search

idkwim / ouspg

Automatically exported from code.google.com/p/ouspg
3 stars 0 forks source link

Radamsa can't fuzz while it fuzz #77

Closed GoogleCodeExporter closed 9 years ago

GoogleCodeExporter commented 9 years ago 
What steps will reproduce the problem?
1. $ yes trolololo | head -n 100000000 > dawg
2. $ cat dawg | radamsa | pv > /dev/null
3. $ cat dawg | radamsa | radamsa | pv > /dev/null

What is the expected output? What do you see instead?
both should push data at >100MB/s, bu the latter drops to 0B/s after a few 
kilobytes.

Please use labels and text to provide additional information.

Original issue reported on code.google.com by aohelin on 30 May 2012 at 7:08

GoogleCodeExporter commented 9 years ago 
Tracked down to IO changes (dumbifications) made to make owl and radamsa work 
in windows. There are small write- and read attempt windows, which causes 
exponential slowdown when owl programs are chained together. Could be worked 
around in radamsa, but this is really an owl regression.

Recommended solution: bring back sane IO and scr^W depend on an emulation dll 
on windows.

Original comment by aohelin on 30 May 2012 at 8:38