diff --git a/crates/cubecl-linalg/src/matmul/kernels/matmul/algorithm/cmma.rs b/crates/cubecl-linalg/src/matmul/kernels/matmul/algorithm/cmma.rs
index 75dbdb93..75ab9665 100644
--- a/crates/cubecl-linalg/src/matmul/kernels/matmul/algorithm/cmma.rs
+++ b/crates/cubecl-linalg/src/matmul/kernels/matmul/algorithm/cmma.rs
@@ -38,7 +38,7 @@ impl<EG: Numeric> base::Algorithm<EG> for Cmma<EG> {
         Self::EG,
         Self::ES,
         Self::GlobalMatmul,
-        batch::TransposedDispatch,
+        batch::SwizzleTransposedDispatch<2>,
     >;
 
     fn cube_count(problem: &MatmulProblem) -> CubeCount {
diff --git a/crates/cubecl-linalg/src/matmul/kernels/matmul/base.rs b/crates/cubecl-linalg/src/matmul/kernels/matmul/base.rs
index faae6a14..01d60d2b 100644
--- a/crates/cubecl-linalg/src/matmul/kernels/matmul/base.rs
+++ b/crates/cubecl-linalg/src/matmul/kernels/matmul/base.rs
@@ -153,8 +153,8 @@ pub(crate) fn matmul_cube_preparation<R: Runtime, EG: Numeric, D: Algorithm<EG>>
     let cube_count = D::cube_count(&problem);
 
     let advanced_config = AdvancedConfig {
-        lhs_tiling_order: matmul::components::stage::TilingOrderConfig::RowMajor,
-        rhs_tiling_order: matmul::components::stage::TilingOrderConfig::ColMajor,
+        lhs_tiling_order: matmul::components::stage::TilingOrderConfig::ColMajor,
+        rhs_tiling_order: matmul::components::stage::TilingOrderConfig::RowMajor,
         enforced_tile_layout: (None, None),
     };