Your GeForce runs AI at half speed. This fixes it for 20/30/40 series cards. Up to 2× FP16 / BF16 / TF32 tensor throughput
C
gpu_unlock: take gfp_t in diag_dma_alloc()
randomizedcoder [email protected]500e48c7c2fd
diff --git a/kernel/gpu_unlock/unlock_610.c b/kernel/gpu_unlock/unlock_610.cindex 5979e5f..33a005b 100644--- a/kernel/gpu_unlock/unlock_610.c+++ b/kernel/gpu_unlock/unlock_610.c@@ -136,7 +136,7 @@ static u32 pci_identity(const struct pci_dev *pdev)}static void *diag_dma_alloc(struct unlock_transaction_work *work, const char *name,- size_t size, dma_addr_t *dma, unsigned int flags)+ size_t size, dma_addr_t *dma, gfp_t flags){void *memory = dma_alloc_coherent(&work->pdev->dev, size, dma, flags);if (!memory)