Synthetics

As always we’ll also take a quick look at synthetic performance. The 290X shouldn’t pack any great surprises here since it’s still GCN, and as such bound to the same general rules for efficiency, but we do have the additional geometry processors and additional ROPs to occupy our attention.

Right off the bat then, the TessMark results are something of a head scratcher. Whereas NVIDIA’s performance here has consistently scaled well with the number of SMXes, AMD’s seeing minimal scaling from those additional geometry processors on Hawaii/290X. Clearly Tessmark is striking another bottleneck on 290X beyond simple geometry throughput, though it’s not absolutely clear what that bottleneck is.

This is a tessellation-heavy benchmark as opposed to a simple massive geometry bencehmark, so we may be seeing a tessellation bottleneck rather than a geometry bottleneck, as tessellation requires its own set of heavy lifting to generate the necessary control points. The 12% performance gain is much closer to the 11% memory bandwidth gain than anything else, so it may be that the 280X and 290X are having to go off-chip to store tessellation data (we are after all using a rather extreme factor), in which case it’s a memory bandwidth bottleneck. Real world geometry performance will undoubtedly be better than this – thankfully for AMD this is the pathological tessellation case – but it does serve of a reminder of how much more tessellation performance NVIDIA is able to wring out of Kepler. Though the nearly 8x increase in tessellation performance since 5870 shows that AMD has at least gone a long way in 4 years, and considering the performance in our tessellation enabled games AMD doesn’t seem to be hurting for tessellation performance in the real world right now.

Moving on, we have our 3DMark Vantage texture and pixel fillrate tests, which present our cards with massive amounts of texturing and color blending work. These aren’t results we suggest comparing across different vendors, but they’re good for tracking improvements and changes within a single product family.

Looking first at texturing performance, we can see that texturing performance is essentially scaling 1:1 with what the theoretical numbers say it should. 36% better texturing performance over 280X is exactly in line with the increased number of texture units versus 280X, at the very least proving that 290X isn’t having any trouble feeding the increased number of texture units in this scenario.

Meanwhile for our pixel fill rates the results are a bit more in the middle, reflecting the fact that this test is a mix of ROP bottlenecking and memory bandwidth bottlenecking. Remember, AMD doubled the ROPs versus 280X, but only gave it 11% more memory bandwidth. As a result the ROPs’ ability to perform is going to depend in part on how well color compression works and what can be recycled in the L2 cache, as anything else means a trip to the VRAM and running into those lesser memory bandwidth gains. Though the 290X does get something of a secondary benefit here, which is that unlike the 280X it doesn’t have to go through a memory crossbar and any inefficiencies/overhead it may add, since the number of ROPs and memory controllers is perfectly aligned on Hawaii.

GRID 2 Compute
POST A COMMENT

396 Comments

View All Comments

  • Sandcat - Friday, October 25, 2013 - link

    That depends on what you define as 'acceptable frame rates'. Yeah, you do need a $500 card if you have a high refresh rate monitor and use it for 3d games, or just improved smoothness in non-3d games. A single 780 with my brothers' 144hz Asus monitor is required to get ~90 fps (i7-930 @ 4.0) in BF3 on Ultra with MSAA.

    The 290x almost requires liduid...the noise is offensive. Kudos to those with the equipment, but really, AMD cheaped out on the cooler in order to hit the price point. Good move, imho, but too loud for me.
    Reply
  • hoboville - Thursday, October 24, 2013 - link

    Yup, and it's hot. It will be worth buying once the manufacturers can add their own coolers and heat pipes.

    AMD has always been slower at lower res, but better in the 3x1080p to 6x1080p arena. They have always aimed for high-bandwidth memory, which is always performs better at high res. This is good for you as a buyer because it means you'll get better scaling at high res. It's essentially forward-looking tech, which is good for those who will be upgrading monitors in the new few years when 1440p IPS starts to be more affordable. At low res the bottleneck isn't RAM, but computer power. Regardless, buying a Titan / 780 / 290X for anything less than 1440p is silly, you'll be way past the 60-70 fps human eye limit anyway.
    Reply
  • eddieveenstra - Sunday, October 27, 2013 - link

    Maybe 60-70fps is the limit. but at 120Hz 60FPS will give noticable lag. 75 is about the minimum. That or i'm having eagle eyes. The 780gtx still dips in the low framerates at 120Hz (1920x1080). So the whole debate about titan or 780 being overkill @1080P is just nonsense. (780gtx 120Hz gamer here) Reply
  • hoboville - Sunday, October 27, 2013 - link

    That really depends a lot on your monitor. When they talked about Gsync and frame lag and smoothness, they mentioned when FPS doesn't exactly match the refresh rate you get latency and bad frame timing. That you have this problem with a 120 Hz monitor is no surprise as at anything less than 120 FPS you'll see some form of stuttering. When we talk about FPS > refresh rate then you won't notice this. At home I use a 2048x1152 @ 60 Hz and beyond 60 FPS all the extra frames are dropped, where as in your case you'll have some frames "hang" when you are getting less than 120 FPS, because the frames have to "sit" on the screen for an interval until the next one is displayed. This appears to be stuttering, and you need to get a higher FPS from the game in order for the frame delivery to appear smoother. This is because apparent delay decreases as a ratio of [delivered frames (FPS) / monitor refresh speed]. Once the ratio is small enough, you can no longer detect apparent delay. In essence 120 Hz was a bad idea, unless you get Gsync (which means a new monitor).

    Get a good 1440p IPS at 60 Hz and you won't have that problem, and the image fidelity will make you wonder why you ever bought a monitor with 56% of 1440p pixels in the first place...
    Reply
  • eddieveenstra - Sunday, October 27, 2013 - link

    To be honnest. I would never think about going back to 60Hz. I love 120Hz but don't know a thing about IPS monitors. Thanks for the response....

    Just checked it and that sounds good. When becoming more affordable i will start thinking about that. Seems like the IPS monitors are better with colors and have less blur@60Hz than TN. link:http://en.wikipedia.org/wiki/IPS_panel
    Reply
  • Spunjji - Friday, October 25, 2013 - link

    Step 1) Take data irrespective of different collection methods.

    Step 2) Perform average of data.

    Step 3) Completely useless results!

    Congratulations, sir; you have broken Science.
    Reply
  • nutingut - Saturday, October 26, 2013 - link

    But who cares if you can play at 90 vs 100 fps? Reply
  • MousE007 - Thursday, October 24, 2013 - link

    Very true, but remember, the only reason nvidia prices their cards where they are is because they could. (Eg Intel CPUs v AMD) Having said that, I truly welcome the competition as it makes it better for all of us, regardless of which side of the fence you sit. Reply
  • valkyrie743 - Thursday, October 24, 2013 - link

    the card runs at 95C and sucks power like no tomorrow. only only beats the 780 by a very little. does not overclock well.

    http://www.youtube.com/watch?v=-lZ3Z6Niir4
    and
    http://www.youtube.com/watch?v=3OHKWMgBhvA

    http://www.overclock3d.net/reviews/gpu_displays/am...

    i like his review. its pure honest and shows the facts. im not a nvidia fanboy nore am i a amd fanboy. but ill take nvidia right how over amd.

    i do like how this card is priced and the performance for the price. makes the titan not worth 1000 bucks (or the 850 bucks it goes used on forums) but as for the 780. if you get a non reference 780. it will be faster than the 290x and put out LESS heat and LESS noise. as well as use less power.

    plus gtx 780 TI is coming out in mid November which will probably cut the cost of the current 780 too 550 and and this card would be probably aorund 600 and beat this card even more.
    Reply
  • jljaynes - Friday, October 25, 2013 - link

    you say the review sticks with the facts - he starts off talking about how ugly the card is so it needs to beat a titan. and then the next sentence he says the R9-290X will cost $699.

    he sure seems to stick with the facts.
    Reply

Log in

Don't have an account? Sign up now