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
Document GA106 support in README
jesvs28f71e926a28
diff --git a/README.md b/README.mdindex d69fdd1..cb658e5 100644--- a/README.md+++ b/README.md@@ -4,7 +4,7 @@ This Linux kernel module combines the CMP unlocker's V67/PLM technique with theThe benefit of this early kmod driver is that it does not require manual patching each time nvidia releases a new driver version.-The effect of this unlock is that it doubles the instruction issuing speed of FP/BF16 FMA into FP32 operations on GA102 and GA104 cards.+The effect of this unlock is that it doubles the instruction issuing speed of FP/BF16 FMA into FP32 operations on GA102, GA104 and GA106 cards.[3090 Ti dense 8192³ GEMM at 450 W, five-sample medians with FP32 accumulation](https://github.com/Zagitta/gpu-unlock/issues/1#issue-5905622687):@@ -100,14 +100,14 @@ There is no `softdep nvidia pre:` rule: kmod honours only the firstThe module does nothing unless the kernel command line opts in:```-gpu_unlock.targets=all # every unbound GA102/GA104 card+gpu_unlock.targets=all # every unbound GA102/GA104/GA106 cardgpu_unlock.targets=0000:21:00.0 # only these cardsgpu_unlock.targets=0000:21:00.0,0000:22:00.0 # comma-separated```Addresses are full PCI names as `lspci -D` prints them (hex digits in eithercase), independent of GPU numbering and display role. An entry that is not-a GA102/GA104 card, including `all` mixed with addresses, is logged and ignored.+a GA102/GA104/GA106 card, including `all` mixed with addresses, is logged and ignored.The cards are unlocked concurrently, about a second in all however manythere are, and each on its own: a card that is already bound or powereddown is skipped. Bus mastering that the board firmware left on (Supermicro