Sophie

Sophie

distrib > Mageia > 7 > x86_64 > media > nonfree-updates > by-pkgid > b86a85131cc739c1c53d0b55840a4328 > files > 3708

nvidia-cuda-toolkit-devel-10.1.168-1.2.mga7.nonfree.x86_64.rpm

.TH CUDA-GDB 1 2013-05-06 "NVIDIA" "CUDA Toolkit Documentation"
.sp

.SH NAME

.sp
cuda-gdb \- The NVIDIA CUDA Debugger
.sp
cuda-gdbserver \- The NVIDIA CUDA-GDB server for remote debugging
.sp

.SH SYNOPSIS

.sp
\fBcuda-gdb\fR [gdb-options] [\fB--cuda-use-lockfile\fR
\fIVALUE\fR] \fI[prog [core|procID]]\fR
.sp
\fBcuda-gdb\fR [gdb-options] [\fB--cuda-use-lockfile\fR
\fIVALUE\fR] --args prog \fI[arguments]\fR
.sp

.SH DESCRIPTION

.sp
\fBcuda-gdb\fR is the NVIDIA CUDA debugger for Unix platforms. It
is \fBgdb\fR with extra commands specific to CUDA debugging.
.sp
All the existing \fBgdb\fR commands are supported. On top of
those commands, the following CUDA-specific commands are also
supported.
.SS Info Commands

.sp
Commands to query information about CUDA activities:
.sp

.TP
\fBinfo cuda devices\fR
information about all the devices
.TP
\fBinfo cuda sms\fR
information about all the SMs in the current device
.TP
\fBinfo cuda warps\fR
information about all the warps in the current SM
.TP
\fBinfo cuda lanes\fR
information about all the lanes in the current warp
.TP
\fBinfo cuda kernels\fR
information about all the active kernels
.TP
\fBinfo cuda contexts\fR
information about all the contexts
.TP
\fBinfo cuda blocks\fR
information about all the active blocks in the current kernel
.TP
\fBinfo cuda threads\fR
information about all the active threads in the current kernel
.TP
\fBinfo cuda launch trace\fR
information about the parent kernels of the kernel in focus
.TP
\fBinfo cuda launch children\fR
information about the kernels launched by the kernels in focus
.SS Focus Commands

.sp
Commands to query or switch the focus of the debugger:
.sp

.TP
\fBcuda block\fR
Print or select the current CUDA block
.TP
\fBcuda device\fR
Print or select the current CUDA device
.TP
\fBcuda grid\fR
Print or select the current CUDA grid
.TP
\fBcuda kernel\fR
Print or select the current CUDA kernel
.TP
\fBcuda lane\fR
Print or select the current CUDA lane
.TP
\fBcuda sm\fR
Print or select the current CUDA SM
.TP
\fBcuda thread\fR
Print or select the current CUDA thread
.TP
\fBcuda warp\fR
Print or select the current CUDA warp
.SS Configuration Commands

.sp
Commands to configure the CUDA-specific commands:
.sp

.TP
\fBset cuda api_failures\fR
Set the api_failures to ignore/stop/hide on CUDA driver API call
errors
.TP
\fBset cuda break_on_launch\fR
Automatically set a breakpoint at the entrance of kernels
.TP
\fBset cuda coalescing\fR
Turn on/off coalescing of the CUDA commands output
.TP
\fBset cuda context_events\fR
Turn on/off context events (push/pop/create/destroy) output
messages
.TP
\fBset cuda defer_kernel_launch_notifications\fR
Turn on/off deferral of kernel launch messages
.TP
\fBset cuda disassemble_from\fR
Choose whether to disassemble from the device memory (slow) or
the ELF image (fast)
.TP
\fBset cuda gpu_busy_check\fR
Turn on/off GPU busy check the next time the inferior application
is run
.TP
\fBset cuda hide_internal_frame\fR
Set hiding of the internal CUDA frames when printing the call
stack
.TP
\fBset cuda kernel_events\fR
Turn on/off kernel events (launch/termination) output messages
.TP
\fBset cuda launch_blocking\fR
Turn on/off CUDA kernel launch blocking (effective starting from
the next run)
.TP
\fBset cuda memcheck\fR
Turn on/off CUDA Memory Checker next time the inferior
application is run
.TP
\fBset cuda notify\fR
Thread to notify about CUDA events when no other known candidate
.TP
\fBset cuda software_preemption\fR
Turn on/off CUDA software preemption debugging the next time the
inferior application is run
.TP
\fBset cuda thread_selection\fR
Set the automatic thread selection policy to use when the current
thread cannot be selected
.sp

.SH OPTIONS

.sp
CUDA-GDB supports the same options as GDB, with the addition of
the following options:
.sp

.TP
\fB--cuda-use-lockfile=VALUE\fR
If VALUE is 0, CUDA-GDB will not create a lock file and will not
enforce the exclusive usage of GPUs.
.sp

.SH ENVIRONMENT

.sp

.TP
\fBTMPDIR\fR
If non-null, the full pathname for an alternate temporary
directory. Note that the debugger must have write and exec
permission to that directory.
.sp

.SH FILES

.sp

.TP
\fB~/.cuda-gdbinit\fR
Per user configuration file. The file format is indentical to
~/.gdbinit. See \fBgdb\fR(5) for further details.
.TP
\fB/tmp/cuda-dbg/\fR
Temporary directory where the debugger stores intermediate files.
The directory must have write and and exec permissions. Use
$TMPDIR to override the location.
.sp

.SH NOTES

.sp
For more information about cuda-gdb, please see the online
documentation at http://docs.nvidia.com/cuda/cuda-gdb/index.html.
.sp
For more information about CUDA-related features, type \fBhelp
cuda\fR or \fBhelp info cuda\fR from within CUDA-GDB, or consult
the CUDA-GDB online manual.
.sp

.SH SEE ALSO

.sp
\fBcuda-binaries\fR(1), \fBcuda-libraries\fR(7), \fBgdb\fR(1)
.sp

.SH REPORTING BUGS

.sp
Email bugs to cuda-debugger-bugs@nvidia.com
.sp

.SH COPYRIGHT

.sp
©2013 NVIDIA Corporation. All rights reserved.