Integrate BSR matrix in GPU2 gemm_gen using bsrgemm.

src/faust_linear_operator/GPU2/faust_linear_algebra_gpu.hpp

@@ -109,21 +109,44 @@ namespace Faust
 			const MatSparse<FPP, GPU2>* spB;
 			const MatDense<FPP, GPU2>* dsA;
 			const MatDense<FPP, GPU2>* dsB;
+			const MatBSR<FPP, GPU2>* bsrA;
+			const MatBSR<FPP, GPU2>* bsrB;
 			// downcast and call the proper function
 			spA = dynamic_cast<const Faust::MatSparse<FPP,GPU2>*>(&A);
 			if(! spA)
+			{
 				dsA = dynamic_cast<const Faust::MatDense<FPP,GPU2>*>(&A);
+				if (! dsA)
+					bsrA = dynamic_cast<const Faust::MatBSR<FPP, GPU2>*>(&A);
+			}
 			spB = dynamic_cast<const Faust::MatSparse<FPP,GPU2>*>(&B);
 			if(! spB)
+			{
 				dsB = dynamic_cast<const Faust::MatDense<FPP,GPU2>*>(&B);
+				if (! dsB)
+					bsrB = dynamic_cast<const Faust::MatBSR<FPP, GPU2>*>(&B);
+			}
 			if(spA && spB)
 				spgemm(*spA, MatDense<FPP, GPU2>(*spB), C, alpha, beta, typeA, typeB);
-			else if(spA)
+			else if(spA && dsB)
 				spgemm(*spA, *dsB, C, alpha, beta, typeA, typeB);
-			else if(spB)
+			else if(spB && dsA)
 				spgemm(*dsA, *spB, C, alpha, beta, typeA, typeB);
-			else
+			else if (dsA && dsB)
 				gemm(*dsA, *dsB, C, alpha, beta, typeA, typeB);
+			else if (bsrA && dsB)
+				bsrgemm(*bsrA, *dsB, C, alpha, beta, typeA, typeB);
+			else if (bsrA && spB)
+				bsrgemm(*bsrA, MatDense<FPP, GPU2>(*spB), C, alpha, beta, typeA, typeB);
+			else if (bsrA && bsrB)
+				bsrgemm(*bsrA, bsrB->to_dense(), C, alpha, beta, typeA, typeB); // TODO: consider also converting bsrB to MatDense, depending on the weight
+			else if (bsrB && dsA)
+				bsrgemm(*dsA, *bsrB, C, alpha, beta, typeA, typeB);
+			else if (bsrB && spA)
+				bsrgemm(MatDense<FPP, GPU2>(*spA), *bsrB, C, alpha, beta, typeA, typeB); // TODO: consider also converting bsrB to MatDense, depending on the weight // to test
+			else
+				throw std::runtime_error("Unsupported matrix type in faust_linear_algebra_gpu gemm_gen");
+
 		}
 
 	template<typename FPP>
@@ -141,7 +164,7 @@ namespace Faust
 				// transpose / adjoint the product to rely on other signature of bsrgemm (MatSparse B as lhs matrix -- i.e. A)
 				char nopA, nopB;
 				MatDense<FPP, GPU2> nA(A);
-				MatDense<FPP, GPU2> nB(B);
+				MatBSR<FPP, GPU2> nB(B);
 				if(opA == 'N' && opB == 'N')
 				{
 					nopA = 'T';
@@ -210,7 +233,7 @@ namespace Faust
 				}
 			}
 			else {
-				bsrgemm(MatDense<FPP, GPU2>(A), MatDense<FPP, GPU2>(B), C, alpha, beta, opA, opB);
+				spgemm(A, B.to_sparse(), C, alpha, beta, opA, opB);
 			}
 
 		}