[GitHub] [incubator-mxnet] TristonC edited a comment on issue #17855: TensorRT does not work on other GPUs except GPU0
TristonC edited a comment on issue #17855: TensorRT does not work on other GPUs except GPU0 URL: https://github.com/apache/incubator-mxnet/issues/17855#issuecomment-604091749 @Caenorst Looks like the TRTCreateState function in [src/operator/subgraph/tensorrt/tensorrt.cc](https://github.com/apache/incubator-mxnet/blob/56e79853ad5cf98baf84454eb595c7658bef6ee6/src/operator/subgraph/tensorrt/tensorrt.cc#L263) file not passing the ctx to the TRT. Add ```c++ CUDA_CALL(cudaSetDevice(ctx.dev_id)); ``` to the top of this function will fix this issue, but not sure it is a good fix. This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org With regards, Apache Git Services
[GitHub] [incubator-mxnet] TristonC edited a comment on issue #17855: TensorRT does not work on other GPUs except GPU0
TristonC edited a comment on issue #17855: TensorRT does not work on other GPUs except GPU0 URL: https://github.com/apache/incubator-mxnet/issues/17855#issuecomment-604039572 It seems the onnx-trt not setting the device properly. @SayHiRay If you accept a dirty workaround try replace ctx = mx.gpu(1) with these: ```python from ctypes import cdll, c_char_p libcudart = cdll.LoadLibrary('libcudart.so') libcudart.cudaGetErrorString.restype = c_char_p def cudaSetDevice(device_idx): ret = libcudart.cudaSetDevice(device_idx) if ret != 0: error_string = libcudart.cudaGetErrorString(ret) raise RuntimeError("cudaSetDevice: " + error_string) device_id = 1 cudaSetDevice(device_id) ctx = mx.gpu(device_id) ``` This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org With regards, Apache Git Services
[GitHub] [incubator-mxnet] TristonC edited a comment on issue #17855: TensorRT does not work on other GPUs except GPU0
TristonC edited a comment on issue #17855: TensorRT does not work on other GPUs except GPU0 URL: https://github.com/apache/incubator-mxnet/issues/17855#issuecomment-604039572 It seems the onnx-trt not setting the device properly. If you accept a dirty workaround try replace ctx = mx.gpu(1) with these: ```python from ctypes import cdll, c_char_p libcudart = cdll.LoadLibrary('libcudart.so') libcudart.cudaGetErrorString.restype = c_char_p def cudaSetDevice(device_idx): ret = libcudart.cudaSetDevice(device_idx) if ret != 0: error_string = libcudart.cudaGetErrorString(ret) raise RuntimeError("cudaSetDevice: " + error_string) device_id = 1 cudaSetDevice(device_id) ctx = mx.gpu(device_id) ``` This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org With regards, Apache Git Services
[GitHub] [incubator-mxnet] TristonC edited a comment on issue #17855: TensorRT does not work on other GPUs except GPU0
TristonC edited a comment on issue #17855: TensorRT does not work on other GPUs except GPU0 URL: https://github.com/apache/incubator-mxnet/issues/17855#issuecomment-604039572 It seems the TRT setting the device properly. If you accept a dirty workaround try replace ctx = mx.gpu(1) with these: ```python from ctypes import cdll, c_char_p libcudart = cdll.LoadLibrary('libcudart.so') libcudart.cudaGetErrorString.restype = c_char_p def cudaSetDevice(device_idx): ret = libcudart.cudaSetDevice(device_idx) if ret != 0: error_string = libcudart.cudaGetErrorString(ret) raise RuntimeError("cudaSetDevice: " + error_string) device_id = 1 cudaSetDevice(device_id) ctx = mx.gpu(device_id) ``` This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: us...@infra.apache.org With regards, Apache Git Services