# \#pycuda **URL:** https://forums.developer.nvidia.com/tag/pycuda/87.md [Latest](https://forums.developer.nvidia.com/latest.md) · [Categories](https://forums.developer.nvidia.com/categories.md) · [Tags](https://forums.developer.nvidia.com/tags.md) --- ## [Problems with float4 (and other 4 element vectors) using PyCUDA](https://forums.developer.nvidia.com/t/problems-with-float4-and-other-4-element-vectors-using-pycuda/341566)
**Author:** [@mauricioandrade2002](https://forums.developer.nvidia.com/u/mauricioandrade2002)\ **Replies:** 1\ **Last updated:** [August 8, 2025, 3:43am UTC](https://forums.developer.nvidia.com/t/problems-with-float4-and-other-4-element-vectors-using-pycuda/341566 "2025-08-08T03:43:41Z")
I am trying to pass a float4 as argument to my cuda kernel (by value) using PyCUDA’s make\_float4(). But there seems to be some misalignment when the data is transferred to the kernel. If I read the output for an input (1… --- ## [Passing decoded frame to cuda kernel (cuda-python)](https://forums.developer.nvidia.com/t/passing-decoded-frame-to-cuda-kernel-cuda-python/338557)
**Author:** [@mrutuy](https://forums.developer.nvidia.com/u/mrutuy)\ **Replies:** 1\ **Last updated:** [July 22, 2025, 9:37am UTC](https://forums.developer.nvidia.com/t/passing-decoded-frame-to-cuda-kernel-cuda-python/338557 "2025-07-22T09:37:34Z")
New here. Trying pass decoded frame to a cuda Kernel and getting the error. Appreciate any suggestion. Decoding frame code: decoder = nvc.SimpleDecoder(enc\_file\_path=input\_video\_file, gpu\_id=cudaD… --- ## [DeepStream 7.1 CUDA Synchronization Issue on WSL2 with RTX 4060](https://forums.developer.nvidia.com/t/deepstream-7-1-cuda-synchronization-issue-on-wsl2-with-rtx-4060/333466)
**Author:** [@mertadgn](https://forums.developer.nvidia.com/u/mertadgn)\ **Replies:** 1\ **Last updated:** [May 16, 2025, 9:08am UTC](https://forums.developer.nvidia.com/t/deepstream-7-1-cuda-synchronization-issue-on-wsl2-with-rtx-4060/333466 "2025-05-16T09:08:19Z")
DeepStream 7.1 CUDA Synchronization Issue on WSL2 with RTX 4060 Hi everyone, I’m running into a persistent issue with my DeepStream 7.1 pipeline on WSL2 (Ubuntu 22.04) using an NVIDIA RTX 4060 GPU. Despite experimenting… --- ## [Illegal memory access Error](https://forums.developer.nvidia.com/t/illegal-memory-access-error/236761)
**Author:** [@user148610](https://forums.developer.nvidia.com/u/user148610)\ **Replies:** 1\ **Last updated:** [May 3, 2023, 7:40am UTC](https://forums.developer.nvidia.com/t/illegal-memory-access-error/236761 "2023-05-03T07:40:03Z")
Hello, I’m trying to do inference for a Convolutional Network on RTX 3060. When I ran my python script, I got a pyCUDA error: Illegal Memory Access. My test env is Ubuntu 22.04, and RTX 3060 GPU. I have tried to recompil… --- ## [Launching Conda environment + Python for Cuda profiling](https://forums.developer.nvidia.com/t/launching-conda-environment-python-for-cuda-profiling/324305)
**Author:** [@jcnelson2185](https://forums.developer.nvidia.com/u/jcnelson2185)\ **Replies:** 2\ **Last updated:** [March 21, 2025, 2:45am UTC](https://forums.developer.nvidia.com/t/launching-conda-environment-python-for-cuda-profiling/324305 "2025-03-21T02:45:22Z")
Hello, I am trying to profile my CUDA application with NSight Compute UI. While I can find examples of how to run a simple executable, I am now trying to use it with Python as PyCuda is what is calling my code. To furth… --- ## [Account disabled](https://forums.developer.nvidia.com/t/account-disabled/307535)
**Author:** [@joel101.alfred](https://forums.developer.nvidia.com/u/joel101.alfred)\ **Replies:** 0\ **Last updated:** [September 22, 2024, 10:01am UTC](https://forums.developer.nvidia.com/t/account-disabled/307535 "2024-09-22T10:01:01Z")
Hello, When I try to log in to https://learn.nvidia.com for a instructer-led event , the site either doesn’t load and when I use a different browser does, I get a message saying, “Your account has been disabled. If you … --- ## [使用pycuda运行项目时提示ImportError: DLL load failed while importing \_driver: 找不到指定的模块。](https://forums.developer.nvidia.com/t/pycuda-importerror-dll-load-failed-while-importing-driver/299566)
**Author:** [@1229837510](https://forums.developer.nvidia.com/u/1229837510)\ **Replies:** 1\ **Last updated:** [July 12, 2024, 7:47am UTC](https://forums.developer.nvidia.com/t/pycuda-importerror-dll-load-failed-while-importing-driver/299566 "2024-07-12T07:47:52Z")
File “D:\\新建文件夹\\yolov8\_9.24\\yolov8\_9.24\\ultralytics\\yolo\\v8\_init\_.py”, line 3, in from ultralytics.yolo.v8 import classify, detect, segment File “D:\\新建文件夹\\yolov8\_9.24\\yolov8\_9.24\\ultralytics\\yolo\\v8\\classify\_init\_.py”,… --- ## [Error during installation of pycuda from pip. Getting error while building wheel for pycuda](https://forums.developer.nvidia.com/t/error-during-installation-of-pycuda-from-pip-getting-error-while-building-wheel-for-pycuda/298067)
**Author:** [@hameed.ibrahim](https://forums.developer.nvidia.com/u/hameed.ibrahim)\ **Replies:** 2\ **Last updated:** [July 1, 2024, 10:09am UTC](https://forums.developer.nvidia.com/t/error-during-installation-of-pycuda-from-pip-getting-error-while-building-wheel-for-pycuda/298067 "2024-07-01T10:09:50Z")
Hi, I am new to Jetson orin nano. Installed cuda 12.5 and trying to install pycuda. Getting the below error when the pip trying to build the wheel for pycuda. Tried all the steps mentioned in the other threads and nothi… --- ## [Unable to Execute Forward and Backward Functions in Parallel Across Different CUDA Streams(pytorch1.9)](https://forums.developer.nvidia.com/t/unable-to-execute-forward-and-backward-functions-in-parallel-across-different-cuda-streams-pytorch1-9/297959)
**Author:** [@lssyes\_shuai](https://forums.developer.nvidia.com/u/lssyes_shuai)\ **Replies:** 3\ **Last updated:** [June 27, 2024, 6:10pm UTC](https://forums.developer.nvidia.com/t/unable-to-execute-forward-and-backward-functions-in-parallel-across-different-cuda-streams-pytorch1-9/297959 "2024-06-27T18:10:19Z")
Hello, I’m experiencing issues trying to execute my forward and backward propagation functions in parallel using PyTorch and CUDA streams. My code creates two different streams and attempts to execute the forward and ba… --- ## [How to install TensorRT?](https://forums.developer.nvidia.com/t/how-to-install-tensorrt/295745)
**Author:** [@redrishlol](https://forums.developer.nvidia.com/u/redrishlol)\ **Replies:** 3\ **Last updated:** [June 12, 2024, 11:14am UTC](https://forums.developer.nvidia.com/t/how-to-install-tensorrt/295745 "2024-06-12T11:14:04Z")
I have Jetpack 5.1.1 installed, and I am using a Nvidia Jetson AGX Orin 32GB H01. I want to use TensorRT to optimize and speed up YoloP, so I used the command sudo apt-get install tensorrt nvidia-tensorrt-dev python3-l… --- ## [Fail to install pycuda](https://forums.developer.nvidia.com/t/fail-to-install-pycuda/287681)
**Author:** [@1075471989](https://forums.developer.nvidia.com/u/1075471989)\ **Replies:** 1\ **Last updated:** [April 8, 2024, 5:24am UTC](https://forums.developer.nvidia.com/t/fail-to-install-pycuda/287681 "2024-04-08T05:24:01Z")
I want to install pycuda but get failed. My jetpack is 5.1.3. The cuda version is 11.4. I already added path: export PATH=/usr/local/cuda-11.4/bin${PATH:+:${PATH}} export LD\_LIBRARY\_PATH=/usr/local/cuda-11.4/lib64/{LD\_L… --- ## [Installing Pycuda for TensorRT Dorcker Image](https://forums.developer.nvidia.com/t/installing-pycuda-for-tensorrt-dorcker-image/283449)
**Author:** [@MDS\_Eddy](https://forums.developer.nvidia.com/u/MDS_Eddy)\ **Replies:** 17\ **Last updated:** [March 4, 2024, 8:04am UTC](https://forums.developer.nvidia.com/t/installing-pycuda-for-tensorrt-dorcker-image/283449 "2024-03-04T08:04:50Z")
Hello, We have to set docker environment on Jetson TX2. The important point is we want TenworRT(\>=8.2) and pycuda. I’ve checked pycuda can install on local as below: But it doesn’t work on docker that it is l4t-tens… --- ## [Error installing Pycuda in WSL2](https://forums.developer.nvidia.com/t/error-installing-pycuda-in-wsl2/279386)
**Author:** [@jefbutar](https://forums.developer.nvidia.com/u/jefbutar)\ **Replies:** 1\ **Last updated:** [February 6, 2024, 7:53am UTC](https://forums.developer.nvidia.com/t/error-installing-pycuda-in-wsl2/279386 "2024-02-06T07:53:33Z")
I find an error when trying to install pycuda on WSL2 (Windows 10) following the command: pip install pycuda the error log is huge but it closes with: ERROR: Failed building wheel for pycuda Failed to build pycuda ERR… --- ## [Help Improving Performance using cuSolver/cuSparse Routines](https://forums.developer.nvidia.com/t/help-improving-performance-using-cusolver-cusparse-routines/276250)
**Author:** [@joao.nizer](https://forums.developer.nvidia.com/u/joao.nizer)\ **Replies:** 0\ **Last updated:** [December 15, 2023, 5:05pm UTC](https://forums.developer.nvidia.com/t/help-improving-performance-using-cusolver-cusparse-routines/276250 "2023-12-15T17:05:39Z")
I wanted to report and ask for help when using CUDA cuSolver/cuSparse GPU routines that are slower than CPU versions (Python → Scipy Sparse Solvers). And, of course, ask for help if something is being done incorrectly i… --- ## [Segmentation Fault while using ROS melodic for Inferencing using TensorRT on Jetson Nano](https://forums.developer.nvidia.com/t/segmentation-fault-while-using-ros-melodic-for-inferencing-using-tensorrt-on-jetson-nano/274634)
**Author:** [@rushabh21rdp](https://forums.developer.nvidia.com/u/rushabh21rdp)\ **Replies:** 4\ **Last updated:** [December 2, 2023, 3:14am UTC](https://forums.developer.nvidia.com/t/segmentation-fault-while-using-ros-melodic-for-inferencing-using-tensorrt-on-jetson-nano/274634 "2023-12-02T03:14:34Z")
Hello everyone, I am currently working on a project using a Jetson Nano where I am trying to perform inference using TensorRT (Yolov5 object detection) in a ROS (Melodic) node. I have encountered a Bus Error(Core dumped… --- ## [Is it possible display directly from GPU memory to HDMI](https://forums.developer.nvidia.com/t/is-it-possible-display-directly-from-gpu-memory-to-hdmi/271375)
**Author:** [@zxl66467611](https://forums.developer.nvidia.com/u/zxl66467611)\ **Replies:** 5\ **Last updated:** [November 2, 2023, 4:08am UTC](https://forums.developer.nvidia.com/t/is-it-possible-display-directly-from-gpu-memory-to-hdmi/271375 "2023-11-02T04:08:29Z")
Hi, I wanna to render or transfer the deep learning model output (rgb image) from cuda (gpu memory) directly to a display using HDMI or DP cable without moving it to cpu memory? Actually I wanna transfer data via HDMI(DP… --- ## [Feature request to implement function pyds.get\_nvds\_buf\_surface\_gpu(hash(gst\_buffer) for Jetson devices](https://forums.developer.nvidia.com/t/feature-request-to-implement-function-pyds-get-nvds-buf-surface-gpu-hash-gst-buffer-for-jetson-devices/268223)
**Author:** [@prabal.malviya](https://forums.developer.nvidia.com/u/prabal.malviya)\ **Replies:** 3\ **Last updated:** [October 17, 2023, 5:37am UTC](https://forums.developer.nvidia.com/t/feature-request-to-implement-function-pyds-get-nvds-buf-surface-gpu-hash-gst-buffer-for-jetson-devices/268223 "2023-10-17T05:37:37Z")
Please provide complete information as applicable to your setup. • Hardware Platform (Jetson / GPU) : Jetson Orin Agx • DeepStream Version : Deepstream 6.2 • JetPack Version (valid for Jetson only) : 5.1-b147 • Tenso… --- ## [Jetson AGX Orin: Error running pycuda](https://forums.developer.nvidia.com/t/jetson-agx-orin-error-running-pycuda/269008)
**Author:** [@md.anowar](https://forums.developer.nvidia.com/u/md.anowar)\ **Replies:** 5\ **Last updated:** [October 16, 2023, 6:09am UTC](https://forums.developer.nvidia.com/t/jetson-agx-orin-error-running-pycuda/269008 "2023-10-16T06:09:33Z")
I am trying to run the following program on NVIDIA AGX Orin, Ubuntu 20.04: import NumPy import pycuda.autoinit import pycuda.driver as drive from pycuda.compiler import SourceModule SourceModule compiles C code for … --- ## [Problems with Pycuda on Jetson Agx Orin](https://forums.developer.nvidia.com/t/problems-with-pycuda-on-jetson-agx-orin/267175)
**Author:** [@marinkovicivan](https://forums.developer.nvidia.com/u/marinkovicivan)\ **Replies:** 9\ **Last updated:** [September 22, 2023, 5:21pm UTC](https://forums.developer.nvidia.com/t/problems-with-pycuda-on-jetson-agx-orin/267175 "2023-09-22T17:21:33Z")
Hello, I am having problems with PyCuda installation on Jetson Agx Orin 64gb Dev Kit. nvcc–version result: nvcc: NVIDIA (R) Cuda compiler driver Copyright (c) 2005-2022 NVIDIA Corporation Built on Sun\_Oct\_23\_22:16:0… --- ## [Segmentation fault when importing pycuda drivers](https://forums.developer.nvidia.com/t/segmentation-fault-when-importing-pycuda-drivers/258401)
**Author:** [@smaiah.sarah](https://forums.developer.nvidia.com/u/smaiah.sarah)\ **Replies:** 5\ **Last updated:** [July 18, 2023, 8:42am UTC](https://forums.developer.nvidia.com/t/segmentation-fault-when-importing-pycuda-drivers/258401 "2023-07-18T08:42:35Z")
Hello every one, I am trying to import pycuda drivers but i get a segmentation fault. When i tried to find the error using faulthandler i get this : $ python3 Python 3.6.9 (default, Mar 10 2023, 16:46:00) \[GCC 8.4.0\]… --- ## [Getting "Failed to build pycuda ERROR: Could not build wheels for pycuda" when trying to set up a Nvidia Jeston AGX](https://forums.developer.nvidia.com/t/getting-failed-to-build-pycuda-error-could-not-build-wheels-for-pycuda-when-trying-to-set-up-a-nvidia-jeston-agx/259791)
**Author:** [@michael399](https://forums.developer.nvidia.com/u/michael399)\ **Replies:** 1\ **Last updated:** [July 14, 2023, 6:31am UTC](https://forums.developer.nvidia.com/t/getting-failed-to-build-pycuda-error-could-not-build-wheels-for-pycuda-when-trying-to-set-up-a-nvidia-jeston-agx/259791 "2023-07-14T06:31:57Z")
when running pip3 install pycuda, we get that error. any help would be greatly appreciated. Will provide extra details to responses. --- ## [Error using nsight system to analyze pycuda](https://forums.developer.nvidia.com/t/error-using-nsight-system-to-analyze-pycuda/256996)
**Author:** [@jychen2021](https://forums.developer.nvidia.com/u/jychen2021)\ **Replies:** 10\ **Last updated:** [July 3, 2023, 7:51pm UTC](https://forums.developer.nvidia.com/t/error-using-nsight-system-to-analyze-pycuda/256996 "2023-07-03T19:51:50Z")
Using the nsight system CLI to analyze the pycuda program in Windows, an error occurs as shown in the following figure. And the analysis is incomplete. Does anyone know how to fix it? --- ## [PyCuda Cuda Context error encountered when running a model using TensorRT framework on Jetson Orin Nano for inferenced through CSI camera](https://forums.developer.nvidia.com/t/pycuda-cuda-context-error-encountered-when-running-a-model-using-tensorrt-framework-on-jetson-orin-nano-for-inferenced-through-csi-camera/255850)
**Author:** [@karthik.a](https://forums.developer.nvidia.com/u/karthik.a)\ **Replies:** 4\ **Last updated:** [June 26, 2023, 4:17am UTC](https://forums.developer.nvidia.com/t/pycuda-cuda-context-error-encountered-when-running-a-model-using-tensorrt-framework-on-jetson-orin-nano-for-inferenced-through-csi-camera/255850 "2023-06-26T04:17:54Z")
Hi, I have a Jetson Orin Nano device and recently I have been developing a project which involves running an object detection inference on a converted Yolo V7 .pt model to tensorrt on the Jetson Orin Nano device. I hav… --- ## [How to use pycuda to implement concurrency of multiple streams](https://forums.developer.nvidia.com/t/how-to-use-pycuda-to-implement-concurrency-of-multiple-streams/255234)
**Author:** [@jychen2021](https://forums.developer.nvidia.com/u/jychen2021)\ **Replies:** 2\ **Last updated:** [June 1, 2023, 12:26pm UTC](https://forums.developer.nvidia.com/t/how-to-use-pycuda-to-implement-concurrency-of-multiple-streams/255234 "2023-06-01T12:26:57Z")
I’m new to cuda and want to use pycuda for multi-stream concurrent execution, but I don’t know how to do it. The following is my pycuda program and report, I hope to get your help. report36.qdrep (652.3 KB) import pycu… --- ## [Optimal way to read video frames in vpi using open CV and GStreamer](https://forums.developer.nvidia.com/t/optimal-way-to-read-video-frames-in-vpi-using-open-cv-and-gstreamer/249135)
**Author:** [@surajgampa95](https://forums.developer.nvidia.com/u/surajgampa95)\ **Replies:** 4\ **Last updated:** [May 12, 2023, 4:48pm UTC](https://forums.developer.nvidia.com/t/optimal-way-to-read-video-frames-in-vpi-using-open-cv-and-gstreamer/249135 "2023-05-12T16:48:46Z")
Hi, I am trying to read an mp4 file in Python using OpenCV CUDA and GStreamer, and use the captured frame in a VPI pipeline. Is there a way to achieve zero mem copy in the process of capturing the frame and sending it t… --- ## [How to pass a numba cuda DeviceNDArray or GPUArray to VideoCapture render method in jetson\_utils to display streaming video](https://forums.developer.nvidia.com/t/how-to-pass-a-numba-cuda-devicendarray-or-gpuarray-to-videocapture-render-method-in-jetson-utils-to-display-streaming-video/250546)
**Author:** [@surajgampa95](https://forums.developer.nvidia.com/u/surajgampa95)\ **Replies:** 5\ **Last updated:** [May 4, 2023, 2:22pm UTC](https://forums.developer.nvidia.com/t/how-to-pass-a-numba-cuda-devicendarray-or-gpuarray-to-videocapture-render-method-in-jetson-utils-to-display-streaming-video/250546 "2023-05-04T14:22:22Z")
Hi, I am currently using jetson\_utils.VideoSource Capture method to read frames from an mp4 file, perform some operations on those frames using vpi, PyCUDA and/or Numba Cuda, and ultimately pass those frames to jetson\_u… --- ## [Optimal way to copy data between a VPI pipeline and CUDA Kernel in PyCUDA](https://forums.developer.nvidia.com/t/optimal-way-to-copy-data-between-a-vpi-pipeline-and-cuda-kernel-in-pycuda/249136)
**Author:** [@surajgampa95](https://forums.developer.nvidia.com/u/surajgampa95)\ **Replies:** 6\ **Last updated:** [April 24, 2023, 6:19am UTC](https://forums.developer.nvidia.com/t/optimal-way-to-copy-data-between-a-vpi-pipeline-and-cuda-kernel-in-pycuda/249136 "2023-04-24T06:19:44Z")
Hi, I am trying to use an image in vpi format in a CUDA kernel. Currently, I am having to convert the image to a numpy.float32 format. Is there a way to achieve zero mem copy in this pipeline, without having to copy the… --- ## [Code profiling tool for VPI Python and PyCUDA application](https://forums.developer.nvidia.com/t/code-profiling-tool-for-vpi-python-and-pycuda-application/249138)
**Author:** [@surajgampa95](https://forums.developer.nvidia.com/u/surajgampa95)\ **Replies:** 1\ **Last updated:** [April 10, 2023, 3:05am UTC](https://forums.developer.nvidia.com/t/code-profiling-tool-for-vpi-python-and-pycuda-application/249138 "2023-04-10T03:05:17Z")
Hi, I wanted to know if there’s a tool similar to NSight systems and/or NSight compute that I can use to profile my python code where I used VPI, OpenCV Cuda and PyCUDA to monitor latency figures for memory copies and c… --- ## [TensorRT 8.5.2-1+cuda11.8: pycuda.\_driver.LogicError: cuMemcpyHtoDAsync failed: invalid argument](https://forums.developer.nvidia.com/t/tensorrt-8-5-2-1-cuda11-8-pycuda-driver-logicerror-cumemcpyhtodasync-failed-invalid-argument/244906)
**Author:** [@shahizat](https://forums.developer.nvidia.com/u/shahizat)\ **Replies:** 2\ **Last updated:** [March 8, 2023, 11:05am UTC](https://forums.developer.nvidia.com/t/tensorrt-8-5-2-1-cuda11-8-pycuda-driver-logicerror-cumemcpyhtodasync-failed-invalid-argument/244906 "2023-03-08T11:05:12Z")
Dear Nvidia community, Could you please help me to fix the below issue? Thank in advance! un.py:53: DeprecationWarning: Use get\_tensor\_shape instead. size = trt.volume(engine.get\_binding\_shape(binding)) \* batch\_size … --- ## [Saving images on a loop in python using Gstreamer](https://forums.developer.nvidia.com/t/saving-images-on-a-loop-in-python-using-gstreamer/243015)
**Author:** [@billyt2011](https://forums.developer.nvidia.com/u/billyt2011)\ **Replies:** 2\ **Last updated:** [February 16, 2023, 7:17pm UTC](https://forums.developer.nvidia.com/t/saving-images-on-a-loop-in-python-using-gstreamer/243015 "2023-02-16T19:17:27Z")
Hi, I am attempting to use gstreamer and cv2 in order to live stream from two 4096x3040 cameras. Below is the code used to set up the video as well as save the file to an external hardrive. I need to save the image tak… [Next page](https://forums.developer.nvidia.com/tag/pycuda/87.md?match_all_tags=true&page=1&tags%5B%5D=pycuda)