Question

I got information from CUDA Profiler. I am so confused why
Replays Instruction != Grobal memory replay + Local memory replay + Shared bank conflict replay?

See the following information I got from profiler:

Replayed Instructions(%): 81.60
Global memory replay(%): 21.80
Local memory replays(%): 0.00
Shared bank conflict replay(%): 0.00

Could you help me explain this? Is there any other case causing instruction replay?

Answer 1

Because The SM can replay instructions due to other factors, like different branching logic.

So I can assume that 60% of your code is being reissued due to branching and 20% due to global memory. Can you post a snippet ?

From the F1 Help menu of the Cuda 4.0 profiler:

Replayed Instructions (%) This gives the percentage of
instructions replayed during kernel execution. Replayed instructions
are the difference between the numbers of instructions that are
actually issued by the hardware to the number of instructions that are
to be executed by the kernel. Ideally this should be zero. This is
calculated as 100 * (instructions issued – instruction executed) /
instruction issued

Global memory replay (%) Percentage of replayed instructions
caused due to global memory accesses. This is calculated as 100 * (l1
global load miss) / instructions issued

Local memory replay (%) Percentage of replayed instructions caused
due to local memory accesses. This is calculated as 100 * (l1 local
load miss + l1 local store miss) / instructions issued

Shared bank conflict replay (%) Percentage of replayed
instructions caused due to shared memory bank conflicts. This is
calculated as 100 * (l1 shared conflict)/ instructions issued