gltracelogger

Syntax:

gltracelogger [-a=frame]
              [-b=count]
              [-d]
              [-i]
              [-o=path]
              [-p=process_name | pid]
              [-S=size] 
              [-s=seconds]
              [-z=frame]

Runs on:

QNX Neutrino

Options:

-a=frame: Specify the frame number on which to start tracing.
-b=count: Specify the number of buffers. Default is 16 buffers.
-d: Start gltracelogger in daemon mode, which means tracing is not started until the specified frame (using -a option). By default, tracing is started right away.
-i: Immediate mode means the shared memory buffers will be flushed immediately after each trace event (useful if you're sending trace data over network for remote replay).
-o=path: Specify the output file .glev location.
-p=process_name | pid: Tell gltracelogger to connect to the named process or pid.
-S=size: Specify the maximum size of the trace file.
-s=seconds: Specify the number of seconds to trace.
-z=frame: Specify the frame number on which to stop tracing.

Description:

This tracing tool is used to help debug graphical issues by providing a trace of all OpenGL function calls made by an application. The gltracelogger is started with a command-line argument specifying the target application by name or pid. The gltracelogger receives the trace data which is then compressed and written into the .glev file, which can be read by gltraceprinter.

To run gltracelogger:

Ensure that screen is running.
Run gltracelogger from a shell.

Examples:

# gltracelogger -p gles3-gears -d -a1 -z1

Environment variables:

GLTRACE_WAIT_FOR_CLIENT

To start tracing the application from the beginning, you need to use the GLTRACE_WAIT_FOR_CLIENT environment variable that forces the application to wait for the gltracelogger to start. The application will resume when the connection with gltracelogger is completed. It can be used along with the application name as following:

GLTRACE_WAIT_FOR_CLIENT=1 gles3-gears