76-fold regression GHC 7.10->7.11 in T9961 byte-allocation

Herbert Valerio Riedel hvriedel at gmail.com
Fri Jan 30 11:02:30 UTC 2015


Hello *,

I noticed something odd while validating the GHC 7.10 branch:

  bytes allocated value is too low:
  (If this is because you have improved GHC, please
  update the test so that GHC doesn't regress again)
      Expected    T9961(normal) bytes allocated: 772510192 +/-5%
      Lower bound T9961(normal) bytes allocated: 733884682 
      Upper bound T9961(normal) bytes allocated: 811135702 
      Actual      T9961(normal) bytes allocated:   9766160 
      Deviation   T9961(normal) bytes allocated:     -98.7 %
  *** unexpected stat test failure for T9961(normal)


...then I also ran ./validate against today's GHC HEAD, and re-ran the
T9961 test:

    Expected    T9961(normal) bytes allocated: 772510192 +/-5%
    Lower bound T9961(normal) bytes allocated: 733884682 
    Upper bound T9961(normal) bytes allocated: 811135702 
    Actual      T9961(normal) bytes allocated: 748225848 
    Deviation   T9961(normal) bytes allocated:      -3.1 %


I'm not sure if it's just the test-case being broken, or there's
something real regression between 7.10 and HEAD...

However, I don't have time to investigate this.

Cheers,
  hvr


More information about the ghc-devs mailing list