for their OC-SKUs.
In one point the standard 4890 allready seems to beat the GTX 285 - power consumption. :LOL:

there was an edit by OBR:

That dude is a fanboy. For many times in the past it lied about products (by RV770 time said that it was only a bit faster then RV670 and other lies).
He is the only with RV790 in the hands in the world :rolleyes:

Problem 1: If it were 220W it would definitely need 8pin + 8pin.
Problem 2: 100Mhz bump does not justify 220W (50W EXTRA) neither.
 
Last edited by a moderator:
I stopped reading OBRs post after he called people imbeciles and told people to cry because RV790 was slower than GTX 260 Core 216. I stopped believing his posts long before that.
 
No it wouldn't, 6+8pin is max the PCIE specs allow, and they allow up to 300W (75W slot 75W 6pin 150W 8pin)

true. I taked statement from Xtremesys.

But price will be very cool for that performance ... BUT YOU ALL will be surprised, what will release NV that same DAY! 6.4.2009 against HD 4890 .... hehehe

A renamed Geforce4 MX?
For the same day Nvidia has nothing. Only a new wave of price drops.
 
Oh man, I haven't been here in ages. Posted the article, then thought... wonder what the B3D fellows are having to say about it all. Then of course you all have 25 pages worth of a topic on RV790...

The PCPer article I wrote was merely speculative, and I of course could be so wrong as to be laughable (when am I not?). Just seems to me that it would be very strange if AMD were to design a new chip that just improved clockspeed and power consumption, especially when looking at their past design history.

As for how double precision is handled... my understanding is that in the group of 5 stream processors, one was double precision and the other 4 were single precision (plus that group of 5 has the general registers assigned to them and the branch prediction unit). Is my understanding of this incorrect?
 
Or..., one or more of the GT21x GPU's... ;)
GT212? The last we heard of, didnt tape out.

I'm not sure how Nvidia will fight a GTX285 equalling RV790 with anything other than the GT212.

The only logicial, possible reaction from Nvidia at this point seems to be .. price cuts. :sly:
 
Back
Top