Enable dequant+matmul 8bit path for Intel CPU and XPU (#1484)

* new matmul8bit

Signed-off-by: jiqing-feng <[email protected]>

* fix cxb

Signed-off-by: jiqing-feng <[email protected]>

---------

Signed-off-by: jiqing-feng <[email protected]>

bitsandbytes/autograd/_functions.py

@@ -563,6 +563,29 @@ def backward(ctx, grad_output):
         return grad_A, grad_B, None, grad_bias, None
 
 
+class MatMul8bitFp(torch.autograd.Function):
+    # For Intel CPU and XPU, the double quant has many unsafe operations which will breaks the finetune.
+    # We'd like to use dequant + matmul to run finetune currently.
+
+    @staticmethod
+    def forward(ctx, A, B, out=None, bias=None, state=MatmulLtState):
+        CB = B.data.to(A.dtype).mul_(state.SCB.unsqueeze(1).mul(1.0 / 127.0)).t()
+        output = torch.matmul(A, CB).to(A.dtype)
+        ctx.state = state
+        ctx.dtype_A = A.dtype
+        ctx.grad_shape = A.shape
+        return output
+
+    @staticmethod
+    def backward(ctx, grad_output):
+        state = ctx.state
+        B = state.CxB if state.CxB is not None else state.CB
+        CB = B.to(ctx.dtype_A).mul_(state.SCB.unsqueeze(1).mul(1.0 / 127.0))
+        grad_A = torch.matmul(grad_output, CB).view(ctx.grad_shape).to(ctx.dtype_A)
+
+        return grad_A, None, None, None, None
+
+
 def matmul(
     A: torch.Tensor,
     B: torch.Tensor,
@@ -574,6 +597,8 @@ def matmul(
     state = state or MatmulLtState()
     if threshold > 0.0:
         state.threshold = threshold
+    if A.device.type in ("cpu", "xpu") and state.is_training:
+        return MatMul8bitFp.apply(A, B, out, bias, state)
     return MatMul8bitLt.apply(A, B, out, bias, state)