Yeah, you know after sending the email (never a better time) I noticed
that the C version wasn't spitting out the right answer. I'm not
really sure why, I just replaced bigint with int64_t from stdint.h.
-Ross
On Jan 28, 2009, at 8:32 PM, Benedikt Huber wrote:
Ross Mellgren schrieb:
Duncan, I think you must have some magics -- on my machine the
original code also takes forever.
Running with +RTS -S indicates it's allocating several gig of
memory or more.
Applying some bang patterns gives me ~8s for 10^8 and somewhat more
than a minute for 10^9:
Hi,
same here, with bang patterns ~100s / 1Mb but
The C program is quite fast:
r...@hugo:~$ time ./circ-orig
1302219321
looks wrong to me
real 0m1.073s
cafe(0) $ time ./gauss
3141592649589764829
real 0m17.894s
So my 32-bit machine is really slow ...
benedikt
On Wed, 2009-01-28 at 16:42 -0800, drblanco wrote:
I do already have the number I wanted, but was wondering how this
could be
made faster, or even why it's so slow. This is all on GHC 6.8.3
under OS X
Intel, using ghc -O2.
I'm not exactly sure what's different, but for me it works pretty
well.
I put back in the Int64 type signature.
For comparison, the C code below runs in <1 second.
You've got a faster machine than me :-)
I compiled both the Haskell and C versions to standalone executables
with ghc/gcc -O2 and ran them with time.
C version:
$ time ./circ
3141592649589764829
real 0m2.430s
user 0m2.428s
sys 0m0.000s
Haskell version:
time ./circ2
3141592653589764829
real 0m2.753s
user 0m2.756s
sys 0m0.000s
Not too bad I'd say! :-)
I was using ghc-6.10 for this test. It would appear that ghc-6.8
is a
bit slower, I get:
3141592653589764829
real 0m5.767s
user 0m5.768s
sys 0m0.000s
Now the other difference is that I'm using a 64bit machine so
perhaps
ghc just produces terrible code for Int64 on 32bit machines.
Duncan
_______________________________________________
Haskell-Cafe mailing list
[email protected]
http://www.haskell.org/mailman/listinfo/haskell-cafe
_______________________________________________
Haskell-Cafe mailing list
[email protected]
http://www.haskell.org/mailman/listinfo/haskell-cafe
