# CUDA-GDB **URL:** https://forums.developer.nvidia.com/c/developer-tools/cuda-developer-tools/cuda-gdb/110.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) --- ## [Cuda-gdb can't debug fortran on an A100 card](https://forums.developer.nvidia.com/t/cuda-gdb-cant-debug-fortran-on-an-a100-card/379988)
**Author:** [@bba](https://forums.developer.nvidia.com/u/bba)\ **Replies:** 8\ **Last updated:** [October 16, 2026, 3:12am UTC](https://forums.developer.nvidia.com/t/cuda-gdb-cant-debug-fortran-on-an-a100-card/379988 "2026-10-16T03:12:07Z")
I am working on Fortran on our cluster, on the tutorial from the book CUDA Fortran for Scienctists and Engineers, specifically on increment. This works but I want to look at a simple example using cuda-gdb, but I can’t r… --- ## [Const changing while moving out of function in debugger](https://forums.developer.nvidia.com/t/const-changing-while-moving-out-of-function-in-debugger/366424)
**Author:** [@l.e.v.d.wal](https://forums.developer.nvidia.com/u/l.e.v.d.wal)\ **Replies:** 5\ **Last updated:** [May 29, 2026, 8:56am UTC](https://forums.developer.nvidia.com/t/const-changing-while-moving-out-of-function-in-debugger/366424 "2026-05-29T08:56:13Z")
While attempting to debug a CUDA program I was inspecting a specific function call inside a kernel. I was inspecting the sudden changing of a variable value which was marked as a const. Now, I noticed that when I continu… --- ## [Use CoreDump to capture GPU out-of-bounds memory access](https://forums.developer.nvidia.com/t/use-coredump-to-capture-gpu-out-of-bounds-memory-access/370891)
**Author:** [@271732480](https://forums.developer.nvidia.com/u/271732480)\ **Replies:** 1\ **Last updated:** [May 27, 2026, 2:23pm UTC](https://forums.developer.nvidia.com/t/use-coredump-to-capture-gpu-out-of-bounds-memory-access/370891 "2026-05-27T14:23:28Z")
Hi, In the debugging environment, compute-sanitizer can effectively capture GPU memory out-of-bounds issues. However, it is difficult to apply compute-sanitizer in actual production environments because it severely impa… --- ## [Internal-error: create\_module: Assertion 'context' failed in cuda-gdb 13.2 when debugging a program with Unified Memory](https://forums.developer.nvidia.com/t/internal-error-create-module-assertion-context-failed-in-cuda-gdb-13-2-when-debugging-a-program-with-unified-memory/366265)
**Author:** [@ededwards129](https://forums.developer.nvidia.com/u/ededwards129)\ **Replies:** 1\ **Last updated:** [April 17, 2026, 3:20am UTC](https://forums.developer.nvidia.com/t/internal-error-create-module-assertion-context-failed-in-cuda-gdb-13-2-when-debugging-a-program-with-unified-memory/366265 "2026-04-17T03:20:39Z")
Describe the bug When attempting to debug a simple CUDA program using Unified Memory (cudaMallocManaged) with cuda-gdb, the debugger crashes with an internal error: internal-error: create\_module: Assertion 'context' fail… --- ## [Three concurrent cudaMemcpyAsync(HtoD) calls still serialized on H100/H20 even when asyncEngineCount == 3](https://forums.developer.nvidia.com/t/three-concurrent-cudamemcpyasync-htod-calls-still-serialized-on-h100-h20-even-when-asyncenginecount-3/365145)
**Author:** [@2218887102](https://forums.developer.nvidia.com/u/2218887102)\ **Replies:** 1\ **Last updated:** [April 2, 2026, 2:46am UTC](https://forums.developer.nvidia.com/t/three-concurrent-cudamemcpyasync-htod-calls-still-serialized-on-h100-h20-even-when-asyncenginecount-3/365145 "2026-04-02T02:46:09Z")
Hi all, I am trying to understand the real meaning of asyncEngineCount == 3 on Hopper GPUs such as H100 and H20. In my test, I launch three host threads, and each thread asynchronously submits cudaMemcpyAsync(..., cuda… --- ## [Breakpoints in \*.cpp files are not hit, while breakpoints in \*.cu files are working](https://forums.developer.nvidia.com/t/breakpoints-in-cpp-files-are-not-hit-while-breakpoints-in-cu-files-are-working/360560)
**Author:** [@xitlk](https://forums.developer.nvidia.com/u/xitlk)\ **Replies:** 0\ **Last updated:** [February 13, 2026, 8:08pm UTC](https://forums.developer.nvidia.com/t/breakpoints-in-cpp-files-are-not-hit-while-breakpoints-in-cu-files-are-working/360560 "2026-02-13T20:08:11Z")
Hi, I am trying to debug a proof-of-concept small C++ app. The app consists from one cpp file (main.cpp) which calls a wrapper function defined in a \*.cu file. This wrapper function then calls another function defined i… --- ## [No support for Python 3.13 in cuda-gdb?](https://forums.developer.nvidia.com/t/no-support-for-python-3-13-in-cuda-gdb/354832)
**Author:** [@utkarsh.verma](https://forums.developer.nvidia.com/u/utkarsh.verma)\ **Replies:** 1\ **Last updated:** [January 30, 2026, 7:34am UTC](https://forums.developer.nvidia.com/t/no-support-for-python-3-13-in-cuda-gdb/354832 "2026-01-30T07:34:39Z")
I just noticed that cuda-gdb (v13.0) could not pick up Python support on my machine because I have Python 3.13 installed. $ ls -la /opt/cuda/bin/cuda-gdb-python\* 2\>/dev/null -rwxr-xr-x 1 root root 15970544 Nov … --- ## [CUDA coredump file corrupted](https://forums.developer.nvidia.com/t/cuda-coredump-file-corrupted/355162)
**Author:** [@cun.zhang](https://forums.developer.nvidia.com/u/cun.zhang)\ **Replies:** 4\ **Last updated:** [December 22, 2025, 11:58am UTC](https://forums.developer.nvidia.com/t/cuda-coredump-file-corrupted/355162 "2025-12-22T11:58:09Z")
Platform: RTX 5090 export CUDA\_ENABLE\_COREDUMP\_ON\_EXCEPTION=1 then run my command: coredump generated, but it’s corrupted: cuda-gdb \`target cudacore core\_1766111917\_9180be2fb883\_1125.nvcudmp\` Opening GPU coredump: … --- ## [Cannot debug application using WSL2 and cuda-gdb](https://forums.developer.nvidia.com/t/cannot-debug-application-using-wsl2-and-cuda-gdb/354566)
**Author:** [@utkarsh.verma](https://forums.developer.nvidia.com/u/utkarsh.verma)\ **Replies:** 4\ **Last updated:** [December 17, 2025, 7:54pm UTC](https://forums.developer.nvidia.com/t/cannot-debug-application-using-wsl2-and-cuda-gdb/354566 "2025-12-17T19:54:03Z")
I have a Windows machine where I wanted to use cuda-gdb within WSL2 using VS Code. When I launch my application, I am greeted with this: WARNING: Debug interface is not enabled. Please see https://docs.nvidia.com/cuda/c… --- ## [On rtx5060, with 576.52 driver and 12.9 cuda toolkit, cuda-gdb cannot step into kernel function. Please Help!](https://forums.developer.nvidia.com/t/on-rtx5060-with-576-52-driver-and-12-9-cuda-toolkit-cuda-gdb-cannot-step-into-kernel-function-please-help/353731)
**Author:** [@chenxlei23](https://forums.developer.nvidia.com/u/chenxlei23)\ **Replies:** 2\ **Last updated:** [December 8, 2025, 6:17am UTC](https://forums.developer.nvidia.com/t/on-rtx5060-with-576-52-driver-and-12-9-cuda-toolkit-cuda-gdb-cannot-step-into-kernel-function-please-help/353731 "2025-12-08T06:17:35Z")
My device is rtx5060 and I’m trying to debug the following demo program in Ubuntu-24.04 WSL: #include \ #include \ \_\_global\_\_ void testKernel(int \*data) { int idx = threadIdx.x; data\[idx\] … --- ## [Why my gdb shows ?() when i use CUDA COREDUMP?](https://forums.developer.nvidia.com/t/why-my-gdb-shows-when-i-use-cuda-coredump/350437)
**Author:** [@qq934603493](https://forums.developer.nvidia.com/u/qq934603493)\ **Replies:** 2\ **Last updated:** [November 11, 2025, 10:22am UTC](https://forums.developer.nvidia.com/t/why-my-gdb-shows-when-i-use-cuda-coredump/350437 "2025-11-11T10:22:08Z")
Normally, if I not set CUDA\_ENABLE\_COREDUMP\_ON\_EXCEPTION and my program crashed. I can use gdb my\_program coredump\_file to see the stack perfectly(using gdb 8.3). But when i set CUDA\_ENABLE\_COREDUMP\_ON\_EXCEPTION=1, then … --- ## ['cudafe++' died with status 0xC0000005 (ACCESS\_VIOLATION)](https://forums.developer.nvidia.com/t/cudafe-died-with-status-0xc0000005-access-violation/350686)
**Author:** [@temur.tskhvedadze](https://forums.developer.nvidia.com/u/temur.tskhvedadze)\ **Replies:** 1\ **Last updated:** [November 10, 2025, 4:17am UTC](https://forums.developer.nvidia.com/t/cudafe-died-with-status-0xc0000005-access-violation/350686 "2025-11-10T04:17:48Z")
hello, i have this problem “‘cudafe++’ died with status 0xC0000005 (ACCESS\_VIOLATION)” windows 11. laptop rtx 3050 6gb. visual studio 2022. now i have cuda 12.4.1 ( also i try vs 2019. i try from cuda 11.8 to lates… --- ## [\[Jetson Orin AGX | CUDA 12.6\] cuda-gdb causes SIGSEGV in libcudadebugger.so.1 when entering kernel](https://forums.developer.nvidia.com/t/jetson-orin-agx-cuda-12-6-cuda-gdb-causes-sigsegv-in-libcudadebugger-so-1-when-entering-kernel/349503)
**Author:** [@yuki.akimoto.sub](https://forums.developer.nvidia.com/u/yuki.akimoto.sub)\ **Replies:** 2\ **Last updated:** [November 7, 2025, 12:44am UTC](https://forums.developer.nvidia.com/t/jetson-orin-agx-cuda-12-6-cuda-gdb-causes-sigsegv-in-libcudadebugger-so-1-when-entering-kernel/349503 "2025-11-07T00:44:36Z")
Hello, I’m encountering a segmentation fault when debugging even the simplest CUDA program using cuda-gdb on Jetson Orin AGX. The program crashes immediately when entering the kernel function. Reproducible code #inc… --- ## [Attach failed due to the internal driver error 0x414300000002e with cuda-gdb](https://forums.developer.nvidia.com/t/attach-failed-due-to-the-internal-driver-error-0x414300000002e-with-cuda-gdb/349680)
**Author:** [@odyssey471](https://forums.developer.nvidia.com/u/odyssey471)\ **Replies:** 7\ **Last updated:** [October 31, 2025, 8:04am UTC](https://forums.developer.nvidia.com/t/attach-failed-due-to-the-internal-driver-error-0x414300000002e-with-cuda-gdb/349680 "2025-10-31T08:04:40Z")
On a Linux server, when I try to attach a running process with cuda-gdb, I would receive such an error after lots of reading symbol from a .so lines. But the cuda-gdb seems attached to the target process, however it los… --- ## [WSL2 cuda-gdb cannot get into kernel](https://forums.developer.nvidia.com/t/wsl2-cuda-gdb-cannot-get-into-kernel/344679)
**Author:** [@Akiam](https://forums.developer.nvidia.com/u/Akiam)\ **Replies:** 5\ **Last updated:** [October 15, 2025, 1:41pm UTC](https://forums.developer.nvidia.com/t/wsl2-cuda-gdb-cannot-get-into-kernel/344679 "2025-10-15T13:41:53Z")
Hi, I’m developing CUDA application on Windows 11 24H2 using WSL2 with Ubuntu 22.04, and debugging with VS Code. currently I can compile and run the program normally, but if I run it with cuda-gdb, it gets stuck. Then … --- ## [Cuda-gdb can't step into kernel code in WSL2](https://forums.developer.nvidia.com/t/cuda-gdb-cant-step-into-kernel-code-in-wsl2/347034)
**Author:** [@1154330861](https://forums.developer.nvidia.com/u/1154330861)\ **Replies:** 9\ **Last updated:** [October 10, 2025, 7:52am UTC](https://forums.developer.nvidia.com/t/cuda-gdb-cant-step-into-kernel-code-in-wsl2/347034 "2025-10-10T07:52:15Z")
Problem I encountered As the topic says, I keep encountering this issue: \[Detaching after vfork from child process 5146\] WARNING: Debug interface is not enabled. Please see https://docs.nvidia.com/cuda/cuda-gdb/index.ht… --- ## [Built a CUDA-native code editor with real-time profiling and GPU emulation](https://forums.developer.nvidia.com/t/built-a-cuda-native-code-editor-with-real-time-profiling-and-gpu-emulation/347255)
**Author:** [@jaberjaber2](https://forums.developer.nvidia.com/u/jaberjaber2)\ **Replies:** 0\ **Last updated:** [October 10, 2025, 1:05am UTC](https://forums.developer.nvidia.com/t/built-a-cuda-native-code-editor-with-real-time-profiling-and-gpu-emulation/347255 "2025-10-10T01:05:20Z")
Hi everyone, I’m part of the NVIDIA Inception Program and wanted to share something I’ve been building for CUDA developers. It’s called RightNow AI Code Editor, a GPU-native environment that lets you • Profile and ben… --- ## [Hit rate is not as expected for a simple kernel](https://forums.developer.nvidia.com/t/hit-rate-is-not-as-expected-for-a-simple-kernel/347064)
**Author:** [@mossaied2](https://forums.developer.nvidia.com/u/mossaied2)\ **Replies:** 2\ **Last updated:** [October 8, 2025, 11:24am UTC](https://forums.developer.nvidia.com/t/hit-rate-is-not-as-expected-for-a-simple-kernel/347064 "2025-10-08T11:24:03Z")
Hello all; I have a simple kernel (I am sharing below) which consists of 1 warp and 1 block and also 1 thread. I coded that kernel just to study l1 cache hit rate that is why I wanted it to be very simple. I have a vec… --- ## [About Async(Copy) Engines](https://forums.developer.nvidia.com/t/about-async-copy-engines/346710)
**Author:** [@zetpop10](https://forums.developer.nvidia.com/u/zetpop10)\ **Replies:** 2\ **Last updated:** [October 7, 2025, 10:55am UTC](https://forums.developer.nvidia.com/t/about-async-copy-engines/346710 "2025-10-07T10:55:29Z")
I want to know the number of asynchronous engines of the graphics cards I’m asking about. And if the number of asynchronous engines is 1 instead of 2 or more, I’d like to know how to perform HtoD / DtoH simultaneously. … --- ## [Coredump on CUDBG\_EXCEPTION\_WARP\_ILLEGAL\_ADDRESS dont show lane/warp info](https://forums.developer.nvidia.com/t/coredump-on-cudbg-exception-warp-illegal-address-dont-show-lane-warp-info/344581)
**Author:** [@goutham24693](https://forums.developer.nvidia.com/u/goutham24693)\ **Replies:** 4\ **Last updated:** [October 7, 2025, 3:42am UTC](https://forums.developer.nvidia.com/t/coredump-on-cudbg-exception-warp-illegal-address-dont-show-lane-warp-info/344581 "2025-10-07T03:42:48Z")
Platform: windows 11 $ nvidia-smi.exe Wed Sep 10 11:01:46 2025 ±----------------------------------------------------------------------------------------+ | NVIDIA-SMI 581.15 Driver Version: 581.15 … --- ## [How to debug cuda in python(use cupy to call kernels, not lib file)](https://forums.developer.nvidia.com/t/how-to-debug-cuda-in-python-use-cupy-to-call-kernels-not-lib-file/346039)
**Author:** [@S200331082](https://forums.developer.nvidia.com/u/S200331082)\ **Replies:** 2\ **Last updated:** [September 26, 2025, 1:19am UTC](https://forums.developer.nvidia.com/t/how-to-debug-cuda-in-python-use-cupy-to-call-kernels-not-lib-file/346039 "2025-09-26T01:19:31Z")
I use cupy to call cuda kernels, but I don’t know how to debug cuda code, here is my wrapper file: wrapper.py import math from pathlib import Path import cupy as cp import numpy as np with open(Path(\_\_file\_\_).parents… --- ## [CUDA coredumps not being generated](https://forums.developer.nvidia.com/t/cuda-coredumps-not-being-generated/341420)
**Author:** [@jonathan.cameron](https://forums.developer.nvidia.com/u/jonathan.cameron)\ **Replies:** 14\ **Last updated:** [September 18, 2025, 9:23am UTC](https://forums.developer.nvidia.com/t/cuda-coredumps-not-being-generated/341420 "2025-09-18T09:23:30Z")
I’m trying to generate .nvcudmp core dump files on a linux x86\_64 target. I’ve done the following: export CUDA\_ENABLE\_COREDUMP\_ON\_EXCEPTION=1 export CUDA\_COREDUMP\_SHOW\_PROGRESS=1 before running running the process. I… --- ## [Cannot step into CUDA kernel while debugging with VSCode on WSL2 Ubuntu 22.04](https://forums.developer.nvidia.com/t/cannot-step-into-cuda-kernel-while-debugging-with-vscode-on-wsl2-ubuntu-22-04/335284)
**Author:** [@lhbran316](https://forums.developer.nvidia.com/u/lhbran316)\ **Replies:** 5\ **Last updated:** [September 15, 2025, 10:33am UTC](https://forums.developer.nvidia.com/t/cannot-step-into-cuda-kernel-while-debugging-with-vscode-on-wsl2-ubuntu-22-04/335284 "2025-09-15T10:33:41Z")
Hi, I’m developing a CUDA application on Windows 11 using WSL2 with Ubuntu 22.04, and debugging with VS Code. My current setup allows me to: Successfully run CUDA programs (nvcc and cuda-gdb are working), Launch and … --- ## [Troubleshooting a Multi-GPU Program](https://forums.developer.nvidia.com/t/troubleshooting-a-multi-gpu-program/341371)
**Author:** [@keyur.joshi](https://forums.developer.nvidia.com/u/keyur.joshi)\ **Replies:** 1\ **Last updated:** [August 12, 2025, 2:50am UTC](https://forums.developer.nvidia.com/t/troubleshooting-a-multi-gpu-program/341371 "2025-08-12T02:50:13Z")
OS:Ubuntu 22.04LTS CPU: Intel Core i9 - 7900 X GPU Cards : Titan Xp No. of GPU Cards: 4 (All identical) CUDA Toolkit version: 12.9 Hi, I am trying to write a program using multiple GPU cards. The code is given belo… --- ## [Unable to set breakpoints in cuda kernel when logging into wsl using remote ssh](https://forums.developer.nvidia.com/t/unable-to-set-breakpoints-in-cuda-kernel-when-logging-into-wsl-using-remote-ssh/340280)
**Author:** [@user34585](https://forums.developer.nvidia.com/u/user34585)\ **Replies:** 1\ **Last updated:** [August 5, 2025, 8:03pm UTC](https://forums.developer.nvidia.com/t/unable-to-set-breakpoints-in-cuda-kernel-when-logging-into-wsl-using-remote-ssh/340280 "2025-08-05T20:03:23Z")
First, I can debug the cuda kernel on vscode using the nsight-vscode-edition extension on top of the host where wsl is located(wsl is running on a win11 pc)(I refer to this link 1. Introduction — CUDA-GDB 12.9 documentat… --- ## [Different results from printf and cuda-gdb](https://forums.developer.nvidia.com/t/different-results-from-printf-and-cuda-gdb/340730)
**Author:** [@peiqi.zhang97](https://forums.developer.nvidia.com/u/peiqi.zhang97)\ **Replies:** 1\ **Last updated:** [August 1, 2025, 8:27am UTC](https://forums.developer.nvidia.com/t/different-results-from-printf-and-cuda-gdb/340730 "2025-08-01T08:27:45Z")
I’m probing register c\[0\], c\[1\], c\[2\], c\[3\] in a CUDA program. The strange thing is that the results shown in CUDA-GDB is correct, but the results from printf is wrong. The program finally write back wrong results. … --- ## [\[BUG REPORT\] CUDA-GDB refuses to step into GPU Kernel](https://forums.developer.nvidia.com/t/bug-report-cuda-gdb-refuses-to-step-into-gpu-kernel/339460)
**Author:** [@areenraj](https://forums.developer.nvidia.com/u/areenraj)\ **Replies:** 1\ **Last updated:** [July 23, 2025, 6:24pm UTC](https://forums.developer.nvidia.com/t/bug-report-cuda-gdb-refuses-to-step-into-gpu-kernel/339460 "2025-07-23T18:24:21Z")
Hello Everyone! I have a simple kernel that performs some linear algebra operations like Gauss Elimination, Matrix Vector Product, etc. A few days ago, I began to notice some issues while debugging my case with cuda-gdb… --- ## [CUDA 11.8 Build issues --\> MSB3721](https://forums.developer.nvidia.com/t/cuda-11-8-build-issues-msb3721/308439)
**Author:** [@manvanth.gs](https://forums.developer.nvidia.com/u/manvanth.gs)\ **Replies:** 5\ **Last updated:** [June 24, 2025, 10:13am UTC](https://forums.developer.nvidia.com/t/cuda-11-8-build-issues-msb3721/308439 "2025-06-24T10:13:43Z")
Hi, I am trying to build a CUDA library, while trying to build the library I’m getting MSB3721 error. This issue is very ambiguous, the issue is not occurring in one code but is occurring in a different code. The config… --- ## [Cannot find python3.13-tui in last version CUDA Toolkit](https://forums.developer.nvidia.com/t/cannot-find-python3-13-tui-in-last-version-cuda-toolkit/337016)
**Author:** [@perestoronin](https://forums.developer.nvidia.com/u/perestoronin)\ **Replies:** 1\ **Last updated:** [June 24, 2025, 7:39am UTC](https://forums.developer.nvidia.com/t/cannot-find-python3-13-tui-in-last-version-cuda-toolkit/337016 "2025-06-24T07:39:47Z")
Cannot find python3.13-tui in last version CUDA Toolkit, exists only last python3.12-tui, but new stable python version 3.13 --- ## [Cuda gdb cannot load coredump file, error: Assertion \`m\_instance.m\_num\_devices \> 0' failed](https://forums.developer.nvidia.com/t/cuda-gdb-cannot-load-coredump-file-error-assertion-m-instance-m-num-devices-0-failed/334412)
**Author:** [@Prayer\_Anna](https://forums.developer.nvidia.com/u/Prayer_Anna)\ **Replies:** 1\ **Last updated:** [June 5, 2025, 4:23am UTC](https://forums.developer.nvidia.com/t/cuda-gdb-cannot-load-coredump-file-error-assertion-m-instance-m-num-devices-0-failed/334412 "2025-06-05T04:23:37Z")
An error occurs when loading coredump in cuda-gdb: root@yy-lite-decode-0:/models# cuda-gdb NVIDIA (R) cuda-gdb 12.4 Portions Copyright (C) 2007-2023 NVIDIA Corporation Based on GNU gdb 13.1 Copyright (C) 2023 Free Softw… [Next page](https://forums.developer.nvidia.com/c/developer-tools/cuda-developer-tools/cuda-gdb/110.md?page=1)