I think this was mentioned in another forum thread. When you use those options, the cache usage modification is applied by ptxas. You won’t see the modified instructions when you look at the PTX output of nvcc because ptxas has not run yet. Compile a .cubin and use cuobjdump to see what instructions were actually generated after ptxas ran.
Related topics
| Topic | Replies | Views | Activity | |
|---|---|---|---|---|
| Setting cache control, when compiling into PTX with NVCC Xptxas -dlcm doesn't work | 4 | 1974 | November 7, 2011 | |
| Help me understand "-Xptxas -dlcm=cg" (take 2) | 1 | 7047 | November 24, 2010 | |
| Turn off L1 caching on Fermi GPUs via the driver API? | 2 | 697 | September 28, 2011 | |
| Compiling with non-caching loads | 0 | 983 | November 23, 2011 | |
| setting ptxas options Another simple question | 1 | 2955 | February 20, 2008 | |
| Disabling cache on Fermi architectures Try to disable L1 and L2 | 11 | 9361 | August 30, 2013 | |
| Cuda: How do I disable global memory caching? | 1 | 2143 | June 30, 2011 | |
| GMEM loads: caching vs. non-caching | 4 | 5332 | February 10, 2014 | |
| --ptxas-options=-v is not working | 1 | 1668 | April 30, 2011 | |
| Compiler -Xptxas flag has no effect | 6 | 4751 | December 19, 2011 |