Need help with CommandBuffer?
Click the “chat” button below for chat support from the developer who created it, or find similar developers for support.

About the developer

142 Stars 8 Forks 34 Commits 0 Opened issues


A lock-free CommandBuffer implementation designed for multi-threaded rendering applications.

Services available


Need anything else?

Contributors list

No Data

Lock-free CommandBuffer

Windows status Language License

A lock-free CommandBuffer implementation designed for multi-threaded rendering applications. Useful to reduce state switches(i.e. materials, shaders), depth sorting(for transparency or hardware early-z rejection) and also for dispatching commands from multiple threads.


  • lock-free, designed for high-congestion
  • graphics API agnostic(see cb::RenderContext)
  • fast and configurable allocation via a linear allocator
  • optional material binder with multiple material passes support
  • chainable/appendable commands
  • configurable key type for sorting of commands(opaque, transparent, depth sorting)
  • easy to use and configurable draw key via bitfields
  • debug utilities, tag commands
  • basic GL commands implementation(see GLCommands.h)
  • lightweight, header only


The implementation is header only, except GL commands, requires at least C++11 support.


Creating a command:

    struct DrawArrays {
        static const cb::RenderContext::function_t kDispatchFunction;
        GLuint          vao;
        uint32_t        base;
        uint32_t        count;
        GLenum          primitive;
    //in cpp
    void drawArrays(const void* data, cb::RenderContext*) {
        auto& cmd = *reinterpret_cast(data);
        glDrawArrays(cmd.primitive, cmd.base, cmd.count);
    const cb::RenderContext::function_t DrawArrays::kDispatchFunction = &drawArrays;
Dispatching a command: ```cpp //create the command key first cb::DrawKey key(0); key.setViewLayer(cb::ViewLayerType::e3D, cb::TranslucencyType::eOpaque); key.setDepth(depth); key.setMaterial(materialId);
//then create command which will dispatch it
cmds::DrawArrays& cmd = *commandBuffer.addCommand<:drawarrays>(key);
cmd.vao = vaoId;
cmd.base = 0;
cmd.count = 4;
cmd.primitive = GL_TRIANGLE_STRIP;
Tagging a command for debug purposes, use CB_DEBUG_TAG_COMMANDS in config.h to enable/disable tagging:
void testFunction() {
    cb::DrawKey key = cb::DrawKey::makeDefault(viewportId, cb::ViewLayerType::e3D);
    cmds::DrawArrays& cmd = *commandBuffer.addCommand<:drawarrays>(key);
    // will name tag: 'testFunction :  : struct cmds::DrawArrays * __ptr64'
    // will name tag: 'testFunction :  : struct cmds::DrawArrays * __ptr64 : draw quad'
    CB_DEBUG_COMMAND_TAG_MSG(cmd, "draw quad");
    // will name tag: 'draw quad'
    CB_DEBUG_COMMAND_SET_MSG(cmd, "draw quad");

NOTE. Can enable/disable logging of the commands via CBDEBUGCOMMANDS_PRINT in config.h which is enabled by default.

Appending/chaining commands(useful to reduce overhead of redundant material bindings): ```cpp cb::DrawKey key = cb::DrawKey::makeDefault(viewportId, cb::ViewLayerType::e3D); //setup key cmds::DrawArrays* cmd = commandBuffer.addCommandcmds::DrawArrays(key); //fill command data CBDEBUGCOMMANDSETMSG(cmd, "draw first quad");

//append multiple draw calls
for(int i = 1; i < 10; ++i) {
    cmd = commandBuffer.appendCommand<:drawarrays>(cmd);
    //fill command data
    CB_DEBUG_COMMAND_SET_MSG(cmd, FORMAT_STR("draw quad %d", i));
Sometimes you may want to share/reference a command so you don't copy it's data multiple times:
    //create the first the shared command
    CommandPacket* headerPacket = buffer.createCommandPacketData(matrices);

//reference/share the command
 key = cb::DrawKey::makeDefault(viewportId, cb::ViewLayerType::e3D);
CommandPacket* cmd = buffer.addCommandFrom(key, headerPacket);
//append a new command to the shared one
cmds::DrawArrays* draw = buffer.appendCommand<:drawarrays>(cmd);
//fill command data

//reference it again
key = cb::DrawKey::makeDefault(otherViewportId, cb::ViewLayerType::e3D);
cmd = buffer.addCommandFrom(key, headerPacket);
draw = buffer.appendCommand<:drawarrays>(cmd);
//fill command data


NOTE. Since matrices are PODs you can use data copy commands which will automatically allocate auxiliary memory and copy it.


Check the example folder which shows how to use the CommandBuffer in a real use case scenario with more advanced usage, it was done by adapting NVIDIA's Gameworks GL Threading example to a deferred renderer.


The implementation is mainly based on Stefan Reinalter's blog post at Molecular Matters, about a 'Stateless, layered, multi-threaded rendering' which I highly recommend reading.

References: - Stateless, layered, multi-threaded rendering by Stefan Reinalter - Order your graphics draw calls around! by Christer Ericson

Bug reports and pull requests are welcome on GitHub at


The code is available as open source under the terms of the MIT License.

We use cookies. If you continue to browse the site, you agree to the use of cookies. For more information on our use of cookies please see our Privacy Policy.