PAR2 Multithreaded Archive Recovery Performance

Par2 is an application used for reconstructing downloaded archives. It can generate parity data from a given archive and later use it to recover the archive

Chuchusoft took the source code of par2cmdline 0.4 and parallelized it using Intel’s Threading Building Blocks 2.1. The result is a version of par2cmdline that can spawn multiple threads to repair par2 archives. For this test we took a 708MB archive, corrupted nearly 60MB of it, and used the multithreaded par2cmdline to recover it. The scores reported are the repair and recover time in seconds.

Data Recovery - par2cmdline 0.4 Multithreaded

WinRAR - Archive Creation

Our WinRAR test simply takes 300MB of files and compresses them into a single RAR archive using the application's default settings. We're not doing anything exotic here, just looking at the impact of CPU performance on creating an archive:

WinRAR 3.8 Compression - 300MB Archive

3D Rendering Performance Excel & Blu-ray/Flash Creation Performance
Comments Locked

177 Comments

View All Comments

  • Fleeb - Tuesday, October 20, 2009 - link

    Or maybe, these three are just the same person craving for attention in real life he cannot have. Do not hate the guy. Pity him.
  • mapesdhs - Tuesday, October 20, 2009 - link


    Sorry for the double post! The submission form just gave an error
    the first time round, but I guess it went through anyway. Anand,
    please feel free to delete this post and my previous duplicate.

    Ian.

  • mapesdhs - Tuesday, October 20, 2009 - link


    Anand, just curious, the test platform description includes mention
    of an X58 mbd, yet there are no i7 results in the tables. How come?
    Then again, including a couple of data points from a P55 with an
    i5 750 and i7 860 would be more useful. AMD wins on price by a mile
    of course (personally I reckon the 620 is the best buy much of the
    time) but for those occasions where only 1 or 2 threads are running,
    the i5 750 might win on price/performance. If it's deemed appropriate
    to include a top-end Ph2 in the results, then surely at the very
    least the i5 750 should have been included aswell just to put things
    into perspective? I would include an 860 aswell just to show where
    the curves are heading on the Intel side, but nothing above that.

    And btw no, I don't agree with anything maddoctor says. Speaking of
    which, can you please ban the guy? Once again the discussion section
    of an otherwise interesting article is just being filled up with
    junk. To everyone else: please don't reply to his posts, you're just
    making it worse. It's a sad fact of nature that half the population
    have got to be below average. Who _are_ these people? Sheesh, I can
    almost hear the banjo, da da ding ding ding... :D

    Ian.

    PS. One other thing Anand, have you ever tested how high the 620
    can be oc'd with a *good* air cooler? I know the retail AMD cooler
    allows it to reach 3.25, but what about with something better? Someone
    mentioned the Coolermaster Hyper TX2 as being a suitable alternative.

  • rupa - Tuesday, October 20, 2009 - link

    hi ... my 620 is stable till 3.380 (260x13) - noctua nh-u12p default vcore x64 asus m3a78t

  • mapesdhs - Tuesday, October 20, 2009 - link


    Anand, just curious, the test platform description includes mention
    of an X58 mbd, yet there are no i7 results in the tables. How come?
    Then again, including a couple of data points from a P55 with an
    i5 750 and i7 860 would be more useful. AMD wins on price by a mile
    of course (personally I reckon the 620 is the best buy much of the
    time) but for those occasions where only 1 or 2 threads are running,
    the i5 750 might win on price/performance. If it's deemed appropriate
    to include a top-end Ph2 in the results, then surely at the very
    least the i5 750 should have been included aswell just to put things
    into perspective? I would include an 860 aswell just to show where
    the curves are heading on the Intel side, but nothing above that.

    And btw no, I don't agree with anything maddoctor says. Speaking of
    which, can you please ban the guy? Once again the discussion section
    of an otherwise interesting article is just being filled up with
    junk. To everyone else: please don't reply to his posts, you're just
    making it worse. It's a sad fact of nature that half the population
    have got to be below average. Who _are_ these people? Sheesh, I can
    almost hear the banjo, da da ding ding ding... :D

    Ian.

    PS. One other thing Anand, have you ever tested how high the 620
    can be oc'd with a *good* air cooler? I know the retail AMD cooler
    allows it to reach 3.25, but what about with something better? Someone
    mentioned the Coolermaster Hyper TX2 as being a suitable alternative.

  • Zool - Tuesday, October 20, 2009 - link

    For a fast compare with other procesors the test setup is not bad but for some people some the benchmarks could be misleading.
    People will not buy these cpu-s just to put them together with a intel SSD drive and a gtx280.
    I think it wouldnt take much longer to test it with average hdd and a sub 100$ gpu. I would care much less if i cant compare it to other anad tests with high end cpus.
  • Zool - Tuesday, October 20, 2009 - link

    For a fast compare with other procesors the test setup is not bad but for some people some the benchmarks could be misleading.
    People will not buy these cpu-s just to put them together with a intel SSD drive and a gtx280.
    I think it wouldnt take much longer to test it with average hdd and a sub 100$ gpu. I would care much less if i cant compare it to other anad tests with high end cpus.
  • 7Enigma - Tuesday, October 20, 2009 - link

    Zool,

    The purpose of using an SSD is strictly for the variability between tests. What Anand's site never shows (and one of my major complaints) is % error. In any statistical measurement you always present the amount of error in a test. What this might show is there is no REAL clear winner, or very little difference as when you get close (say within 5%), and you have a large variability (say due to a standard HD or run-to-run variability), the numbers become moot.

    I will agree with you on the GTX280, however, as that is a pointless component for this price sector.
  • 7Enigma - Tuesday, October 20, 2009 - link

    As an example I went back and looked at the game data. The Fallout3 data is generated MANUALLY by running through an area and using FRAPS. The X2/X3 chips are bunched up with less than 1 fps between them. I would wager a hefty sum that the % error in this test is greater than 1fps and so any chips within that range are EQUAL. Same goes for Left 4 Dead...
  • maddoctor - Tuesday, October 20, 2009 - link

    I don't see anything wrong for the benchmark setup, but other SSD products are not competitive and have lower performance than Intel products. The most wrong thing about is Anand is compare AMD rubbish product to Intel products.

Log in

Don't have an account? Sign up now