RV740 Taped-Out, First 40nm GPU From AMD

[btarunr]

The manufacturing process technologies for graphics processors that dominated the 2008 were TSMC 65nm and 55nm. While AMD made the transition to 55nm over an year ago, it was a little later when we started seeing 55nm versions of existing NVIDIA GPUs.

A month into TSMC's announcement of of its 40nm bulk production node, AMD has reportedly taped out its first 40nm GPU, the RV740. While this is no high-end GPU, it is supposed to be the first successful port of AMD's GPU architecture to the new node. The RV740 succeeds the RV730, the GPU that went into the making the Radeon HD 4670. It is a mainstream GPU that ideally should make it to the sub-$100 graphics card segment. With RV740, AMD gains some experience as a manufacturer as it works on the RV870 "Lil' Dragon", the next generation GPU from the red camp. A product based on the RV740 can be expected only after Q1 2009.

View at TechPowerUp Main Site

 

[Black Hades]

Good strategy.. test it out in the lower end market, master it by the time RV870 is out.

 

[kylew]

Hopefully we'll start seeing laptops with decent ATi GPUs that don't cost a fortune. A laptop with an RV740 would be a pretty good lower power laptop for games. I'd buy one

 

[DarkMatter]

Good strategy.. test it out in the lower end market, master it by the time RV870 is out.

I'm not so sure actually. Bear in mind that mainstream market is the bulk of their income, of anyone's income. If something goes wrong it could be the bigger market which would suffer. Then again delays on lower end cards might not be so dramatic and widespread onto the media, should they happen, so I don't know...

 

[truehighroller1]

I'm not so sure actually. Bear in mind that mainstream market is the bulk of their income, of anyone's income. If something goes wrong it could be the bigger market which would suffer. Then again delays on lower end cards might not be so dramatic and widespread onto the media, should they happen, so I don't know...

I could care less until I get a patch / Driver, that fixes the memory leak in GTA IV .

 

[Black Hades]

I'm not so sure actually. Bear in mind that mainstream market is the bulk of their income, of anyone's income. If something goes wrong it could be the bigger market which would suffer. Then again delays on lower end cards might not be so dramatic and widespread onto the media, should they happen, so I don't know...

Yes true, but then again the mainstream doesn't know what 40nm or RV870 means anyway...
And I doubt AMD/ATi would roll out a faulty product. Mainstream will be happy with it's new and improved power efficiency and therfore thermal envelope.

Most engineering obstacles they'll get from die shrinkage should be ironed out through harvesting and repair techniques that both ATi and Nvidia rely on. It's not nearly as big of a gamble as the HD 4k concept was to begin with.

We'll see anyway

EDIT:
What I am worried about is the next generation of cards... this thing is pretty neat right now for us with all this competiton. I hope this never changes

 

[lemonadesoda]

The manufacturing process technologies for graphics processors that dominated the 2008 were TSMC 65nm and 55nm. While AMD made the transition to 55nm over an year ago, it was a little later when we started seeing 55nm versions of existing NVIDIA GPUs.

A month into TSMC's announcement of of its 40nm bulk production node, AMD has reportedly taped out its first 40nm GPU, the RV740.

tick-tock after all

 

[btarunr]

tick-tock after all

Sure. The last time we talked about it, it was about NVIDIA

 

[alwayssts]

Am I the only one that still thinks Rv740 is the refresh part? ~3xRv670 (sans ROPs), or more accurately, 1.2x rv770, a rv770 with 12 arrays? We know Rv870 isn't the 960sp part...4900 parts are coming soon...and there is no rv775/rv790.

I also think a chip with ~rv730 (or even 2x) performance on 40nm would be ridiculously small, probably the size of RV710, certainly under 100mm2. It couldn't have a 128-bit bus, let alone a 256-bit bus needed to make an upgrade over rv730. It just doesn't make sense.

I think it becomes the perfect upgrade to rv770, and 40nm should allow it to be a similar size to rv730 (150mm2), at little over a billion transistors. A 128-bit bus with 4000mhz gddr5 could match 4850, and 7gbps gddr5 part could equal 4870 in bandwidth.

If we do the math, 1.5TF would only take 781mhz, 1.25TF 650mhz. A 25% upgrade over existing parts.

Now wouldn't that be a great "value" - oriented part? A upgrade to RV770, but to be placed below RV870. It also would make the value parts roughly follow the "2.5x" strategy I think ATi is going for with new processes vs rv730...while retaining the same die size, as well as floating technology down from performance to mainsteam.

On top of that, I wouldn't be surprised to see such parts, if it is what they are doing, residing in the 75W without-a-connector territory. RV730, at roughly the same size uses 59W @ 750mhz.

Such a part could probably get pretty close to taking on a GTX285 (at 400mm2+), and with a die size probably around 1/3.

It'd look something like this:

RV740pro:

150mm2@40nm
128-bit
650/4000
960 shaders
48 TMUs
16 ROPs
1.25TF
64gbps bandwidth

RV740XT:
150mm2@40nm
128-bit
775/7000
960 shaders
48 TMUs
16 ROPs
~1.5TF
112gbps bandwidth
Now, I'm not saying that's what they're doing...But wouldn't that make sense? It could compete now, and perhaps compete with the 40nm GT200 when it arrives on price, because it would be a smaller die and cheaper to make. I think GT212 will be 200-225mm2, 256-bit, gddr5.

But it's okay to call me crazy, I'm the guy that still thinks 870 is 2000 shaders @ 205mm2... 2.5 and 3TF parts (625 and 750mhz).

It may not be the direction they go, but if it is, I will find it amusing how well they continue to spin the press to confuse Nvidia (just as they did with RV770, and seem to be with RV870.)

/Dream

 

[DarkMatter]

Yes true, but then again the mainstream doesn't know what 40nm or RV870 means anyway...
And I doubt AMD/ATi would roll out a faulty product. Mainstream will be happy with it's new and improved power efficiency and therfore thermal envelope.

Most engineering obstacles they'll get from die shrinkage should be ironed out through harvesting and repair techniques that both ATi and Nvidia rely on. It's not nearly as big of a gamble as the HD 4k concept was to begin with.

We'll see anyway

EDIT:
What I am worried about is the next generation of cards... this thing is pretty neat right now for us with all this competiton. I hope this never changes

Well, Ati does repairability, not harvesting. At least in the first place. If things went wrong they might need to rely on harvesting a lot and that would make them suffer economically, because the extra stuff in the chips due to repairability would not be working (=wasted silicon), not to metion that in the high-end you can rely on chip binning for a time, but I doubt that could work on the mainstream. And all that would happen in a market segment where you can't really fight that back with slightly higher prices...

AFAIK ATi, Nvidia and pretty much every chipmaker has been releasing high-end first for a reason, and the reason they always give is the one I am "defending". It's not as if I knew that first hand, if it's wrong it's because THEY ALL were wrong, until know...

 

[PCpraiser100]

Nice, a 40nm GPU. I like how these cores clock if they OC. If AMD downgrades this should play a good role for AMD to get started to Fusion teasing with media.

 

[Steevo]

I agree, a 40nm should overclocke quite well, and if X-fire scales well with this as well, populating four slots and having a cool running system for cheap might be the way to go.

 

[ShadowFold]

What is an RV740? 4600?

 

[kysg]

What is an RV740? 4600?

As they stated 40nm die shrink of rv 730, don't worry about what number they give it. Time will tell what the red camp plans to pull.

 

[laszlo]

as i see amd is ahead nvidia with shrinkage and i think they'll implement the same tech. regarding power leaks as in phenom 2 so we'll have a few small excellent gpu's on 40nm which don't have to compete with green flagships ones,who need the gold crown when the silver is worth more in sales and price/perf/power ... just my 2 c...

 

[WarEagleAU]

I like starting with the lower segment first. Id love to see some reviews when possible.

 

[insider]

The low end sector is where the high volume sales are at and the lowest profit margins, the 40nm process will improve those tiny margins a bit.

 

[Bjorn_Of_Iceland]

yep.. this is how they catch up.. lower fab process than nvidia's.. sooner or later theyl hit the bottom and no more running for em'..

 

[DarkMatter]

Take into account that new processes are more expensive too. I supose they know what they are doing...

 

[Tatty_Two]

Thing is, it sounds good....a 45nm GPU and all that, it actually HAS gotta run cool with appropriatly decent coolers, look at many of the 55nm cards....they actually ran hotter than the oppositions 65nm cards, ATi really have to get it right otherwise, with the increased costs, they could lose out big time.......lets hope they do get it right......let the wars begin!

 

[wolf]

maybe this is the foundation of what theyre using for the multi core gpu, think of it, four of these babies on one 40nm die

32 rops, 1280 sp's, 512-bit gddr5, hello!

 

[DarkMatter]

maybe this is the foundation of what theyre using for the multi core gpu, think of it, four of these babies on one 40nm die

32 rops, 1280 sp's, 512-bit gddr5, hello!

And what would be the advantage of that versus a monolithic GPU with those specs?? It would use MORE silicon for the same performance and I can't see any benefit on that. Today GPUs are made in a way that harvesting is extremely easy and node independent. The GT200 for example has 10 SP nodes and 8 ROP nodes with a 64bit MC each. If one of those fails the chip is still usable in GTX260 form. Your hypothetical chip would be better because you can scrap 1/4 of the chip at once if one of it's units fail?? If you wanted (go figure why...) you could still do it in any modern GPU with the added value that you can choose which parts you want to scrap.

Multi CPU on a die works for CPUs because you can't scrap its units if they fail, just the cache and because it's the better way of reaching parallelism. But GPUs are inherently parallel already.

Another issue is that if your elemental unit (each of those 4 GPUs) is a low-end part what are you going to do wth the ones that fail? Create an even lower end card? Foresee that will happen and use self-repairability in conjuntion to harvesting? A bad idea in any case IMHO, because you would inevitably end up with too many different options, this kind of chips (following your example):

1 - Chips with 4 full cores.
2 - Chips with 4 cores with 28 ROPS (448 bits) and 1280 SPs.
3 - 4 cores with 32 ROPS and say 800-960 SPs.
4 - 4 cores with 24-28 ROPS and 800-960 SPs.
[5 to 8] Repeat 1-4 but with 3 cores.
[9-12] Repeat 1-4 with 2 cores.
13 - Chips where each core is different.

*Note that the clockability of each core has not been mentioned, but that would just add one more layer of complexity, which is not good.

As you can see, there are too many different chips there, so you would have to scrap a lot of cores in order to create a defined lineup. You think Nvidia has released too many cards lately? That's a joke compared to what would they have to do for the above to be anything close to profitable.

 

[wolf]

ahh well at least the thought was nice to roll with....

sitting on a GTX260 anyway so very happy.

 

[Hayder_Master]

40 nm , coooooooooooool like ice