NEW version of par2, DUAL/QUAD core released, 4/20/2008

[ericab]

http://chuchusoft.com/par2_tbb/index.html

[pilotsnipes]

Is this as easy as just dropping the new file into the altbinz directory?

[Hecks]

The previous multicore version was already modified by Rdl to prevent disk thrashing, and corrected faulty PAR2 count (version 0.4d).  I'd wait for Rdl to issue another custom version for Alt.Binz, if it's even necessary.

[mc00]

to be honest, I don't see any difference between single core and dual core with altbinz. I tested with a single core on my laptop and dual core. turion 64 2.2ghz and a  turion 64 x2 tl 52. I guess laptop are different from desktop or my laptop just suck lol. .

[ericab]

guess so.

there is a huge difference. huge.

[Rdl]

guess so.

there is a huge difference. huge.

I don't see a reason to recompile/support this version. Nothing new.

[Rdl]

Also be warned. This version has, like every previous, strange bug while creating pars. Occasionaly it will create bad pars. Reproducable in 10% of cases on quad, very rare on dual cores.

[DM8Mydog]

there is a huge difference. huge.

PAR2 checking is hardly CPU intensive.
Checking is mostly limited to what average speed your HDD can pull.
I'm also not so sure Repairing something is that much cpu intensive either, it appears to me that it's also relative to HDD speed (faster repairing on my RAID machine than on my single drive machine).
Where a multi-threaded oriented build of par2 would show huge benefit is par2 block creation, which is very CPU intensive.

Take your pick:
There's not enough incentive at this time to implement a new multi thread build of par2 in Alt.Binz cause Alt.Binz doesn't make pars, it only checks & repairs.
-or-
It's a fresh build with a few major changes on it, give it a few weeks/months before it becomes stable and has bugs ironed out before considering using it embedded in something like alt.binz.