NameDateSize

..02-Jul-20154 KiB

Android.mk02-Jul-201515.2 KiB

DM.cpp02-Jul-201510.4 KiB

DMBenchTask.cpp02-Jul-20152.7 KiB

DMBenchTask.h02-Jul-20151.7 KiB

DMCpuGMTask.cpp02-Jul-20152.1 KiB

DMCpuGMTask.h02-Jul-20151 KiB

DMExpectations.h02-Jul-20151.2 KiB

DMExpectationsTask.cpp02-Jul-2015547

DMExpectationsTask.h02-Jul-2015772

DMGpuGMTask.cpp02-Jul-20151.7 KiB

DMGpuGMTask.h02-Jul-20151 KiB

DMGpuSupport.h02-Jul-20151.5 KiB

DMPDFRasterizeTask.cpp02-Jul-2015897

DMPDFRasterizeTask.h02-Jul-2015942

DMPDFTask.cpp02-Jul-20152.9 KiB

DMPDFTask.h02-Jul-20151.1 KiB

DMPipeTask.cpp02-Jul-20152.3 KiB

DMPipeTask.h02-Jul-20151 KiB

DMQuiltTask.cpp02-Jul-20152 KiB

DMQuiltTask.h02-Jul-2015901

DMRecordTask.cpp02-Jul-20151.6 KiB

DMRecordTask.h02-Jul-2015937

DMReplayTask.cpp02-Jul-20151.4 KiB

DMReplayTask.h02-Jul-20151,006

DMReporter.cpp02-Jul-20151.2 KiB

DMReporter.h02-Jul-2015839

DMSerializeTask.cpp02-Jul-20151.2 KiB

DMSerializeTask.h02-Jul-2015753

DMSKPTask.cpp02-Jul-2015791

DMSKPTask.h02-Jul-2015686

DMTask.cpp02-Jul-20152.3 KiB

DMTask.h02-Jul-20151.7 KiB

DMTaskRunner.cpp02-Jul-2015734

DMTaskRunner.h02-Jul-2015719

DMTestTask.cpp02-Jul-20151.9 KiB

DMTestTask.h02-Jul-20151.5 KiB

DMUtil.cpp02-Jul-20153.2 KiB

DMUtil.h02-Jul-20151.2 KiB

DMWriteTask.cpp02-Jul-20156.5 KiB

DMWriteTask.h02-Jul-20151.3 KiB

README02-Jul-20151 KiB

README

1DM is like GM, but multithreaded.  It doesn't do everything GM does.
2
3DM's design is based around Tasks and a TaskRunner.
4
5A Task represents an independent unit of work that might fail.  We make a task
6for each GM/configuration pair we want to run.  Tasks can kick off new tasks
7themselves.  For example, a CpuTask can kick off a ReplayTask to make sure
8recording and playing back an SkPicture gives the same result as direct
9rendering.
10
11The TaskRunner runs all tasks on one of two threadpools, whose sizes are
12configurable by --cpuThreads and --gpuThreads.  Ideally we'd run these on a
13single threadpool but it can swamp the GPU if we shove too much work into it at
14once.  --cpuThreads defaults to the number of cores on the machine.
15--gpuThreads defaults to 1, but you may find 2 or 4 runs a little faster.
16
17So the main flow of DM is:
18
19    for each GM:
20        for each configuration:
21            kick off a new task
22    < tasks run, maybe fail, and maybe kick off new tasks >
23    wait for all tasks to finish
24    report failures
25
26