There have been some talk on some forums regarding the Nvidia GTX 970 memory allocation bug and Nvidia has today released an official statement regarding the issue.


According to earlier discussion regarding the GTX 970 memory allocation bug, which suggested that GTX 970 suffered from a design flaw and could not address more than 3.3GB of VRAM, as in higher memory usage, produces micro-stuttering. The issue was spotted in VRAM-intensive gaming scenarios where the GTX 980 did not suffer same issues. Overclock.net managed to even include an Nvidia engineer into the story.




According to Nvidia, the GTX 970 GPU is fully capable of addressing its 4GB of memory, but does it in a different and unusual way. Nvidia states that the GTX 970 is indeed equipped with 4GB of memory, however, it as a different configuration of SMs than the 980, and fewer crossbar resources to the memory system.

According to Nvidia :"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."

Nvidia has also provided some performance data that shows that there is no further difference in performance between GTX 970 and the GTX 980 when using less than 3.5GB of memory and more than 3.5GB, which suggest that those 0.5GB might be slower but still does the job well.




Of course, Nvidia still did not explain micro-stuttering that some users have been experiencing and hopefully will fix that with the next driver release.



Source: Techpowerup.com.