curl vs bytestring

[Ian Lynagh]

This is quite a dull mail; I've just been checking that copying a local
file with curl isn't significantly slower than using bytestring.
Actually it seems to be slightly faster, if anything. So, no problems
with using curl for everything, which is good news, but not a great
surprise. Anyway, source is attached should anyone be interested. The
file ./data is 1001767061 bytes of /dev/urandom.

Some timings:

With everything in cache:

bs   all         11.705      11.808      11.169
curl all         12.233      14.076      12.480
bs   bits       131.398     132.917     130.922
curl bits       129.153     128.533     130.186
bs   mostbits   131.478     130.558     130.424
curl mostbits   130.386     127.471     129.806

With cache flushed (echo 3 | /proc/sys/vm/drop_caches):

bs   all         29.283      30.789      26.640
curl all         28.175      30.762      27.802
bs   bits       140.743     140.542     139.717
curl bits       139.114     135.823     135.216
bs   mostbits   140.570     140.305     138.657
curl mostbits   137.126     137.847     136.557


Thanks
Ian

-------------- next part --------------
A non-text attachment was scrubbed...
Name: dl.hs
Type: text/x-haskell
Size: 2620 bytes
Desc: not available
Url : http://projects.haskell.org/pipermail/camp/attachments/20090214/7328b698/attachment.hs 
-------------- next part --------------
A non-text attachment was scrubbed...
Name: mkinputs.hs
Type: text/x-haskell
Size: 792 bytes
Desc: not available
Url : http://projects.haskell.org/pipermail/camp/attachments/20090214/7328b698/attachment-0001.hs