Skip to content
This repository

Tools for tracing OpenGL, Direct3D, and other graphics APIs

gui: respect the alpha channel in surfaces

We weren't rendering any background which meant that surfaces
that were fully transparent would just show garbage that was
underneath. Lets make sure that the widget has a painted background
instead of garbage underneath.
latest commit 8d6c36d74b
zackr zackr authored
Octocat-spinner-32 cli apitrace trim: plug memory leak in trim_trace() (cli_trim.cpp) December 10, 2013
Octocat-spinner-32 cmake android.toolchain.cmake: update from upstream (#225) February 10, 2014
Octocat-spinner-32 common Use /data/data/$procname/$procname.trace on Android February 28, 2014
Octocat-spinner-32 dispatch gltrace: Fake GL_EXT_debug_marker/label too. February 04, 2014
Octocat-spinner-32 gui gui: respect the alpha channel in surfaces April 09, 2014
Octocat-spinner-32 helpers gltrace: Fix the computation of image sizes. January 21, 2014
Octocat-spinner-32 image Add includes for algorithm. February 25, 2014
Octocat-spinner-32 inject inject: Provide a helpful message when the wrapper DLL is never actua… October 23, 2013
Octocat-spinner-32 retrace d3dretrace: Force windowed mode on D3D9 traces. April 08, 2014
Octocat-spinner-32 scripts tracediff: Don't highlight differences if the call nos are actually t… April 08, 2014
Octocat-spinner-32 specs glstate: Describe a few more transform feedback states. March 19, 2014
Octocat-spinner-32 thirdparty md5: Fix include for memcpy. February 28, 2014
Octocat-spinner-32 wrappers gltrace: Handle GL_EXT_compiled_vertex_array correctly. March 07, 2014
Octocat-spinner-32 .dir-locals.el Improve the indention in emacs just a bit. October 25, 2011
Octocat-spinner-32 .gitignore qapitrace: ignore more supporting files generated by qtcreator January 23, 2014
Octocat-spinner-32 .lvimrc Set tabstops to 8 April 23, 2013
Octocat-spinner-32 Android.mk Update Android.mk to add support for x86 October 29, 2013
Octocat-spinner-32 BUGS.markdown Aggregate all info on debugging into a single place. July 12, 2013
Octocat-spinner-32 CMakeLists.txt gltrace: Detect the availability of tr1/memory header. February 25, 2014
Octocat-spinner-32 DEVELOPMENT.markdown Several documentation fixes. June 20, 2013
Octocat-spinner-32 Dalvik.markdown egltrace: Drop debug.apitrace.procname support. February 28, 2014
Octocat-spinner-32 INSTALL.markdown Add note about Qt and Windows 64 bits builds. June 14, 2013
Octocat-spinner-32 LICENSE Update copyright in License. September 27, 2012
Octocat-spinner-32 NEWS.markdown Tag a new release 4.0. May 02, 2013
Octocat-spinner-32 README.markdown apitrace documentation: add multiple callsets February 25, 2014
Octocat-spinner-32 TODO.markdown Don't package TODO.markdown anymore. March 14, 2012
README.markdown

About apitrace

apitrace consists of a set of tools to:

  • trace OpenGL, OpenGL ES, Direct3D, and DirectDraw APIs calls to a file;

  • replay OpenGL and OpenGL ES calls from a file;

  • inspect OpenGL state at any call while retracing;

  • visualize and edit trace files.

See the apitrace homepage for more details.

Obtaining apitrace

To obtain apitrace either download the latest binaries for your platform if available, or follow the instructions in INSTALL.markdown to build it yourself. On 64bits Linux and Windows platforms you'll need apitrace binaries that match the architecture (32bits or 64bits) of the application being traced.

Basic usage

Run the application you want to trace as

apitrace trace --api API /path/to/application [args...]

and it will generate a trace named application.trace in the current directory. You can specify the written trace filename by passing the --output command line option.

Problems while tracing (e.g, if the application uses calls/parameters unsupported by apitrace) will be reported via stderr output on Unices. On Windows you'll need to run DebugView to view these messages.

Follow the "Tracing manually" instructions below if you cannot obtain a trace.

View the trace with

apitrace dump application.trace

Replay an OpenGL trace with

apitrace replay application.trace

Pass the --sb option to use a single buffered visual. Pass --help to apitrace replay for more options.

Basic GUI usage

Start the GUI as

qapitrace application.trace

You can also tell the GUI to go directly to a specific call

qapitrace application.trace 12345

Backtrace Capturing

apitrace now has the ability to capture the call stack to an OpenGL call. This can be helpful in determing which piece of code made that glDrawArrays call.

NOTE this feature is currently only available on Android and Linux at the moment.

On linux you need to have libunwind, and libdwarf installed to compile in the feature.

To use the feature you need to set an environment variable with the list of GL call prefixes you wish to capture stack traces to.

export APITRACE_BACKTRACE="glDraw* glUniform*"

The backtrace data will show up in qapitrace in the bottom section as a new tab.

Advanced command line usage

Call sets

Several tools take CALLSET arguments, e.g:

apitrace dump --calls=CALLSET foo.trace
apitrace dump-images --calls=CALLSET foo.trace
apitrace trim --calls=CALLSET1 --calls=CALLSET2 foo.trace

The call syntax is very flexible. Here are a few examples:

  • 4 one call

  • 0,2,4,5 set of calls

  • "0 2 4 5" set of calls (commas are optional and can be replaced with whitespace)

  • 0-100/2 calls 1, 3, 5, ..., 99

  • 0-1000/draw all draw calls between 0 and 1000

  • 0-1000/fbo all fbo changes between calls 0 and 1000

  • frame all calls at end of frames

  • @foo.txt read call numbers from foo.txt, using the same syntax as above

Tracing manually

Linux

On 64 bits systems, you'll need to determine whether the application is 64 bits or 32 bits. This can be done by doing

file /path/to/application

But beware of wrapper shell scripts -- what matters is the architecture of the main process.

Run the GLX application you want to trace as

LD_PRELOAD=/path/to/apitrace/wrappers/glxtrace.so /path/to/application

and it will generate a trace named application.trace in the current directory. You can specify the written trace filename by setting the TRACE_FILE environment variable before running.

For EGL applications you will need to use egltrace.so instead of glxtrace.so.

The LD_PRELOAD mechanism should work with the majority applications. There are some applications (e.g., Unigine Heaven, Android GPU emulator, etc.), that have global function pointers with the same name as OpenGL entrypoints, living in a shared object that wasn't linked with -Bsymbolic flag, so relocations to those global function pointers get overwritten with the address to our wrapper library, and the application will segfault when trying to write to them. For these applications it is possible to trace by using glxtrace.so as an ordinary libGL.so and injecting it via LD_LIBRARY_PATH:

ln -s glxtrace.so wrappers/libGL.so
ln -s glxtrace.so wrappers/libGL.so.1
ln -s glxtrace.so wrappers/libGL.so.1.2
export LD_LIBRARY_PATH=/path/to/apitrace/wrappers:$LD_LIBRARY_PATH
export TRACE_LIBGL=/path/to/real/libGL.so.1
/path/to/application

If you are an application developer, you can avoid this either by linking with -Bsymbolic flag, or by using some unique prefix for your function pointers.

See the ld.so man page for more information about LD_PRELOAD and LD_LIBRARY_PATH environment flags.

Android

To trace standalone native OpenGL ES applications, use LD_PRELOAD=/path/to/egltrace.so /path/to/application as described in the previous section. To trace Java applications, refer to Dalvik.markdown.

Mac OS X

Run the application you want to trace as

DYLD_FRAMEWORK_PATH=/path/to/apitrace/wrappers /path/to/application

Note that although Mac OS X has an LD_PRELOAD equivalent, DYLD_INSERT_LIBRARIES, it is mostly useless because it only works with DYLD_FORCE_FLAT_NAMESPACE=1 which breaks most applications. See the dyld man page for more details about these environment flags.

Windows

When tracing third-party applications, you can identify the target application's main executable, either by:

  • right clicking on the application's icon in the Start Menu, choose Properties, and see the Target field;

  • or by starting the application, run Windows Task Manager (taskmgr.exe), right click on the application name in the Applications tab, choose Go To Process, note the highlighted Image Name, and search it on C:\Program Files or C:\Program Files (x86).

On 64 bits Windows, you'll need to determine ether the application is a 64 bits or 32 bits. 32 bits applications will have a *32 suffix in the Image Name column of the Processes tab of Windows Task Manager window.

You also need to know which graphics API is being used. If you are unsure, the simplest way to determine what API an application uses is to:

  • download and run Process Explorer

  • search and select the application's process in Process Explorer

  • list the DLLs by pressing Ctrl + D

  • sort DLLs alphabetically, and look for the DLLs such as opengl32.dll, d3d9.dll, d3d10.dll, etc.

Copy the appropriate opengl32.dll, d3d8.dll, or d3d9.dll from the wrappers directory to the directory with the application you want to trace. Then run the application as usual.

You can specify the written trace filename by setting the TRACE_FILE environment variable before running.

For D3D10 and higher you really must use apitrace trace -a DXGI .... This is because D3D10-11 API span many DLLs which depend on each other, and once a DLL with a given name is loaded Windows will reuse it for LoadLibrary calls of the same name, causing internal calls to be traced erroneously. apitrace trace solves this issue by injecting a DLL dxgitrace.dll and patching all modules to hook only the APIs of interest.

Emitting annotations to the trace

From whitin OpenGL applications you can embed annotations in the trace file through the following extensions:

apitrace will advertise and intercept these OpenGL extensions regardless of whether the OpenGL implementation supports them or not. So all you have to do is to use these extensions when available, and you can be sure they will be available when tracing inside apitrace.

For example, if you use GLEW to dynamically detect and use OpenGL extensions, you could easily accomplish this by doing:

void foo() {

  if (GLEW_KHR_debug) {
    glPushDebugGroup(GL_DEBUG_SOURCE_APPLICATION, 0, -1, __FUNCTION__);
  }

  ...

  if (GLEW_KHR_debug) {
    glDebugMessageInsert(GL_DEBUG_SOURCE_APPLICATION, GL_DEBUG_TYPE_OTHER,
                         0, GL_DEBUG_SEVERITY_MEDIUM, -1, "bla bla");
  }

  ...

  if (GLEW_KHR_debug) {
    glPopDebugGroup();
  }

}

This has the added advantage of working equally well with other OpenGL debugging tools.

Also, provided that the OpenGL implementation supports GL_KHR_debug, labels defined via glObjectLabel() , and the labels of several objects (textures, framebuffers, samplers, etc. ) will appear in the GUI state dumps, in the parameters tab.

For OpenGL ES applications you can embed annotations in the trace file through the GL_KHR_debug or GL_EXT_debug_marker extensions.

For Direct3D applications you can follow the standard procedure for adding user defined events to Visual Studio Graphics Debugger / PIX:

  • D3DPERF_BeginEvent, D3DPERF_EndEvent, and D3DPERF_SetMarker for D3D9 applications.

  • ID3DUserDefinedAnnotation::BeginEvent, ID3DUserDefinedAnnotation::EndEvent, and ID3DUserDefinedAnnotation::SetMarker for D3D11.1 applications.

Dump OpenGL state at a particular call

You can get a dump of the bound OpenGL state at call 12345 by doing:

apitrace replay -D 12345 application.trace > 12345.json

This is precisely the mechanism the GUI uses to obtain its own state.

You can compare two state dumps by doing:

apitrace diff-state 12345.json 67890.json

Comparing two traces side by side

apitrace diff trace1.trace trace2.trace

This works only on Unices, and it will truncate the traces due to performance limitations.

Recording a video with FFmpeg/Libav

You can make a video of the output with FFmpeg by doing

apitrace dump-images -o - application.trace \
| ffmpeg -r 30 -f image2pipe -vcodec ppm -i pipe: -vcodec mpeg4 -y output.mp4

or Libav (which replaces FFmpeg on recent Debian/Ubuntu distros) doing

apitrace dump-images -o - application.trace \
| avconv -r 30 -f image2pipe -vcodec ppm -i - -vcodec mpeg4 -y output.mp4

Recording a video with gstreamer

You can make a video of the output with gstreamer by doing

glretrace --snapshot-format=RGB -s - smokinguns.trace | gst-launch-0.10 fdsrc blocksize=409600 ! queue \
! videoparse format=rgb width=1920 height=1080 ! queue ! ffmpegcolorspace ! queue \
! vaapiupload direct-rendering=0 ! queue ! vaapiencodeh264 ! filesink location=xxx.264

Trimming a trace

You can truncate a trace by doing:

apitrace trim --exact --calls 0-12345 -o trimed.trace application.trace

If you need precise control over which calls to trim you can specify the individual call numbers in a plain text file, as described in the 'Call sets' section above.

There is also experimental support for automatically trimming the calls necessary for a given frame or call:

apitrace trim --auto --calls=12345 -o trimed.trace application.trace
apitrace trim --auto --frames=12345 -o trimed.trace application.trace

Profiling a trace

You can perform gpu and cpu profiling with the command line options:

  • --pgpu record gpu times for frames and draw calls.

  • --pcpu record cpu times for frames and draw calls.

  • --ppd record pixels drawn for each draw call.

The results from these can then be read by hand or analyzed with a script.

scripts/profileshader.py will read the profile results and format them into a table which displays profiling results per shader.

For example, to record all profiling data and utilise the per shader script:

apitrace replay --pgpu --pcpu --ppd foo.trace | ./scripts/profileshader.py

Advanced usage for OpenGL implementors

There are several advanced usage examples meant for OpenGL implementors.

Regression testing

These are the steps to create a regression test-suite around apitrace:

  • obtain a trace

  • obtain reference snapshots, by doing on a reference system:

    mkdir /path/to/reference/snapshots/
    apitrace dump-images -o /path/to/reference/snapshots/ application.trace
    
  • prune the snapshots which are not interesting

  • to do a regression test, use apitrace diff-images:

    apitrace dump-images -o /path/to/test/snapshots/ application.trace
    apitrace diff-images --output summary.html /path/to/reference/snapshots/ /path/to/test/snapshots/
    

Automated git-bisection

With tracecheck.py it is possible to automate git bisect and pinpoint the commit responsible for a regression.

Below is an example of using tracecheck.py to bisect a regression in the Mesa-based Intel 965 driver. But the procedure could be applied to any OpenGL driver hosted on a git repository.

First, create a build script, named build-script.sh, containing:

#!/bin/sh
set -e
export PATH=/usr/lib/ccache:$PATH
export CFLAGS='-g'
export CXXFLAGS='-g'
./autogen.sh --disable-egl --disable-gallium --disable-glut --disable-glu --disable-glw --with-dri-drivers=i965
make clean
make "$@"

It is important that builds are both robust, and efficient. Due to broken dependency discovery in Mesa's makefile system, it was necessary to invoke make clean in every iteration step. ccache should be installed to avoid recompiling unchanged source files.

Then do:

cd /path/to/mesa
export LIBGL_DEBUG=verbose
export LD_LIBRARY_PATH=$PWD/lib
export LIBGL_DRIVERS_DIR=$PWD/lib
git bisect start \
    6491e9593d5cbc5644eb02593a2f562447efdcbb 71acbb54f49089b03d3498b6f88c1681d3f649ac \
    -- src/mesa/drivers/dri/intel src/mesa/drivers/dri/i965/
git bisect run /path/to/tracecheck.py \
    --precision-threshold 8.0 \
    --build /path/to/build-script.sh \
    --gl-renderer '.*Mesa.*Intel.*' \
    --retrace=/path/to/glretrace \
    -c /path/to/reference/snapshots/ \
    topogun-1.06-orc-84k.trace

The trace-check.py script will skip automatically when there are build failures.

The --gl-renderer option will also cause a commit to be skipped if the GL_RENDERER is unexpected (e.g., when a software renderer or another OpenGL driver is unintentionally loaded due to a missing symbol in the DRI driver, or another runtime fault).

Side by side retracing

In order to determine which draw call a regression first manifests one could generate snapshots for every draw call, using the -S option. That is, however, very inefficient for big traces with many draw calls.

A faster approach is to run both the bad and a good OpenGL driver side-by-side. The latter can be either a previously known good build of the OpenGL driver, or a reference software renderer.

This can be achieved with retracediff.py script, which invokes glretrace with different environments, allowing to choose the desired OpenGL driver by manipulating variables such as LD_LIBRARY_PATH, LIBGL_DRIVERS_DIR, or TRACE_LIBGL.

For example, on Linux:

./scripts/retracediff.py \
    --ref-env LD_LIBRARY_PATH=/path/to/reference/OpenGL/implementation \
    --retrace /path/to/glretrace \
    --diff-prefix=/path/to/output/diffs \
    application.trace

Or on Windows:

python scripts\retracediff.py --retrace \path\to\glretrace.exe --ref-env TRACE_LIBGL=\path\to\reference\opengl32.dll application.trace

Advanced GUI usage

qapitrace has rudimentary support for replaying traces on a remote target device. This can be useful, for example, when developing for an embedded system. The primary GUI will run on the local host, while any replays will be performed on the target device.

In order to target a remote device, use the command-line:

qapitrace --remote-target <HOST> <trace-file>

In order for this to work, the following must be available in the system configuration:

  1. It must be possible for the current user to initiate an ssh session that has access to the target's window system. The command to be exectuted by qapitrace will be:

    ssh <HOST> glretrace
    

    For example, if the target device is using the X window system, one can test whether an ssh session has access to the target X server with:

    ssh <HOST> xdpyinfo
    

    If this command fails with something like "cannot open display" then the user will have to configure the target to set the DISPLAY environment variable, (for example, setting DISPLAY=:0 in the .bashrc file on the target or similar).

    Also, note that if the ssh session requires a custom username, then this must be configured on the host side so that ssh can be initiated without a username.

    For example, if you normally connect with ssh user@192.168.0.2 you could configure ~/.ssh/config on the host with a block such as:

    Host target
      HostName 192.168.0.2
      User user
    

    And after this you should be able to connect with ssh target so that you can also use qapitrace --remote-target target.

  2. The target host must have a functional glretrace binary available

  3. The target host must have access to at the same path in the filesystem as the path on the host system being passed to the qapitrace command line.

githalytics.com alpha

Something went wrong with that request. Please try again.