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

  • mr_tawan - Tuesday, November 05, 2013 - link

    AMD card may suffer from loud cooler. Let's just hope that the OEM versions would be shipped with quieter coolers. Reply
  • 1Angelreloaded - Monday, November 11, 2013 - link

    I have to be Honest here, it is beast, in fact the only thing in my mind holding this back is lack of feature sets compared to NVidia, namely PhysX, to me this is a bit of a deal breaker compared for 150$ more the 780 Ti gives me that with lower TDP/and sound profile, as we are only able to so much pull from 1 120W breaker without tripping it and modification for some people is a deal breaker due to wear they live and all. Honestly What I really need to see from a site is 4k gaming at max, 1600p/1200p/1080p benchmarks with single cards as well as SLI/Crossfire to see how they scale against each other. To be clear as well a benchmark using Skyrim Modded to the gills in texture resolutions as well to fully see how the VRAM might effect the cards in future games from this next Gen era, where the Consoles can manage a higher texture resolution natively now, and ultimately this will affect PC performance when the standard was 1-2k texture resolutions now becomes double to 4k or even in a select few up to 8k depth. With a native 64 bit architecture as well you will be able to draw more system RAM into the equation where Skyrim can use a max of 3.5 before it dies with Maxwell coming out and a shared memory pool with a single core microprocessor on the die itself with Gsync for smoothness we might see an over engineered GPU card capable of much much more than we thought, ATI as well has their own ideas which will progress, I have a large feeling Hawaii is actually a reject of sorts because they have to compete with Maxwell and engineer more into the cards themselves. Reply
  • marceloviana - Monday, November 25, 2013 - link

    I Just wondering why does this card came with 32Gb gddr5 and see only 4Gb. The PCB show 16 Elpida EDW2032BBBG (2G each). This amount of memory will help a lot in large scenes wit Vray-RT. Reply
  • Mat3 - Thursday, March 13, 2014 - link

    I don't get it. It's supposed to have 11 compute units per shader engine, making 44 on the entire chip. But the 2nd picture says each shader engine can only have up to 9 compute units....? Reply
  • Mat3 - Thursday, March 13, 2014 - link

    2nd picture on page three I mean. Reply
  • sanaris - Monday, April 14, 2014 - link

    Who cares? This card was never meant to compute something.

    It supposed to be "cheap but decent".
    Initially they made this ridiculous price, but now it is around 200-350 at ebay.
    For $200 it worth its price, because it can be used only to play games.
    Who wants to play games at medium quality (not the future ones), may prefer it.
    Reply

Log in

Don't have an account? Sign up now