I found that the runtime stats output was incorrect for the one-line stats
(i.e. the +RTS -t -RTS output). The output was giving zeros for several of the
times, for example the mutator time. The problem was that the time values were
being computed inside a conditional so they were not correctly set for the
one-line outputs, but they were set for the more verbose outputs (i.e. -s).
I wrote a small patch that fixes the problem by moving the computation outside
of the conditional. The patch is available on the stats-fix branch here:
git://github.com/dmpots/ghc.git stats-fix
Unfortunately I am unable to validate, because the validate script fails even
without my patch. It gives this error:
compiler/nativeGen/AsmCodeGen.lhs:71:1:
Warning: The import of `Config' is redundant
except perhaps to import instances from `Config'
To import instances alone, use: import Config()
<no location info>:
Failing due to -Werror.
For those interested, here is a small example that shows the problem.
$ cat Main.hs
module Main where
import System.Environment
loop 0 n = putStrLn $ "Done: " ++ n
loop m n = loop (m-1) n
main = do
[n] <- getArgs
loop (read n) n
In the head version, notice that the mutator time is zero (0.0 MUT).
$ time ./head 20000000 +RTS -tstderr -RTS
Done: 20000000
<<ghc: 3200064664 bytes, 6111 GCs, 2088/2088 avg/max bytes residency (1
samples), 1M in use, 0.00 INIT (0.00 elapsed), 0.00 MUT (0.00 elapsed), 0.06 GC
(0.08 elapsed) :ghc>>
real 0m4.202s
user 0m4.162s
sys 0m0.037s
In the fixed version, it looks correct (4.90 MUT).
$ time ./fixed 20000000 +RTS -tstderr -RTS
Done: 20000000
<<ghc: 3200064664 bytes, 6111 GCs, 2088/2088 avg/max bytes residency (1
samples), 1M in use, 0.00 INIT (0.00 elapsed), 4.90 MUT (4.94 elapsed), 0.08 GC
(0.09 elapsed) :ghc>>
real 0m5.041s
user 0m4.981s
sys 0m0.050s
_______________________________________________
Cvs-ghc mailing list
Cvs-ghc@haskell.org
http://www.haskell.org/mailman/listinfo/cvs-ghc
