Merge branch 'release_v2.8' of...

Merge branch 'release_v2.8' of http://10.16.6.30/dcutoolkit/deeplearing/TransformerEngine into release_v2.8

transformer_engine/common/gemm/cublaslt_gemm.cu

@@ -1395,7 +1395,7 @@ void nvte_grouped_gemm(const NVTETensor *A, const NVTETensor *B, NVTETensor *D,
     NVTE_ERROR("MOE nvte_grouped_gemm not surpport bias or gelu.");
   }
 
-  hipblaslt_goupedgemm(inputA, inputB, outputD, m, n, k, b,
+  hipblaslt_groupedgemm(inputA, inputB, outputD, m, n, k, b,
                       (transa) ? HIPBLAS_OP_T : HIPBLAS_OP_N,
                       (transb) ? HIPBLAS_OP_T : HIPBLAS_OP_N, 
                       wspace->data.dptr, wspace->data.shape[0],

transformer_engine/common/gemm/rocm_gemm.cu

@@ -1834,7 +1834,7 @@ class d_userArgsManager {
 static userArgsManager UAManager;
 static d_userArgsManager d_UAManager;
 
-void hipblaslt_goupedgemm(std::vector<const Tensor*>& inputA, std::vector<const Tensor*>& inputB,
+void hipblaslt_groupedgemm(std::vector<const Tensor*>& inputA, std::vector<const Tensor*>& inputB,
                           std::vector<Tensor*>& outputD, std::vector<int64_t>& m,
                           std::vector<int64_t>& n, std::vector<int64_t>& k, std::vector<int64_t>& b,
                           hipblasOperation_t transa, hipblasOperation_t transb, void* workspace,