Under the Hood of G84

So the quick and dirty summary of the changes is that the G84 is a reduced width G80 with a higher proportion of texture to shader hardware and a reworked PureVideo processing engine (dubbed VP2 as opposed to G80's VP1). Because there are fewer ROPs, fill rate and antialiasing capabilities will be reduced from the G80 as well. This isn't as necessary on a budget card where shader power won't be able to keep up with huge resolutions either.

We expect the target audience of the 8600 series to be running 1280x1024 resolution panels. Of course, some people will be running larger panels and we will test some higher resolutions to see what kind of capabilities the hardware has, but above 1600x1200 tests are somewhat academic. As 1080p TVs become more popular in the coming years, however, we may start putting pressure on graphics makers to target 1920x1200 as their standard resolution for mainstream parts even if average computer monitor sizes weigh in with fewer pixels.

In order to achieve playable performance at 1280x1024 with good quality settings, NVIDIA has gone with 32 shaders, 16 texture address units, and 8 ROPs. Here's the full breakdown:

GeForce 8600/8500 Hardware
GeForce 8600 GTS GeForce 8600 GT GeForce 8500
Stream Processors 32 32 16
Texture Address / Filtering 16/16 16/16 8/8
ROPs 8 8 4
Core Clock 675 MHz 540 MHz 450 MHz
Shader Clock 1.45 GHz 1.19 GHz 900 MHz
Memory Clock (Data Rate) 2 GHz 1.4 GHz 800 MHz
Memory Bus Width 128-bit 128-bit 128-bit
Frame Buffer 256 MB 256 MB 256MB / 512MB
Outputs 2x dual-link DVI 2x dual-link DVI ?
Transistor count 289 M 289 M ?
Price $200 - $230 $150 - $160 $90 - $130

We'll tackle the 8500 in more depth when we have hardware. For now, we'll include the data as reference. As for the 8600, right out of the gate, 32 SPs mean one third the clock for clock shader power of the 8800 GTS. At the same time, NVIDIA has increased the ratio of Texture address units to SPs from 1:4 to 1:2. We also see a 1:1 ratio of texture address and filter units. These changes prompted NVIDIA to further optimize their scheduling algorithms.

The combination of greater resource availability and improved scheduling allow for increased efficiency. In other words, clock for clock, G84 SPs are more efficient than G80 SPs. This makes it harder to compare performance based on specifications. Apparently stencil culling performance has also been improved, which should help boost algorithms like the Doom 3 engine's shadowing technique. NVIDIA didn't give us any detail on how stencil culling performance was improved, but indicated that this, among other things, was also tweaked with the new hardware.

Top this off with the fact that G84 has also been enhanced for higher clock speeds than G80 and we can expect much more work to be done by each SP per second than on 8800 hardware. Exactly how much is something we don't have an easy way of measuring as changes in efficiency will vary by the algorithms running on the hardware as well.

With 256 MB of memory on a 128-bit bus, we can expect a little more memory pressure than on the 8800 series. The 2 x 64-bit wide channels provide 40% of the bus width of an 8800 GTS. This isn't as cut down as the number of SPs; remember that the texture address units have only been reduced from 24 on the 8800 GTS to 16 on the 8600 series. Certainly the reduction of 20 ROPs to 8 will help cut down on memory traffic, but that extra texturing power won't be insignificant. While we don't have quantitative measurements, our impression is that memory bandwidth is more important in NVIDIA's more finely grained unified architecture than it was with the GeForce 7 series pipelined architecture. Sticking with a 128-bit memory interface for their mainstream part might work this time around, but depending on what we see from game developers over the next six months, this could easily change in the near future.

Let's round out our architectural discussion with a nice block diagram for the 8600 series:



We can see very clearly that this is a cut down G80. As we have discussed, many of these blocks have been tweaked and enhanced to provide more efficient processing. The fundamental function of each block remains the same, and the inside of each SP remains unchanged as well. The features supported are also the same as G80. For 8500 hardware, based on G86, we drop down from two blocks of Shaders and ROPs to one each.

Two full dual-link DVI ports on a $150 card is a very nice addition. With the move from analog to digital displays, seeing a reduction in maximum resolution on budget parts because of single-link bandwidth limitations, while not devastating, isn't desirable. There are tradeoffs in moving from analog to digital display hardware, and now an additional issue has a resolution. Now we just need to see display makers crank up pixel density and improve color space without reducing response time and this old Sony GDM-F520 can finally rest in peace.

In the video output front, G84 makes a major improvement over all other graphics cards on the market: G84 based hardware supporting HDCP will be capable of HDCP over dual-link connections. This is a major feature, as a handful of larger widescreen monitors like Dell's 30" only support 1920x1080 with a dual-link connection. Unless both links are protected with HDCP, software players will refuse to play AACS protected HD content. NVIDIA has found a way around the problem by using one key ROM but sending the key over both links. The monitor is able to handle HDCP connections on both links, and is able to display the video properly at the right resolution.

As for manufacturing, the G84 is still an 80 nm part. While G80 is impressively huge at 681M transistors, G84 is "only" 289M transistors. This puts it at nearly the same transistor count as G71 (7900 GTX). While performance of the 8600 series doesn't quite compare to the 7900 GTX, the 80 nm process makes smaller die sizes (and lower prices) possible.

In addition to all this, PureVideo has received a significant boost this time around.

Index The New Face of PureVideo HD
POST A COMMENT

60 Comments

View All Comments

  • deathwalker - Wednesday, April 18, 2007 - link

    So, whats the word on the Ultra version of the 8600? Has that fallen to the wayside? Reply
  • crystal clear - Wednesday, April 18, 2007 - link

    Interview: NVIDIA's Keita Iida
    The future of Direct X, Crysis and PS3 under the spotlight.

    Keita Iida, Director of Content Management at NVIDIA sat down with IGN AU to discuss all things Direct X 10 and the evolution of their Geforce graphics cards. Iida goes into detail on the differences between developing for the PS3's RSX graphics processor, and the latest development tools to hit the scene.


    quote:
    Selected portions of the interview-


    IGN AU: What are your thoughts on Microsoft effectively forcing gamers to upgrade to Vista in order to run Direct X 10 - when there's no real reason why it can't run on Windows XP?

    Keita Iida: It's a business and marketing decision.

    IGN AU: Can you comment on what happened with NVIDIA's Vista drivers? You guys have had access to Vista for years to build drivers and at the launch of Vista there were no drivers. The ones that are out now are still basically crippled. Why did this happen?

    Keita Iida: On a high level, we had to prioritise. In our case, we have DX9, DX10, multiple APIs, Vista and XP - the driver models are completely different, and the DX9 and 10 drivers are completely different. Then you have single- and multi-card SLI - there are many variables to consider. Given that we were so far ahead with DX10 hardware, we've had to make sure that the drivers, although not necessarily available to a wide degree, or not stable, were good enough from a development standpoint.

    If you compare our situation to our competitor's, we have double the variables to consider when we write the drivers; they have much more time to optimise and make sure their drivers work well on their DX10 hardware when it comes out. We've had to balance our priorities between making sure we have proper DX10 feature-supported drivers to facilitate development of DX10 content, but also make sure that the end user will have a good experience on Vista. To some degree, I think that we may have underestimated how many resources were necessary to have a stable Vista driver off the bat. I can assure you and your readers that our first priority right now is not performance, not anything else; it's stability and all the features supported on Vista.

    IGN AU: So what kind of timeline are we looking at until the end user can be comfortable with Vista drivers? With DX9 drivers that work as stably and quickly as they do with XP?

    Keita Iida: We're ramping up the frequency of our Vista driver releases. Users will probably understand that we release a number of beta drivers on our site, so we're making incremental progress. We believe that, in a very short time we will have addressed the vast majority, if not all of the issues. We've had teams who were working on other projects who have mobilised to make sure that as quickly as possible we have the drivers fixed. I'm not going to give you an exact timeframe, but it's going to be very soon. We're disappointed that we couldn't do it right off the bat, but we hear what everyone is saying and we're willing to fix it.

    http://pc.ign.com/articles/780/780314p1.html">http://pc.ign.com/articles/780/780314p1.html
    Reply
  • xpose - Tuesday, April 17, 2007 - link

    This next gen purevideo stuff sounds amazing. I thought I was gonna have to get a new motherboard and dual core cpu to play some HD-DVD content smoothly. Please, do try and rush testing the purevideo stuff ASAP. Blu-ray and hd-dvd is growing. . . Reply
  • shabby - Tuesday, April 17, 2007 - link

    128bit/256meg for $200 bucks? Gimme a break. Reply
  • Sunrise089 - Tuesday, April 17, 2007 - link

    Unless these cards are majically fast under DX10 (and we all know they won't be, they will play Crysis, but not quickly) they offer less performance than even midrange parts from the last get.

    Anyone remember how a 6600GT offreed 9800pro beating performance, and how nVidia sold millions of them. I don't see that happening here. What I do see is a wait-and-see attitude. Does anyone else think it's VERY suspicious that there are no 64 shader cards? Here is what may happen: nVidia waits for the midrange AMD cards to emerge. If they offer better performance, nVidia slashes prices of these and releases a 8800GS with 64 shaders for $200. I won't be surprised at all if that's what we have in 3 months.
    Reply
  • JarredWalton - Tuesday, April 17, 2007 - link

    We've got 128 SP on the GTX, 96 on the GTS... and then 32 on the G84. I'd say there's definitely room for 64 SP from NVIDIA, and possibly 48 SP as well. Will they go that route, though? Unless they've already been working on it, doing a new chip will cost quite a bit of time and effort. I was expecting 8600 to be 64 SP and 8300 to be 32 SP before we had any details, but then the 8600 probably would have been too close to the 8800. Reply
  • kilkennycat - Tuesday, April 17, 2007 - link

    Er, wait (not too long) for nVidia's re-roll of the 8xxx-series on 65nm... You might just get your wish. I believe that nV is copying Intel's 'tic-toc' process strategy - architecture and go to production on a mature process (80nm half-node), then transfer and refine the implementation on the new process. Note the interesting and important tweaks in the implementation of the 8600 vs 8800... which gives a glimpse of the future 65nm 9xxx(??)-family architecture but with higher numbers of stream-processors and high-precision math processing for the expected GPGPU applications.

    nVidia has already hinted that the successor to the 8800 will be available before the end of 2007, and no doubt will be on 65nm for the obvious cost and yield reasons. If the R600 turns out to be a true contender for the 8800 "crown" in the same price-range, then I fully expect nV to accelerate the appearance of the 8800 successor. No doubt the design was started long before the 8800 itself was production-available.
    Reply
  • Toebot - Tuesday, April 17, 2007 - link

    No, nothing to sneeze at, just something to blow my nose on! Utter wretch. This card is NVidia's attempt to milk the Vista market, nothing more. Reply
  • DerekWilson - Tuesday, April 17, 2007 - link

    We should at least wait and see what DX10 performance looks like first. Reply
  • AdamK47 - Tuesday, April 17, 2007 - link

    With what software? Reply

Log in

Don't have an account? Sign up now