Some stack questions

Accelerated Computing CUDA CUDA Programming and Performance
1

I am trying to understand some aspects of the fermi architecture’s stack:

[list=1]

[*]What memory does the stack occupy? I’m assuming local memory, but I haven’t been able to find it spelled out anywhere on the forums or in the documentation; whatever the memory location is, is it cached?

[*]I read somewhere (release notes I think) that by default the stack is set to 1 kilobyte per thread, and that it can be resized by calling a function (cuCtxSetLimit) in the host code. Does anyone know what the maximum stack size is? Is there any documentation on this that I’ve missed?

Thank you to anyone who can offer some insight.

2

I am trying to understand some aspects of the fermi architecture’s stack:

[list=1]

[*]What memory does the stack occupy? I’m assuming local memory, but I haven’t been able to find it spelled out anywhere on the forums or in the documentation; whatever the memory location is, is it cached?

[*]I read somewhere (release notes I think) that by default the stack is set to 1 kilobyte per thread, and that it can be resized by calling a function (cuCtxSetLimit) in the host code. Does anyone know what the maximum stack size is? Is there any documentation on this that I’ve missed?

Thank you to anyone who can offer some insight.