#4928: Add primops for copying/cloning an array
------------------------------+---------------------------------------------
Reporter: tibbe | Owner: simonmar
Type: feature request | Status: patch
Priority: normal | Milestone: 7.4.1
Component: Runtime System | Version: 7.0.1
Resolution: | Keywords:
Testcase: | Blockedby:
Difficulty: | Os: Unknown/Multiple
Blocking: | Architecture: Unknown/Multiple
Failure: None/Unknown |
------------------------------+---------------------------------------------
Comment(by tibbe):
I've attached the beginnings of a benchmark harness. For now it only
exercises `copyArray#` and `thawArray#`. Usage:
{{{
$ ./inplace/bin/ghc-stage2 -O2 ArrayCopyBench.hs -o bench
$ ./bench
Usage: bench BENCHMARK ARRAYSIZE ITERS
$ ./bench thawArray# 32 1000000
thawArray#: 0.040091s
}}}
Using this benchmark and Linux's perf events we now have a better idea of
where the time is spent:
{{{
$ perf record -f ./bench thawArray# 32 10000000
thawArray#: 0.408859s
[ perf record: Woken up 1 times to write data ]
[ perf record: Captured and wrote 0.095 MB perf.data (~4160 samples) ]
$ perf report
# Samples: 4100
#
# Overhead Command Shared Object Symbol
# ........ ............... ................... ......
#
37.05% bench /lib/libc-2.11.1.so [.] memcpy
36.78% bench ./bench [.] stg_thawArrayzh
6.56% bench /lib/libc-2.11.1.so [.] __GI_memset
5.27% bench ./bench [.] allocate
4.34% bench ./bench [.] s224_info
2.02% bench ./bench [.] s24C_info
1.17% bench ./bench [.] memcpy@plt
1.05% bench ./bench [.] memset@plt
1.02% bench ./bench [.] clearNurseries
0.27% bench [kernel] [k] sigprocmask
}}}
--
Ticket URL: <http://hackage.haskell.org/trac/ghc/ticket/4928#comment:19>
GHC <http://www.haskell.org/ghc/>
The Glasgow Haskell Compiler
_______________________________________________
Glasgow-haskell-bugs mailing list
[email protected]
http://www.haskell.org/mailman/listinfo/glasgow-haskell-bugs
