Is GPU L1 cache enable in Jetson x1?

Hello!
I am running a code in Jetson X1. When i compile my code with the flag -Xptxas -dlcm=ca (enable L1 cache), the execution time of my application is the same as if i compile it with the flag -Xptxas -dlcm=cg (disabled L1 cache). In addition, when i profile my application with nvvp there are not metrics for L1 cache to profile. Is L1 cache enable in Jetson x1?
Thank you in advance!

Hi Sinay,

Not sure about nvvp but according to http://docs.nvidia.com/cuda/profiler-users-guide/index.html#metrics-reference-5x , “local_memory_overhead” is the only metric related to L1.

Hi,

The Maxwell and Pascal architecture combined the texture and L1 cache into a single unified cache.
All global, local, surface, and texture operations go through this cache.

You can profile it via this command:

nvprof -m tex_utilization ./vectorAdd

Read more: http://docs.nvidia.com/cuda/maxwell-tuning-guide/index.html#memory-throughput

Thank you very much both of you!