Do not enable gpu-gpu direct transfer, it seems to just slow down cuda ATM...

src/drivers/cuda/driver_cuda.c

@@ -197,7 +197,7 @@ static void init_context(unsigned devid)
 	starpu_cuda_set_device(devid);
 
 #ifdef HAVE_CUDA_MEMCPY_PEER
-	if (starpu_get_env_number("STARPU_DISABLE_CUDA_GPU_GPU_DIRECT") <= 0) {
+	if (starpu_get_env_number("STARPU_DISABLE_CUDA_GPU_GPU_DIRECT") == 0) {
 		int nworkers = starpu_worker_get_count();
 		for (workerid = 0; workerid < nworkers; workerid++) {
 			struct _starpu_worker *worker = _starpu_get_worker_struct(workerid);
@@ -207,7 +207,7 @@ static void init_context(unsigned devid)
 				if (!cures && can) {
 					cures = cudaDeviceEnablePeerAccess(worker->devid, 0);
 					if (!cures)
-						_STARPU_DEBUG("GPU-Direct %d -> %d\n", worker->devid, devid);
+						_STARPU_DEBUG("Enabled GPU-Direct %d -> %d\n", worker->devid, devid);
 				}
 			}
 		}