CUDA recursion depth

Question

When using Dynamic Parallelism in CUDA, you can implement recursive algorithms like mergeSort. I have implemented it and my program don't work for inputs greater than blah.
My question is how many depth in the recursion tree the implementation can go? Is there any limitation? (My program is just fine for smaller inputs.)

stackoverflow.com/questions/14301903/…

void_ptr
– void_ptr

2015-01-03 17:06:28 +00:00
Commented Jan 3, 2015 at 17:06 — void_ptr
– void_ptr, Commented Jan 3, 2015 at 17:06

user14717 · Accepted Answer · 2015-01-03 20:27:03Z

4

From Professional CUDA C Programming:

The maximum nesting depth of dynamic parallelism is limited to 24, but in reality most kernels will be limited by the amount of memory required by the device runtime system at each new level . . .

answered Jan 3, 2015 at 20:27

user14717

5,3114 gold badges50 silver badges84 bronze badges

Sign up to request clarification or add additional context in comments.

2 Comments

Robert Crovella Over a year ago

This is documented in the programming guide as well.

user14717 Over a year ago

Seems like something that will need to go in the cudaDeviceProp struct eventually.

Collectives™ on Stack Overflow

CUDA recursion depth

1 Answer 1

2 Comments

Your Answer

Linked

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

2 Comments

Your Answer

Sign up or log in

Post as a guest

Linked

Related