Graphics, Hardware, VR World

Nvidia Denies Any Design Flaws With The GTX 970 in VRAM Allocation Issue

After several users found that the GTX 970 was running into severe framerate issues after crossing 3.5GB VRAM in intensive games, Nvidia (NASDAQ:NVDA) has released a statement mentioning that the card itself is not at fault.

The issue arose out of the fact that the GTX 970 is different from the GTX 980, which uses a fully unlocked GM204 Maxwell GPU with four GPCs. Each GPC contains four Streaming Multiprocessor (SM) units, each in turn housing 128 CUDA cores. The GTX 980, then has 16 SMs and 2,048 CUDA cores. To differentiate the card from the high-end GTX 980, Nvidia disabled a few clusters in the GTX 970, which houses 13 SMs and 1,664 CUDA cores. Along with the decrease in tessellation units, Nvidia lowered the TDP of the card from 165W on the GTX 980 to 145W on the GTX 970.

In doing so, it has been found that the GTX 970 — which features 4GB GDDR5 memory (same as the GTX 980) — is unplayable during high VRAM usage. In games like Shadow of Mordor, Watch Dogs and other recent titles, gamers have found that while the GTX 980 was able to utilize 4GB VRAM, the GTX 970 failed to cross 3.5GB. In situations where the card does have to access the last 500 to 700MB of memory, users noted that the performance was taking a significant hit, with DRAM bandwidth going from 150GB/s to as low as 20GB/s:

Going beyond 3.5GB vram usage in games like Hitman Absolution, Call of Duty Advanced Warfare severely degrades the performance, as if the last 512mb is actually being swapped from the RAM.

In certain scenarios, such as when gaming at 5K resolutions with 8xMSAA, the GTX 970 does indeed seem to be utilizing all of its VRAM, but once again, the framerates were being affected.

Here’s the statement Nvidia provided to the press in full:

The GeForce GTX 970 is equipped with 4GB of dedicated graphics memory.  However the 970 has a different configuration of SMs than the 980, and fewer crossbar resources to the memory system. To optimally manage memory traffic in this configuration, we segment graphics memory into a 3.5GB section and a 0.5GB section.  The GPU has higher priority access to the 3.5GB section.  When a game needs less than 3.5GB of video memory per draw command then it will only access the first partition, and 3rd party applications that measure memory usage will report 3.5GB of memory in use on GTX 970, but may report more for GTX 980 if there is more memory used by other commands.  When a game requires more than 3.5GB of memory then we use both segments.

We understand there have been some questions about how the GTX 970 will perform when it accesses the 0.5GB memory segment.  The best way to test that is to look at game performance.  Compare a GTX 980 to a 970 on a game that uses less than 3.5GB.  Then turn up the settings so the game needs more than 3.5GB and compare 980 and 970 performance again.

Here’s an example of some performance data:

GTX980 GTX970
Shadows of Mordor
<3.5GB setting = 2688×1512 Very High 72fps 60fps
>3.5GB setting = 3456×1944 55fps (-24%) 45fps (-25%)
Battlefield 4
<3.5GB setting = 3840×2160 2xMSAA 36fps 30fps
>3.5GB setting = 3840×2160 135% res 19fps (-47%) 15fps (-50%)
Call of Duty: Advanced Warfare
<3.5GB setting = 3840×2160 FSMAA T2x
Supersampling off
82fps 71fps
>3.5GB setting = 3840×2160 FSMAA T2x
Supersampling on
48fps (-41%) 40fps (-44%)


On GTX 980, Shadows of Mordor drops about 24% on GTX 980 and 25% on GTX 970, a 1% difference.  On Battlefield 4, the drop is 47% on GTX 980 and 50% on GTX 970, a 3% difference.  On CoD: AW, the drop is 41% on GTX 980 and 44% on GTX 970, a 3% difference.  As you can see, there is very little change in the performance of the GTX 970 relative to GTX 980 on these games when it is using the 0.5GB segment.

While Nvidia’s statement focuses on the numbers, it does not highlight why several users were having issues when accessing more than 3.5GB VRAM. At this stage, it is unclear if this is an issue that can be solved by issuing new drivers, or whether the problems are due to a design flaw with the card itself. We’ll share further details as and when we have them.