Description Hi I am tring to update my code from TRT 7.1 to TRT 7.2. I found add_matrix_multiply give right answer on TensorRT 7.1.3.4,but fail on TensorRT 7.2.1.6 with error log: [TensorRT] INTERNAL ERROR: Assertion failed: cublasStatus == CUBLAS_STATUS_SUCCESS ../rtSafe/cublas/cublasLtWrapper.…

Hi TRT 7.2.1 switches to use cuBLASLt (previously it was cuBLAS). cuBLASLt is the defaulted choice for SM version >= 7.0. However,you may need CUDA-10.2 Patch 1 (Released Aug 26, 2020) to resolve some cuBLASLt issues. Another option is to use the new TacticSource API and disable cuBLASLt tactics if…

MatrixMultiply failed on TensorRT 7.2.1

AI & Data Science Deep Learning (Training & Inference) TensorRT

kel1 November 11, 2020, 4:15am 4

Getting same issue here. Tensorrt 7.2.1

1 Like

Topic		Replies	Views
[TensorRT] INTERNAL ERROR: Assertion failed: cublasStatus == CUBLAS_STATUS_SUCCESS ../rtSafe/cublas/cublasLtWrapper.cpp:279 TensorRT tensorrt	1	1985	March 29, 2021
Tensor RT7.2: Assertion `a.biasWeights.count() == a.widthB' failed TensorRT	2	304	October 12, 2021
[E] 2: [ltWrapper.cpp::setupHeuristic::349] Error Code 2: Internal Error (Assertion cublasStatus == CUBLAS_STATUS_SUCCESS failed. ) TensorRT tensorrt , cuda , ubuntu , python	1	1133	July 8, 2022
CUDA error when running matrixMulCUBLAS sample - Ubuntu 16.04 CUDA Setup and Installation	19	13492	May 4, 2018
TensorRT was linked against cuBLAS/cuBLAS LT 11.3.0 but loaded cuBLAS/cuBLAS LT 11.2.1 TensorRT tensorrt , cuda , cudnn	10	9472	October 12, 2021
Trying to run TensorRT project on AWS, getting error TensorRT	6	1878	May 19, 2021
Using IMatrixMultiplyLayer for batch matrix multiplication TensorRT tensorrt	4	1142	October 12, 2021
tensorRt installation fails w cuda9.1 / cuDNN 7.0.5 GPU-Accelerated Libraries	0	520	January 23, 2018
TensorRT was linked against cuBLAS/cuBLAS LT 10.2.3 but loaded cuBLAS/cuBLAS LT 10.2.2 CUDA Setup and Installation	9	3263	May 20, 2022
[TensorRT] ERROR: 2: [ltWrapper.cpp::setupHeuristic::327] TensorRT	4	2027	June 17, 2022

MatrixMultiply failed on TensorRT 7.2.1

Related topics