NVIDIA GeForce GTX 690 Review: Ultra Expensive, Ultra Rare, Ultra Fast
by Ryan Smith on May 3, 2012 9:00 AM ESTCompute Performance
For our look at compute performance this is going to be a brief look. Our OpenGL AES and DirectCompute Fluid Simulation benchmarks simply don’t scale with multiple GPUs, so we’ll skip though (though the data is still available in Bench).
Our first compute benchmark comes from Civilization V, which uses DirectCompute to decompress textures on the fly. Civ V includes a sub-benchmark that exclusively tests the speed of their texture decompression algorithm by repeatedly decompressing the textures required for one of the game’s leader scenes. Note that this is a DX11 DirectCompute benchmark.
Given the nature of the benchmark, it’s not surprising that we see a performance regression here with some setups. The nature of this benchmark is that it doesn’t split across multiple GPUs well, though that doesn’t stop AMD and NVIDIA from tying. This doesn’t impact real game performance as we’ve seen, but it’s a good reminder of the potential pitfalls of multi-GPU configurations. Though AMD does deserve some credit here for gaining on their single GPU performance, pushing their lead even higher.
Our other compute benchmark is SmallLuxGPU, the GPU ray tracing branch of the open source LuxRender renderer. We’re now using a development build from the version 2.0 branch, and we’ve moved on to a more complex scene that hopefully will provide a greater challenge to our GPUs.
Unlike the Civ V compute benchmark, SLG scales very well with multiple GPUs, nearly doubling in performance. Unfortunately for NVIDIA GK104 shows its colors here as a compute-weak GPU, and even with two of them we’re nowhere close to one 7970, let alone the monster that is two. If you’re looking at doing serious GPGPU compute work, you should be looking at Fermi, Tahiti, or the future Big Kepler.
Facebook
Twitter
RSS
200 Comments
View All Comments
InsaneScientist - Sunday, May 6, 2012 - link
Or don't...It's 2 days later, and you've been active in the comments up through today. Why'd you ignore this one, Cerise?
CeriseCogburn - Sunday, May 6, 2012 - link
Because you idiots aren't worth the time and last review the same silverblue stalker demanded the links to prove my points and he got them, and then never replied.It's clear what providing proof does for you people, look at the sudden 100% ownership of 1920x1200 monitors..
ROFL
If you want me to waste my time, show a single bit of truth telling on my point on the first page.
Let's see if you pass the test.
I'll wait for your reply - you've got a week or so.
KompuKare - Thursday, May 3, 2012 - link
It is indeed sad. AMD comes up with really good hardware features like eyefinity but then never polishes up the drivers properly. Looking some of crossfire results is sad too: in Crysis and BF3 CF scalling is better than SLI (unsure but I think the trifire and quadfire results for those games are even more in AMD's favour), but in Skyrim it seems that CF is totally broken.Of course compared to Intel, AMD's drivers are near perfect but with a bit more work they could be better than Nvidia's too rather than being mostly at 95% or so.
Tellingly, JHH did once say that Nvidia were a software company which was a strange thing for a hardware manufacturer to say. But this also seems to mean that they forgotten the most basic primary thing which all chip designers should know: how to design hardware which works. Yes I'm talking about bumpgate.
See despite all I said about AMD's drivers, I will never buy Nvidia hardware again after my personal experience of their poor QA. My 8800GT, my brother's 8800GT, this 8400M MXM I had, plus number of laptops plus one nForce motherboard: they all had one thing in common, poorly made chips made by BigGreen and they all died way before they were obsolete.
Oh, and as pointed out in the Anand VC&G forums earlier today:
"Well, Nvidia has the title of the worst driver bug in history at this point-
http://www.zdnet.com/blog/hardware/w...hics-card/7... "
killing cards with a driver is a record.
Filiprino - Thursday, May 3, 2012 - link
Yep, that's true. They killed cards with a driver. They should implement hardware auto shutdown, like CPUs. As for the nForce, I had one motherboard, the best nForce they made: nForce 2 for AMD Athlon. The rest of mobo chipsets were bullshit, including nForce 680.The QA I don't think is NVIDIA's fault but videocard manufacturers.
KompuKare - Thursday, May 3, 2012 - link
No, 100% Nvidia's fault. Although maybe QA isn't the right word. I was referring to Nvidia using the wrong solder underfil for a few million chips (the exact number is unknown): they were mainly mobile parts and Nvidia had to put $250 million aside to settle a class action.
http://en.wikipedia.org/wiki/GeForce_8_Series#Prob...
Although that wiki article is rather lenient towards Nvidia since that bit about fan speeds is red herring: more accurately it was Nvidia which spec'ed their chips to a certain temperature and designs which run way below that will have put less stress on the solder but to say it was poor OEM and AIB design which lead to the problem is not correct. Anyway, the proper expose was by Charlie D. in the Inquirer and later SemiAccurate
CeriseCogburn - Friday, May 4, 2012 - link
But in fact it was a bad heatsink design, thank HP, and view the thousands of heatsink repairs, including the "add a copper penny" method to reduce the giant gap between the HS and the NV chip.Charlie was wrong, a liar, again, as usual.
KompuKare - Friday, May 4, 2012 - link
Don't be silly. While HP's DV6000s were the most notorious failures and that was due to HP's poorly designed heatsink / cooling bumpgate also saw Dells, Apples and others:http://www.electronista.com/articles/10/09/29/suit...
http://www.nvidiadefect.com/nvidia-settlement-t874...
The problem was real, continues to be real and also affects G92 desktop parts and certain nForce chipsets like the 7150.
Yes, the penny shim trick will fix it for a while but if you actually were to read up on technicians forums who fix laptops, that plus reflows are only a temporary fix because the actual chips are flawed. Re-balling with new, better solder is a better solution but not many offer those fixes since it involves 100s of tiny solder balls per chip.
Before blindly leaping to Nvidia's defence like a fanboy, please do some research!
CeriseCogburn - Saturday, May 5, 2012 - link
Before blindly taking the big lie from years ago repeated above to attack nvidia for no reason at all other than all you have is years old misinformation, then wail on about it, while telling someone else some more lies about it, check your own immense bias and lack of knowledge, since I had to point out the truth for you to find, and you forgot DV9000, dv2000 and dell systems with poor HS design, let alone apple amd console video chip failings, and the fact that payment was made and restitution was delivered, which you also did not mention, because of your fanboy problems, obviously in amd's favor.Ashkal - Thursday, May 3, 2012 - link
In price comparison in Final words you are not referring with AMD products. I think AMD is better in price performance ratio.prophet001 - Thursday, May 3, 2012 - link
I agree