X-Git-Url: http://git.megacz.com/?a=blobdiff_plain;f=docs%2Fusers_guide%2Fruntime_control.xml;h=c482a28a91d3f3d4dfb6b6089681176015e4dd99;hb=65277a1c9ff86c28c656849d6f6cbb392f1eb3e7;hp=69e26bc9208ea67c35b1a81d01e065f05a459324;hpb=9ee105bd1f977a8fd68281e7658383d5a0c86156;p=ghc-hetmet.git diff --git a/docs/users_guide/runtime_control.xml b/docs/users_guide/runtime_control.xml index 69e26bc..c482a28 100644 --- a/docs/users_guide/runtime_control.xml +++ b/docs/users_guide/runtime_control.xml @@ -298,51 +298,58 @@ - - RTS + + RTS option - [New in GHC 6.12.1] Disable the parallel GC. - The parallel GC is turned on automatically when parallel - execution is enabled with the option; - this option is available to turn it off if - necessary. + [New in GHC 6.12.1] [Default: 0] + Use parallel GC in + generation gen and higher. + Omitting gen turns off the + parallel GC completely, reverting to sequential GC. - Experiments have shown that parallel GC usually - results in a performance improvement given 3 cores or - more; with 2 cores it may or may not be beneficial, - depending on the workload. Bigger heaps work better with - parallel GC, so set your value high (3 - or more times the maximum residency). Look at the timing - stats with to see whether you're - getting any benefit from parallel GC or not. If you find - parallel GC is significantly slower - (in elapsed time) than sequential GC, please report it as - a bug. - - In GHC 6.10.1 it was possible to use a different - number of threads for GC than for execution, because the GC - used its own pool of threads. Now, the GC uses the same - threads as the mutator (for executing the program). + The default parallel GC settings are usually suitable + for parallel programs (i.e. those + using par, Strategies, or with multiple + threads). However, it is sometimes beneficial to enable + the parallel GC for a single-threaded sequential program + too, especially if the program has a large amount of heap + data and GC is a significant fraction of runtime. To use + the parallel GC in a sequential program, enable the + parallel runtime with a suitable -N + option, and additionally it might be beneficial to + restrict parallel GC to the old generation + with -qg1. - - RTS + + RTS option - [Default: 1] [New in GHC 6.12.1] - Enable the parallel GC only in - generation n and greater. - Parallel GC is often not worthwhile for collections in - generation 0 (the young generation), so it is enabled by - default only for collections in generation 1 (and higher, - if applicable). + [New in GHC 6.12.1] [Default: 1] Use + load-balancing in the parallel GC in + generation gen and higher. + Omitting gen disables + load-balancing entirely. + + + Load-balancing shares out the work of GC between the + available cores. This is a good idea when the heap is + large and we need to parallelise the GC work, however it + is also pessimal for the short young-generation + collections in a parallel program, because it can harm + locality by moving data from the cache of the CPU where is + it being used to the cache of another CPU. Hence the + default is to do load-balancing only in the + old-generation. In fact, for a parallel program it is + sometimes beneficial to disable load-balancing entirely + with -qb. @@ -810,6 +817,92 @@ + + Tracing + + tracing + events + eventlog files + + + When the program is linked with the + option (), runtime events can + be logged in two ways: + + + + + + In binary format to a file for later analysis by a + variety of tools. One such tool + is ThreadScopeThreadScope, + which interprets the event log to produce a visual parallel + execution profile of the program. + + + + + As text to standard output, for debugging purposes. + + + + + + + + + RTS option + + + + Log events in binary format to the + file program.eventlog, + where type indicates the type + of events to log. Currently there is only one type + supported: -ls, for scheduler events. + + + + The format of the log file is described by the header + EventLogFormat.h that comes with + GHC, and it can be parsed in Haskell using + the ghc-events + library. To dump the contents of + a .eventlog file as text, use the + tool show-ghc-events that comes with + the ghc-events + package. + + + + + + + + RTS option + + + + Log events as text to standard output, instead of to + the .eventlog file. + + + + + + + + The debugging + options also + generate events which are logged using the tracing framework. + By default those events are dumped as text to stdout + ( + implies ), but they may instead be stored in + the binary eventlog file by using the + option. + + + RTS options for hackers, debuggers, and over-interested souls @@ -846,14 +939,28 @@ - num + x -DRTS option - An RTS debugging flag; varying quantities of output - depending on which bits are set in - num. Only works if the RTS was - compiled with the option. + + An RTS debugging flag; only availble if the program was + linked with the option. Various + values of x are provided to + enable debug messages and additional runtime sanity checks + in different subsystems in the RTS, for + example +RTS -Ds -RTS enables debug + messages from the scheduler. + Use +RTS -? to find out which + debug flags are supported. + + + + Debug messages will be sent to the binary event log file + instead of stdout if the option is + added. This might be useful for reducing the overhead of + debug tracing. + @@ -866,20 +973,13 @@ Produce “ticky-ticky” statistics at the - end of the program run. The file - business works just like on the RTS - option (above). - - “Ticky-ticky” statistics are counts of - various program actions (updates, enters, etc.) The program - must have been compiled using - - (a.k.a. “ticky-ticky profiling”), and, for it to - be really useful, linked with suitable system libraries. - Not a trivial undertaking: consult the installation guide on - how to set things up for easy “ticky-ticky” - profiling. For more information, see . + end of the program run (only available if the program was + linked with ). + The file business works just like + on the RTS option, above. + + For more information on ticky-ticky profiling, see + .