results

README-CS521.md

@@ -3,6 +3,11 @@ To benchmark matrix multiply with CUDA, run:
 cargo bench -p juno_matmul --features=cuda
 ```
 
+To see what code is specific to the CS 521 project, run:
+```
+git diff main matmul_gpu_opt
+```
+
 Times for (512, 512) X (512, 512) (ms):
 Base: 65.554
 Larger blocks: 4.887