- Joined
- Jul 20, 2002
I read somewhere that the GTX 970 used to have slower access to the last 500MiB or so of memory. I've also read that this isn't the first time Nvidia has done something like this (i.e. slowed memory access for some percentage of the total memory of a GPU). I can see why they might do it (so the GTX 970 doesn't perform like the GTX 980).
So, is the rumor true? Why did Nvidia hamstring the GTX 970? How did they manage to slow the access to certain addresses in the memory VRAM but not others? If a game does end up using the slowed memory does it make a significant difference in FPS or frame-time latency?
So, is the rumor true? Why did Nvidia hamstring the GTX 970? How did they manage to slow the access to certain addresses in the memory VRAM but not others? If a game does end up using the slowed memory does it make a significant difference in FPS or frame-time latency?