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
AnnonymousCoward - Sunday, September 28, 2014 - link
So you're saying it doesn't matter to say "clocked at 7GHz" when the actual clock is 1.75GHz. Well it matters to me! What if the chip multiplies the core clock by 2 internally; should we then say the core clock is 2.2GHz instead of 1.1GHz?Black Obsidian - Monday, September 29, 2014 - link
And thus the reporting of EFFECTIVE clock speeds. Not everyone does (or can) understand the complexities of the underlying architecture.jtrdfw - Wednesday, September 30, 2015 - link
yes. heatsinks on memory are pretty much a scamMagickMan - Friday, September 26, 2014 - link
How about a 970 OC vs 290 OC comparison? I don't have a favored GPU, I just care about bang for buck.The_Assimilator - Friday, September 26, 2014 - link
What would be the point? 970 already equals or beats 290X, and you don't get much from overclocking Hawaii GPUs, apart from more heat.poohbear - Friday, September 26, 2014 - link
Thank u for including a full stable of previous gen video cards to compare it to! In particularly the 670 & 770! Gives us a better idea of how it performs!Tetracycloide - Friday, September 26, 2014 - link
Seconded. Still running a 6950 flashed to 6970 so having the stock 6970 as a point of reference made this the easiest buy ever. Roughly twice the performance with lower power, heat, and noise? Yes.roxamis - Monday, September 29, 2014 - link
I have the same card (Sapphire 6950 Dirt 3, unlocked to 6970) and one fan of the 2 died last week (hitting 90-100 deg C in games). So with that I went to see what i can replace it with and the 970 ticks all my marks.krazyfrog - Friday, September 26, 2014 - link
The price-to-performance ratio is strong with this one.Frenetic Pony - Friday, September 26, 2014 - link
Certainly, but considering AMD has implemented some of the same things I'd expect an equivalent price to performance ratio from AMD for their new cards. Cut down the 290's bus to 256bit, clock the ram to 7ghz, and with the bandwidth compression you'd get a cheaper card with the same performance.Still, nice all around to have choices, been looking at cards myself, and am going to build a system for my brother quite soon. Yay price wars!