NVIDIA’s GeForce GTX Titan Review, Part 2: Titan's Performance Unveiled
by Ryan Smith & Rahul Garg on February 21, 2013 9:00 AM ESTSynthetics
As always we’ll also take a quick look at synthetic performance to get a better look at Titan’s underpinnings. These tests are mostly for comparing cards from within a manufacturer, as opposed to directly comparing AMD and NVIDIA cards. We’ll start with 3DMark Vantage’s Pixel Fill test.
Pixel fill is a mix of a ROP test and a test to see if you have enough bandwidth to feed those ROPs. At the same time the smallest increase in theoretical performance for Titan over GTX 680 was in ROP performance, where a 50% increase in ROPs was met with a minor clockspeed reduction for a final increase in ROP performance of 25%.
The end result is that with gains of 28%, Titan’s lead over GTX 680 is just a hair more than its increase in theoretical ROP performance. Consequently at first glance it looks like Titan has enough memory and cache bandwidth to feed its 48 ROPs, which given where we’re at today with GDDR5 is good news as GDDR5 has very nearly run out of room.
Moving on, we have our 3DMark Vantage texture fillrate test, which does for texels and texture mapping units what the previous test does for ROPs.
Oddly enough, despite the fact that Titan’s texture performance improvements over GTX 680 were only on the range of 46%, here Titan is measured as having 62% more texturing performance. This may be how Titan is interplaying with its improved bandwidth, or it may be a case where some of the ancillary changes NVIDIA made to the texture paths for compute are somehow also beneficial to proper texturing performance.
Finally we’ll take a quick look at tessellation performance with TessMark.
Unsurprisingly, Titan is well ahead of anything else NVIDIA produces. At 49% faster it’s just a bit over the 46% theoretical performance improvement we would expect from the increased number of Polymorph Engines the extra 6 SMXes bring. Interestingly, as fast as GTX 580’s tessellation performance was, these results would indicate that Titan offers more than a generational jump in tessellation performance, nearly tripling GTX 580’s tessellation performance. Though at this time it’s not at all clear just what such tessellation performance is good for, as we seem to be reaching increasingly ridiculous levels.
Facebook
Twitter
RSS
337 Comments
View All Comments
vps - Thursday, February 21, 2013 - link
For compute benchmark you might want to take a look at FAHBenchFAHBench is the official Folding@Home GPU benchmark. It measures the compute performance of GPUs for Folding@Home.
http://proteneer.com/blog/?page_id=1671
Some reference scores are here:
http://foldingforum.org/viewtopic.php?f=38&t=2...
Ryan Smith - Thursday, February 21, 2013 - link
FAHBench is primarily an OpenCL benchmark (there's a CUDA path, but it's effectively on its way out). It's on our list, and is one of the things we couldn't run due to the fact that OpenCL is not currently working on Titan.Hrel - Thursday, February 21, 2013 - link
PowerDirector still uses CUDAatlr - Thursday, February 21, 2013 - link
Not sure if this helps. I found CLBenchmark results of a Titan versus a 7970 here.http://clbenchmark.com/compare.jsp?config_0=144702...
atlr - Thursday, February 21, 2013 - link
Ville Timonen posted results running his own code on a Tesla K20 versus the usual suspects. Might be helpful to folks considering options for GPGPU computation.http://wili.cc/blog/gpgpu-faceoff.html
chizow - Thursday, February 21, 2013 - link
That's the first thing that comes to mind now when I think of Nvidia, which is a shame because that name used to be synonymous with Awesome. That's gone and replaced with a disdain for ridiculous levels of usury with their last two high-end product launches. I'm not going to be disingenuous and claim I'm going AMD, because the fact of the matter is, Nvidia products are still in a class of it's own for my usage needs, but I will certainly not be spending as much on Nvidia parts as I used to.Kepler has basically set back Nvidia's product stack back by half a generation, but my price:performance metrics will stay the same. Nvidia has their ultra-premium "Xtreme Edition" GPU this round, but that only came about as a result of AMD's ridiculous pricing and overall lackluster performance of the 7970 for a "flagship" card. Either way, I think it will be difficult for Nvidia to sustain this price point, as expectations just got higher at that $1K range.
@Ryan: I'm disappointed you didn't write a harsher commentary on the fact Nvidia is now charging 2x for the same class of GPU, pricing that has stayed true since 2006. Even the G80 Ultra didn't approach this $1K mark. Given how many times Nvidia has had to backpedal and apologize about the Ultra's pricing, you would think they would learn from their mistakes. I guess not, I hope they are prepared to deal with the long-term ramifications, backlash, and loss of goodwill stemming from this pricing decision.
CeriseCogburn - Thursday, February 21, 2013 - link
7.1 billion transistors and 6G of ram.I for one am sick of you people constantly whining.
If we check the whine log from the ATI32 days you were doing it then, too.
It's all you people do. Every time, all the time.
chizow - Friday, February 22, 2013 - link
And all you do is post inane, barely intelligible nonsense in defense of Nvidia. If you check your "whine logs" you'll see I've done my fair share of defending Nvidia, but I can't and won't give them a pass for what they've done with Kepler. AMD started it for sure with the terribad price:performance of Tahiti, but Nvidia has taken it to a new level of greed.And for all the idiots who are going to reply "herr dueerr Nvidai need make money not a charity derrr", my advanced reply is that Nvidia has made money in all but 2-3 quarters since 2006 without selling a single $1K desktop GPU. In fact, they enjoyed record profits, margin and revenue on the back of a $250 GPU, the 8800GT in 2007-2008.
CeriseCogburn - Saturday, February 23, 2013 - link
Nope, you are the crying whining baby who says the same thing in 100 different posts here, and has no clue what "the economy" of "the world" has been doing for the past several years.Whatever.
Please go cry about all the other computer part prices that are doing the same thing.
CeriseCogburn - Saturday, February 23, 2013 - link
I mean you idiots want the same price for more perf over a DECADE. Meanwhile, the rest of the worlds pricing has DOUBLED.Now, computer prices used to drop across the board, but they just aren't doing it anymore, and IDIOTS like yourself continue on your stupid frikkin rants, ignoring the world itself, not to mention the LACK of production in that world for the GPU's you whine about. It's really funny how stupid you are.