CUDA-GDB
CUDA-GDB is NVIDIA’s debugger for CUDA applications on Linux and QNX.
What CUDA-GDB is
CUDA-GDB is NVIDIA’s debugger for CUDA applications on Linux and QNX. It can debug CPU and GPU code together in one application, including CUDA C/C++ and Fortran; Fortran debugging is limited to 64-bit Linux. Developers can set breakpoints, single-step, and inspect or modify thread memory and variables. The debugger handles kernels compiled for specific CUDA architectures and kernels built at runtime through JIT compilation. Features include source-level and GPU-assembly debugging, conditional breakpoints, memory-access violation detection, autostep, multiple sessions, automatic breaks on kernel launches, core dump debugging, and remote debugging for x86 and ARM. Its command-line interface works locally or over Telnet or SSH, with listed integrations for DDD, EMACS, Nsight Eclipse Edition, and Nsight Visual Studio Code Edition. CUDA-GDB is supplied through the free CUDA Toolkit. macOS host support was dropped with CUDA 12.5.
Good
- Debugs CPU and GPU code together
- Supports breakpoints and single-stepping
- Handles JIT-compiled CUDA kernels
- Offers remote debugging for x86 and ARM
- Free through CUDA Toolkit
Watch out
- Targets Linux and QNX environments
- Fortran debugging limited to 64-bit Linux
- macOS host version dropped with CUDA 12.5
Who it suits
CUDA-GDB suits developers debugging CUDA C/C++, Fortran, and applications using CUDA driver or runtime APIs on Linux or QNX. It is intended for CPU-and-GPU debugging workflows.
Verdict
CUDA-GDB offers detailed source and GPU-assembly debugging with local or remote operation. Its platform limits, especially the lack of current macOS host support, matter when selecting a debugger.
The facts
Each one read from CUDA-GDB's own pages, with the date it was checked.
What it does
- CPU and GPU debugging
- It supports simultaneous debugging of CPU and GPU code within the same application.checked 8 Oct 2026
- CUDA APIs
- It supports applications using the CUDA driver API, runtime API, or both.checked 8 Oct 2026
- Current release features
- The CUDA-GDB 13.4 release notes list CUDA RPC.LO and RPC.HI register access, bindless constant address lookup, and improved barrier information and breakpoint-condition errors.checked 8 Oct 2026
- Debugging controls
- It supports breakpoints, single-stepping, and inspecting or modifying thread memory and variables.checked 8 Oct 2026
- Debugging features
- The product page lists source and GPU assembly debugging, conditional breakpoints, memory access violation detection, autostep, and automatic breaks on kernel launches.checked 8 Oct 2026
- Download and pricing
- NVIDIA presents CUDA Toolkit as free developer tools and provides CUDA-GDB through the toolkit; the opened pages do not state a standalone CUDA-GDB price.checked 8 Oct 2026
- Execution controls
- Users can set breakpoints, single-step applications, and inspect or modify thread memory and variables.checked 8 Oct 2026
- Feature set
- NVIDIA lists source-level and GPU-assembly debugging, conditional breakpoints, memory-access violation detection, autostep, multiple sessions, and remote debugging for x86 and ARM.checked 8 Oct 2026
- GUI integrations
- The product page lists DDD, EMACS, Nsight Eclipse Edition, and Nsight Visual Studio Code Edition as GUI frontend integrations.checked 8 Oct 2026
- Interfaces
- The command-line debugger can run locally or remotely over Telnet or SSH, and NVIDIA lists DDD, EMACS, Nsight Eclipse Edition, and Nsight Visual Studio Code Edition as GUI integrations.checked 8 Oct 2026
- Kernel builds
- It supports kernels compiled for specific CUDA architectures and kernels compiled at runtime with JIT compilation.checked 8 Oct 2026
- Kernel compilation
- It can debug kernels compiled for specific CUDA architectures as well as kernels compiled at runtime with JIT compilation.checked 8 Oct 2026
- Languages
- It supports CUDA C/C++ and Fortran applications, with Fortran debugging limited to 64-bit Linux.checked 8 Oct 2026
- Purpose
- CUDA-GDB is NVIDIA’s debugger for CUDA applications running on Linux and QNX.checked 8 Oct 2026
- Remote debugging
- The feature list includes remote debugging for x86 and ARM.checked 8 Oct 2026
- Security-related capability
- CUDA-GDB release notes state that debugging with Confidential Compute mode is supported with devtools mode.checked 8 Oct 2026
- Support
- NVIDIA directs users with CUDA Tools questions to [email protected].checked 8 Oct 2026
- Support contact
- NVIDIA directs users with CUDA Tools questions or difficulties to [email protected].checked 8 Oct 2026
The company
- Founded
- 1993checked 28 Sep 2026
- Headquarters
- Santa Clara, California, United Stateschecked 28 Sep 2026
Where it runs
- macOS limit
- CUDA Toolkit 11.0 and later no longer supports development or running applications on macOS; the macOS host version of cuda-gdb was dropped as of CUDA 12.5.checked 8 Oct 2026
- macOS support
- NVIDIA states that CUDA Toolkit 11.0 and later no longer supports development or running applications on macOS, and CUDA-GDB macOS host versions were dropped as of CUDA 12.5.checked 8 Oct 2026
Alternatives to CUDA-GDB
- 1
LLDB
Free planAndroidiOSLinuxmacOS7.3Free plan - 2
IDA Pro
Free planLinuxmacOSWindows7.0From $1,099 a year - 3
PonyDebugger
Free planiOSWebmacOSSelf-hosted6.8Free plan - 4
pyOCD
Free planAPILinuxmacOSSelf-hosted6.8Free plan - 5
Perl Debugger
Free planLinuxmacOSWindows6.5Free plan - 6
Black Magic Debug
Free planLinuxmacOSWindows6.4Free plan