# HG changeset patch # User Steve Borho <[email protected]> # Date 1439526505 -19800 # Fri Aug 14 09:58:25 2015 +0530 # Branch stable # Node ID 1b8ff853ad41b3a033b3c98213e9e51f37401b3e # Parent 567f9b1263a27b9384f10662c169615e62b8e2b3 doc: fix typos and improve readability of threading chapter
diff -r 567f9b1263a2 -r 1b8ff853ad41 doc/reST/threading.rst --- a/doc/reST/threading.rst Fri Aug 14 09:06:13 2015 +0530 +++ b/doc/reST/threading.rst Fri Aug 14 09:58:25 2015 +0530 @@ -28,7 +28,7 @@ providers are recommended to call this method when they make new jobs available. -Worker jobs are not allowed to block except when abosultely necessary +Worker jobs are not allowed to block except when absolutely necessary for data locking. If a job becomes blocked, the work function is expected to drop that job so the worker thread may go back to the pool and find more work. @@ -94,10 +94,10 @@ If a worker thread job has work which can be performed in parallel by many threads, it may allocate a bonded task group and enlist the help of -other idle worker threads in the same pool. Those threads will cooperate -to complete the work of the bonded task group and then return to their -idle states. The larger and more uniform those tasks are, the better the -bonded task group will perform. +other idle worker threads from the same thread pool. Those threads will +cooperate to complete the work of the bonded task group and then return +to their idle states. The larger and more uniform those tasks are, the +better the bonded task group will perform. Parallel Mode Analysis ~~~~~~~~~~~~~~~~~~~~~~ @@ -105,19 +105,20 @@ When :option:`--pmode` is enabled, each CU (at all depths from 64x64 to 8x8) will distribute its analysis work to the thread pool via a bonded task group. Each analysis job will measure the cost of one prediction -for the CU: merge, skip, intra, inter (2Nx2N, Nx2N, 2NxN, and AMP). At -slower presets, the amount of increased parallelism is often enough to -be able to reduce frame parallelism while achieving the same overall CPU -utilization. Reducing frame threads is often beneficial to ABR and VBV -rate control. +for the CU: merge, skip, intra, inter (2Nx2N, Nx2N, 2NxN, and AMP). + +At slower presets, the amount of increased parallelism from pmode is +often enough to be able to reduce or disable frame parallelism while +achieving the same overall CPU utilization. Reducing frame threads is +often beneficial to ABR and VBV rate control. Parallel Motion Estimation ~~~~~~~~~~~~~~~~~~~~~~~~~~ When :option:`--pme` is enabled all of the analysis functions which perform motion searches to reference frames will distribute those motion -searches as jobs for worker threads via a bonded task group (if more -than two motion searches are required). +searches to other worker threads via a bonded task group (if more than +two motion searches are required). Frame Threading =============== @@ -241,7 +242,7 @@ bonded task groups to measure single frame cost estimates using slices. (see :option:`--lookahead-slices`) -The function slicetypeDecide() itself is also be performed by a worker +The main slicetypeDecide() function itself is also performed by a worker thread if your encoder has a thread pool, else it runs within the context of the thread which calls the x265_encoder_encode(). _______________________________________________ x265-devel mailing list [email protected] https://mailman.videolan.org/listinfo/x265-devel
