AMD Radeon HD 7970 Review: 28nm And Graphics Core Next, Together As One
by Ryan Smith on December 22, 2011 12:00 AM EST- Posted in
- GPUs
- AMD
- Radeon
- ATI
- Radeon HD 7000
Battlefield 3
Its popularity aside, Battlefield 3 may be the most interesting game in our benchmark suite for a single reason: it’s the first AAA DX10+ game. It’s been 5 years since the launch of the first DX10 GPUs, and 3 whole process node shrinks later we’re finally to the point where games are using DX10’s functionality as a baseline rather than an addition. Not surprisingly BF3 is one of the best looking games in our suite, but as with past Battlefield games that beauty comes with a high performance cost
How to benchmark BF3 is a point of great contention. Our preference is to always stick to the scientific method, which means our tests need to be perfectly repeatable or very, very close. For that reason we’re using an on-rails section of the single player game, Thunder Run, to do our testing. This isn’t the most strenuous part of Battlefield 3 – multiplayer can get much worse – but it’s the most consistent part of the game. In general we’ve found that minimum framerates in multiplayer are about half of the average framerate in Thunder Run, so it’s important to frame your expectations accordingly.
With that out of the way, Battlefield 3 ends up being one of the worst games for the 7970 from a competitive standpoint. It always maintains a lead over the GTX 580, but the greatest lead is only 13% at 2560 without any MSAA, and everywhere else it’s 3-5%. Of course it goes without saying that realistically BF3 is only playable at 1920 (no MSAA) and below on any of the single-GPU cards in this lineup, so unfortunately for AMD it’s the 5% number that’s the most relevant.
Meanwhile compared to the 6970, the 7970’s performance gains are also a bit below average. 2560 and 1920 with MSAA are quite good at 30% and 34% respectively, but at 1920 without MSAA that’s only a 25% gain, which is one of the smaller gaps between the two cards throughout our entire test suite.
The big question of course is why are we only seeing such a limited lead from the 7970 here? BF3 implements a wide array of technologies so it’s hard to say for sure, but there is one thing we know they implement in the engine that only NVIDIA can use: Driver Command Lists, the same “secret sauce” that boosted NVIDIA’s Civilization V performance by so much last year. So it may be that NVIDIA’s DCL support is helping their performance here in BF3, much like it was in Civ V.
But in any case, this is probably the only benchmark that’s really under delivered for the 7970. 5% is still a performance improvement (and we’ll take it any day of the week), but this silences any reasonable hope of being able to use 1920 at Ultra settings with MSAA on a single-GPU card for the time being.
Facebook
Twitter
RSS
292 Comments
View All Comments
CeriseCogburn - Thursday, March 8, 2012 - link
The gentleman should get one of the high end cards, and run on the common 19x0 x 1xxx monitors that poor people have several of nowadays, then crank up a decent 3d game and get back to us.These cards are not as powerful as the naysayers think - there's a lot of SUCK left in them even in cheap setups, let alone high end everything X2 or X4.
Ph0b0s - Thursday, December 22, 2011 - link
This was one of the most interesting parts of the article. So the only thing that requires new hardware is "Target Independent Rasterization". What does that do, so we know how much we are missing out on from only having Directx 11 hardware.SlyNine - Thursday, December 22, 2011 - link
Second that.Sgt. Stinger - Thursday, December 22, 2011 - link
I would like to call attention to that the reviewer at www.sweclockers.com noticed a full 10° C temperature drop on full load after removing and re-mounting the cards cooler with a quality thermal paste. This is quite significant, and the card should be more quiet after doing this.Chloiber - Thursday, December 22, 2011 - link
Computerbase noticed the exact same thing. They said, that AMD even told them, that there was a problem with the thermal paste (that's why they tested it specifically).It's still as loud as a GTX590 with new thermal paste though... :>
Sgt. Stinger - Thursday, December 22, 2011 - link
Well, yes, unless you can use custom fan profiles. I hope it is possible at least via third party software.If not, the cheapest way to get a reference cooler quiet is to remove the shroud and original fan, and strap on a couple of 120mm fans :D
Ryan Smith - Thursday, December 22, 2011 - link
This is the first I've heard of this. I've pinged AMD, but if they believed it was a problem for us I expect we would have heard about it by now.Sgt. Stinger - Friday, December 23, 2011 - link
Well, maybe, but your load temps seem to correspond to sweclockers load temps as they were before re-mounting the cooler. (here's their temp graph, before: http://www.sweclockers.com/image/diagram/2621?k=3f... ) and after remounting they got 66° C instead.SlyNine - Friday, December 23, 2011 - link
Honestly I'd rather have these benchmarks now than have the benchmarks cut any shorter becuase he had to remount the fan.Worth looking in to moving forward though.
CeriseCogburn - Thursday, March 8, 2012 - link
Oh, so another amd defect in released cards problem... so did we ever get the same treatment for the 480 or the 470 or the 580 or the 570 ?NO !
Of course not silly rabbit, screw nvidia and excuse amd and blame anything and anyone but.....
Thank you this is hilarious...