AMD's Radeon HD 5870: Bringing About the Next Generation Of GPUs
by Ryan Smith on September 23, 2009 9:00 AM EST- Posted in
- GPUs
Meet the Rest of the Evergreen Family
Somewhere on the way to Cypress, AMD’s small die strategy got slightly off-track.
AMD’s small-die strategy for RV770
Cypress is 334mm2, compared to 260mm2 for RV770. In that space they can pack 2.15 billion transistors, versus 956 million on the RV770, and come out at a load power of 188W versus 160W on the RV770. AMD called 256mm2 their sweet spot for the small die strategy, and Cypress missed that sweet spot.
The cost of missing the sweet spot is that by missing the size, they’re missing the price. The Cypress cards are $379 and $259, compared to $299 and $199 that the original small die strategy dictated. This has resulted in a hole in the Evergreen family, which is why we’re going to see one more member than usual.
As Cypress is the base chip, there are 4 designs and 3 different chips that will be derived from it. Above Cypress is Hemlock, which will be the requisite X2 part using a pair of Cypress cores. Hemlock is going to be interesting to watch not just for its performance, but because by missing their sweet spot, AMD is running a bit hot. A literal pair of 5870s is 376W, which is well over the 300W limit of a 6-pin + 8-pin power configuration. AMD saves some power in a single card (which is how they got the 4870 under the limit) but it likely won’t be enough. We’ll be keeping an eye on this matter to see what AMD ends up doing to get Hemlock out the door at the right power load. As scheduled we should see Hemlock before the end of the year, although given the supply problems for Cypress that we mentioned earlier, it’s going to be close.
The “new” member of the Evergreen family is Juniper, a part born out of the fact that Cypress was too big. Juniper is the part that’s going to let AMD compete in the <$200 category that the 4850 was launched in. It’s going to be a cut-down version of Cypress, and we know from AMD’s simulation testing that it’s going to be a 14 SIMD part. We would wager that it’s going to lose some ROPs too. As AMD does not believe they’re particularly bandwidth limited at this time with GDDR5, we wouldn’t be surprised to see a smaller bus too (perhaps 192bit?). Juniper based cards are expected in the November timeframe.
Finally at the bottom we have Redwood and Cedar, the Evergreen family’s compliments to RV710 and RV730. These will be the low-end parts derived from Cypress, and will launch in Q1 of 2010. All told, AMD will be launching 4 chips in less than 6 months, giving them a top-to-bottom range of DX11 parts. The launch of 4 chips in such a short time frame is something their engineering staff is very proud of.
Facebook
Twitter
RSS
327 Comments
View All Comments
BoFox - Friday, November 6, 2009 - link
Yep, that's turning up LOD to -1 or -2 depending on which game. It was done in Crysis, and with LOD at -2, it looked sharp with SSAA.The Wasrad - Wednesday, September 23, 2009 - link
Why are you using 4 gigs of ram with a 920?Do you understand how DDR3 memory works?
Ryan Smith - Wednesday, September 23, 2009 - link
Error when writing the chart. It has been corrected.Sc4freak - Wednesday, September 23, 2009 - link
Do you? The fact that the i7 920 works best in a triple-channel configuration has nothing to do with the fact that it uses DDR3.chizow - Wednesday, September 23, 2009 - link
Agreed and to add to that, the fact the third channel means very little when it comes to actual gaming performance makes it even less signficant. As compared to Lynnfield clock for clock, which is only dual channel:http://www.anandtech.com/cpuchipsets/showdoc.aspx?...">http://www.anandtech.com/cpuchipsets/showdoc.aspx?...
Von Matrices - Wednesday, September 23, 2009 - link
Could someone enlighten me as to why the 4870 X2 could be faster than the 5870 in some situations? It was noted it the article but never really explained. They have the same number of SP's, and one would expect crossfire scaling to be detrimental to the 4870 X2"s performance. Would this be indicative of the 5870 being starved for memory bandwidth in these situations or something else?Dobs - Wednesday, September 23, 2009 - link
4870x2 has 2Gb of DDR55870 only has 1 until the 2Gb edition comes out :)
nafhan - Wednesday, September 23, 2009 - link
Doesn't using dual GPU's effectively halve the onboard memory, as significant portions of the textures, etc. need to be duplicated? So, the 4870x2 has a memory disadvantage by requiring 2x memory to accomplish the same thing.chizow - Wednesday, September 23, 2009 - link
Right, with an X2 each GPU has a copy of the same frame buffer, so the total memory onboard is effectively halved. A 2GB frame buffer with 2 GPU is two of the same 1GB frame buffer mirrored on each.With the 5870 essentially being 2xRV790 on one chip, in order to accomplish the same frame rates on the same sized 1GB frame buffer, you would expect to need additional bandwidth to facilitate the transfers to and from the frame buffer and GPU.
chizow - Wednesday, September 23, 2009 - link
Ya he mentions bandwidth being a potential issue preventing the 5870 from mirroring the 4870X2's results.It could also be that the 5870's scheduler/dispatch processor aren't as efficient at extracting performance as driver forced AFR. Seems pretty incredible, seeing as physically doubling GPU transistors on a single die has always been traditionally better than multi-GPU scaling.
Similarly, it could be a CPU limitation where CF/SLI benefit more from multi-threaded driver performance, whereas a single GPU would be limited to a single fast thread or core's performance. We saw this a bit as well last year with the GT200s compared to G92s in SLI.