The NVIDIA GeForce GTX 970 Review: Featuring EVGA
by Ryan Smith on September 26, 2014 10:00 AM ESTSynthetics
As always we’ll also take a quick look at synthetic performance. These tests mainly serve as a canary for finding important architectural and configuration changes.
Tessellation performance has scaled very closely with the change in SMMs and clock speeds, just as we would expect here.
Texel throughput has also taken a hit in accordance with the loss of SMMs and clock speed. Based on gaming performance the GTX 970 doesn’t appear to be too badly handicapped here, but it definitely doesn’t have much in the way of texel throughput to spare.
Pixel throughput on the other hand ends up being extremely odd and not at all what we were expecting. The GTX 970 takes an incredible dive here, with its pixel fillrate dropping by 26%. At a high level this test is bounded by memory bandwidth and ROP throughput, and both of these factors should be identical between GTX 980 and GTX 970. Instead we see GTX 970 lose more performance than should theoretically be possible, as the 26% drop is more than the accumulated difference between the clock speed and SMM differences.
At this point we’re still trying to figure out exactly what’s going on. We have no other evidence that there’s a difference in ROP throughput or memory bandwidth between the GTX 980 and GTX 970 so it is not clear to us where the difference lies. One possibility is that this is somehow bottlenecked at the Raster Engine level – where each of the four engines accounts for 25% of the work – but the pigeonhole principle means that NVIDIA can’t disable a GPC since at least 1 SMM must be active in each GPC partition. This matter will require further research.
Facebook
Twitter
RSS
155 Comments
View All Comments
JarredWalton - Friday, September 26, 2014 - link
There are plenty of older (but still decent) PSUs that only have 6-pin PEG connectors, and 6-pin to 8-pin adapters are never a good idea IMO.cobalt42 - Friday, September 26, 2014 - link
I believe at least one released card does use a single 8-pin connector.jmke - Tuesday, September 30, 2014 - link
Asus strix 970 DC2OC uses a single 8-pin connectorHrel - Friday, September 26, 2014 - link
Remember you guys saying Nvidia is working from the low end up, no longer top down. Well, they should start releasing cards that way. I don't care about your overpriced space heaters, I care about the cards between $100 and $200. Release those first!cobalt42 - Friday, September 26, 2014 - link
They did -- the 750 and 750Ti are the first generation Maxwell cards, released earlier this year. They didn't break new ground in price/performance, but they did in price/watt.anandreader106 - Friday, September 26, 2014 - link
You mean performance/watt.Houdani - Friday, September 26, 2014 - link
I think it's finally time to upgrade my vintage 460 to one of these 970's. I don't plan on upgrading my ancient i7 930 (Nahelem, Bloomfield) just yet.I think I can eek more life out of my rig by bumping the GPU and keeping the rest of the innards the same for another year or two. I'm just surprised that I was able to get 4 good years out of that little 460 (mated with a 1080P monitor).
CaptainSassy - Friday, September 26, 2014 - link
Exactly the same rig and monitir but i will continue sitting on it :D 329$ is still pricy, i'll wait for true 200$ championwetwareinterface - Saturday, September 27, 2014 - link
there won't be a $200 champion. the bang for buck cards are in the $300 and above price tier right now and for the foreseeable future. it was the r9 290 and now it's the 970. the next one i predict will be the 290's replacement it will have the texture compression of the 285 and be their second down part.just4U - Sunday, September 28, 2014 - link
The $200 champion appears to really be the 280 atm for those sitting on older cards.. I'd like to say the 285 but it's higher up in the price bracket. The 960 should be interesting though when it comes out but I doubt it will be a $200 card.. Looking at the 970.. Im guessing it will sit at the $250 price point.