skia2/dm
2013-10-24 17:44:43 +00:00
..
DM.cpp Fix DM build failures: 2013-10-16 18:21:03 +00:00
DMComparisonTask.cpp DM: add --serialize 2013-10-21 18:40:25 +00:00
DMComparisonTask.h dm is like gm, but faster and with fewer features. 2013-10-16 13:02:15 +00:00
DMCpuTask.cpp DM: add --serialize 2013-10-21 18:40:25 +00:00
DMCpuTask.h dm is like gm, but faster and with fewer features. 2013-10-16 13:02:15 +00:00
DMGpuTask.cpp DM: print cache stats after every GPU GM when GR_CACHE_STATS is set 2013-10-24 17:44:43 +00:00
DMGpuTask.h dm is like gm, but faster and with fewer features. 2013-10-16 13:02:15 +00:00
DMReplayTask.cpp DM: add --serialize 2013-10-21 18:40:25 +00:00
DMReplayTask.h DM: add --serialize 2013-10-21 18:40:25 +00:00
DMReporter.cpp DM: add --writePath 2013-10-18 20:52:44 +00:00
DMReporter.h dm is like gm, but faster and with fewer features. 2013-10-16 13:02:15 +00:00
DMSerializeTask.cpp DM: add --serialize 2013-10-21 18:40:25 +00:00
DMSerializeTask.h DM: add --serialize 2013-10-21 18:40:25 +00:00
DMTask.cpp dm is like gm, but faster and with fewer features. 2013-10-16 13:02:15 +00:00
DMTask.h dm is like gm, but faster and with fewer features. 2013-10-16 13:02:15 +00:00
DMTaskRunner.cpp dm is like gm, but faster and with fewer features. 2013-10-16 13:02:15 +00:00
DMTaskRunner.h dm is like gm, but faster and with fewer features. 2013-10-16 13:02:15 +00:00
DMUtil.cpp DM: add --serialize 2013-10-21 18:40:25 +00:00
DMUtil.h DM: add --serialize 2013-10-21 18:40:25 +00:00
DMWriteTask.cpp DM: add --serialize 2013-10-21 18:40:25 +00:00
DMWriteTask.h DM: add --writePath 2013-10-18 20:52:44 +00:00
README DM: add --serialize 2013-10-21 18:40:25 +00:00

DM is like GM, but multithreaded.  It doesn't do everything GM does yet.

Current approximate list of missing features:
  --mismatchPath
  --missingExpectationsPath
  --writePicturePath

  --deferred / --pipe
  --rtree
  --tiledGrid


DM's design is based around Tasks and a TaskRunner.

A Task represents an independent unit of work that might fail.  We make a task
for each GM/configuration pair we want to run.  Tasks can kick off new tasks
themselves.  For example, a CpuTask can kick off a ReplayTask to make sure
recording and playing back an SkPicture gives the same result as direct
rendering.

The TaskRunner runs all tasks on one of two threadpools, whose sizes are
configurable by --cpuThreads and --gpuThreads.  Ideally we'd run these on a
single threadpool but it can swamp the GPU if we shove too much work into it at
once.  --cpuThreads defaults to the number of cores on the machine.
--gpuThreads defaults to 1, but you may find 2 or 4 runs a little faster.

So the main flow of DM is:

    for each GM:
        for each configuration:
            kick off a new task
    < tasks run, maybe fail, and maybe kick off new tasks >
    wait for all tasks to finish
    report failures