The NVIDIA GeForce RTX 2060 6GB Founders Edition Review: Not Quite Mainstream
by Nate Oh on January 7, 2019 9:00 AM ESTCompute & Synthetics
Shifting gears, we'll look at the compute and synthetic aspects of the RTX 2060 (6GB). As a cut-down configuration of the TU106 GPU found in the RTX 2070, we should expect a similar progression in results.
Starting off with GEMM tests, the RTX 2070's tensor cores are pulled into action with half-precision matrix multiplication, though using binaries originally compiled for Volta. Because Turing is backwards compatible and in the same compute capability family as Volta (sm_75 compared to Volta's sm_70), the benchmark continues to work out-of-the-box, though without any particular Turing optimizations.
For Turing-based GeForce, FP32 accumulation on tensors is capped at half-speed, thus resulting in the observed halved performance. Aside from product segmentation, that higher-precision mode is primarily for deep learning training purposes, something that GeForce cards wouldn't be doing in games or consumer tasks.
Moving on, we have CompuBench 2.0, the latest iteration of Kishonti's GPU compute benchmark suite offers a wide array of different practical compute workloads, and we’ve decided to focus on level set segmentation, optical flow modeling, and N-Body physics simulations.
Moving on, we'll also look at single precision floating point performance with FAHBench, the official Folding @ Home benchmark. Folding @ Home is the popular Stanford-backed research and distributed computing initiative that has work distributed to millions of volunteer computers over the internet, each of which is responsible for a tiny slice of a protein folding simulation. FAHBench can test both single precision and double precision floating point performance, with single precision being the most useful metric for most consumer cards due to their low double precision performance.
Next is Geekbench 4's GPU compute suite. A multi-faceted test suite, Geekbench 4 runs seven different GPU sub-tests, ranging from face detection to FFTs, and then averages out their scores via their geometric mean. As a result Geekbench 4 isn't testing any one workload, but rather is an average of many different basic workloads.
We'll also take a quick look at tessellation performance.
Finally, for looking at texel and pixel fillrate, we have the Beyond3D Test Suite. This test offers a slew of additional tests – many of which use behind the scenes or in our earlier architectural analysis – but for now we’ll stick to simple pixel and texel fillrates.
Facebook
Twitter
RSS
134 Comments
View All Comments
Retycint - Tuesday, January 8, 2019 - link
So you think its performance level is not worthy of its price tag? In that case then you wouldn't find a single GPU on the market right now that can meet your price/performance standardspiroroadkill - Tuesday, January 8, 2019 - link
So.. NVIDIA's back to their old ways of giving cards too little RAM. Their cards have typically had much less RAM than the ones from AMD. So the 2060 is a tiny bit faster than a 1070 Ti... But loses 2GB of VRAM. Once you start using over 6GB VRAM, the 2060 is going to lose hard to the 1070 Ti. Price is also too high. A firm "meh".mkaibear - Wednesday, January 9, 2019 - link
How about identifying a couple of games in which the 1070Ti "wins hard" over the 2060?I think you'll find that the 6Gb of RAM is more than enough to run games at the resolutions that the 1070Ti or 2060 excel at (1080p fast or 1440p 60Hz)
Disagree? provide me with some evidence!
prateekprakash - Tuesday, January 8, 2019 - link
Maybe AIB partners should offer one custom rtx 2060 with 12gb vram by using 2gb gddr6 chips, that may be helpful down the road...evernessince - Tuesday, January 8, 2019 - link
So who's going to bet this card doesn't launch at MSRP like every other Nvidia card this generation? You might as well throw away any price/performance comparison done in this article as well all know there is going to be Nvidia tax thrown on top of that price tag.If this is Nvidia's new mainstream card, I hate to see the price tag on their "budget" $200 2050.
catavalon21 - Thursday, March 7, 2019 - link
I am surprised that the answer to your question is "well, actually, many are" available at or about MSRP. The one thing I'm puzzle by, is even though the article states NVIDIA will sell their own cards (as they have and do for many other models), I haven't seen one listed on NVIDIA's site (other than linking to AIB cards) since day one.catavalon21 - Thursday, March 7, 2019 - link
*puzzled* - need more coffeeoverseer - Tuesday, January 8, 2019 - link
My sense tells me that RTX 2000 series will be a short-lived line, and NV shall come up with the next-gen (be it called 3000 series or not) on 7nm later this year. Why bother buying a half-node-like card now?sing_electric - Tuesday, January 8, 2019 - link
Well, people who are building now, for one. There's still a lot of people who were put off upgrades when mining shot prices through the roof; this is a new entry that's significantly faster than say, $200 cards.Having said that, I tend to agree with you: 7nm will probably offer significantly better performance, and if you care about raytracing, it's really the games that *started* development after the RTX came out that will show a real benefit (instead of seeing a reflection in one pond in one part of one map or something), and those games will be coming out around the time that the next-gen RTX will anyways.
richough3 - Tuesday, January 8, 2019 - link
The reality is that this card, spec wise is a replacement of the 1070, but to call it a 2070 would show small performance increases vs calling it 2060. With it being faster than a 1070 at $30 less MSRP, it's an okay upgrade from a 1060, although I would have rather had 8+ GB RAM. I would expect the 2050 TI to have similar specs as a 1060, although with just 4 GB RAM.