Fix logging and test files which misspell "precision" (pytorch#146113)

Noticed this while working on something, decided to submit a quick fix.
Pull Request resolved: pytorch#146113
Approved by: https://github.com/drisspg

Author: danielvegamyhre

aten/src/ATen/native/cuda/Blas.cpp

@@ -1082,7 +1082,7 @@ _scaled_mm_out_cuda(const Tensor& mat1, const Tensor& mat2,
 #ifndef USE_ROCM
   // We are doing row-wise scaling
   if (scaling_choice == ScalingType::RowWise) {
-    TORCH_CHECK(out.dtype() == kBFloat16, "Only bf16 high precsion output types are supported for row-wise scaling.");
+    TORCH_CHECK(out.dtype() == kBFloat16, "Only bf16 high precision output types are supported for row-wise scaling.");
     at::cuda::detail::f8f8bf16_rowwise(
         mat1,
         mat2,

test/test_linalg.py

@@ -2230,7 +2230,7 @@ def test_norm_old(self, device):
         def gen_error_message(input_size, p, keepdim, dim=None):
             return f"norm failed for input size {input_size}, p={p}, keepdim={keepdim}, dim={dim}"
 
-        # 'nuc' norm uses SVD, and thus its precsion is much lower than other norms.
+        # 'nuc' norm uses SVD, and thus its precision is much lower than other norms.
         # test_svd takes @precisionOverride({torch.float: 1e-4, torch.cfloat: 2e-4}),
         # and here we are doing the same thing for nuc norm.
         class PrecisionContext: