parallel program in haskell in 5 steps

In step 4 of Haskell in 5 Steps [ http://haskell.org/haskellwiki/Haskell_in_5_steps], a parallel program is given. I changed it very slightly so it would run a long time (see below). It compiles and runs but my CPU meter is barely above 50%. I have a dual core processor. What in the world would keep this program from completely saturating the CPU? import Control.Parallel main = a `par` b `pseq` print (a + b) where a = ack 4 10 b = ack 4 10 fac 0 = 1 fac n = n * fac (n-1) ack 0 n = n+1 ack m 0 = ack (m-1) 1 ack m n = ack (m-1) (ack m (n-1)) fib 0 = 0 fib 1 = 1 fib n = fib (n-1) + fib (n-2)

How are you running the program? You have to explicitly tell the
compiler/interpreter to use multiple system threads.
Michael
On Tue, May 5, 2009 at 10:19 PM, Jack Kennedy
In step 4 of Haskell in 5 Steps [ http://haskell.org/haskellwiki/Haskell_in_5_steps], a parallel program is given. I changed it very slightly so it would run a long time (see below).
It compiles and runs but my CPU meter is barely above 50%. I have a dual core processor. What in the world would keep this program from completely saturating the CPU?
import Control.Parallel
main = a `par` b `pseq` print (a + b) where a = ack 4 10 b = ack 4 10
fac 0 = 1 fac n = n * fac (n-1)
ack 0 n = n+1 ack m 0 = ack (m-1) 1 ack m n = ack (m-1) (ack m (n-1))
fib 0 = 0 fib 1 = 1 fib n = fib (n-1) + fib (n-2)
_______________________________________________ Beginners mailing list Beginners@haskell.org http://www.haskell.org/mailman/listinfo/beginners

I am compiling (Windows by the way) using the line from the tutorial:
ghc -O2 --make par.hs -threaded
and running with the line
par +RTS -N2
CPU usage for the process flits around a little, but stays in the 45% - 55%
range.
On Tue, May 5, 2009 at 12:51 PM, Michael Snoyman
How are you running the program? You have to explicitly tell the compiler/interpreter to use multiple system threads.
Michael
On Tue, May 5, 2009 at 10:19 PM, Jack Kennedy
wrote: In step 4 of Haskell in 5 Steps [ http://haskell.org/haskellwiki/Haskell_in_5_steps], a parallel program is given. I changed it very slightly so it would run a long time (see below).
It compiles and runs but my CPU meter is barely above 50%. I have a dual core processor. What in the world would keep this program from completely saturating the CPU?
import Control.Parallel
main = a `par` b `pseq` print (a + b) where a = ack 4 10 b = ack 4 10
fac 0 = 1 fac n = n * fac (n-1)
ack 0 n = n+1 ack m 0 = ack (m-1) 1 ack m n = ack (m-1) (ack m (n-1))
fib 0 = 0 fib 1 = 1 fib n = fib (n-1) + fib (n-2)
_______________________________________________ Beginners mailing list Beginners@haskell.org http://www.haskell.org/mailman/listinfo/beginners

Does this happen for everyone, or just me?
On Tue, May 5, 2009 at 2:05 PM, Jack Kennedy
I am compiling (Windows by the way) using the line from the tutorial: ghc -O2 --make par.hs -threaded
and running with the line
par +RTS -N2
CPU usage for the process flits around a little, but stays in the 45% - 55% range.
On Tue, May 5, 2009 at 12:51 PM, Michael Snoyman
wrote: How are you running the program? You have to explicitly tell the compiler/interpreter to use multiple system threads.
Michael
On Tue, May 5, 2009 at 10:19 PM, Jack Kennedy
wrote: In step 4 of Haskell in 5 Steps [ http://haskell.org/haskellwiki/Haskell_in_5_steps], a parallel program is given. I changed it very slightly so it would run a long time (see below).
It compiles and runs but my CPU meter is barely above 50%. I have a dual core processor. What in the world would keep this program from completely saturating the CPU?
import Control.Parallel
main = a `par` b `pseq` print (a + b) where a = ack 4 10 b = ack 4 10
fac 0 = 1 fac n = n * fac (n-1)
ack 0 n = n+1 ack m 0 = ack (m-1) 1 ack m n = ack (m-1) (ack m (n-1))
fib 0 = 0 fib 1 = 1 fib n = fib (n-1) + fib (n-2)
_______________________________________________ Beginners mailing list Beginners@haskell.org http://www.haskell.org/mailman/listinfo/beginners

Hello Jack,
according to the source code your program should not run for a long
time. If you're using GHC, use the '-threaded' flag when compiling and
also pass the parameters '+RTS -N2' to your program to let it run with
two parallel threads.
However, my experience is that the pure parallel constructs are
currently not really suitable for real world programs. With my dual
core processor I get the expected ~190% CPU time using explicit
parallelism using concurrency (i.e. forkIO and MVars), whereas with pure
constructs like the ones from Control.Parallel and
Control.Parallel.Strategies I get only about 140%.
Greets,
Ertugrul.
Jack Kennedy
In step 4 of Haskell in 5 Steps [ http://haskell.org/haskellwiki/Haskell_in_5_steps], a parallel program is given. I changed it very slightly so it would run a long time (see below).
It compiles and runs but my CPU meter is barely above 50%. I have a dual core processor. What in the world would keep this program from completely saturating the CPU?
import Control.Parallel
main = a `par` b `pseq` print (a + b) where a = ack 4 10 b = ack 4 10
fac 0 = 1 fac n = n * fac (n-1)
ack 0 n = n+1 ack m 0 = ack (m-1) 1 ack m n = ack (m-1) (ack m (n-1))
fib 0 = 0 fib 1 = 1 fib n = fib (n-1) + fib (n-2)
-- nightmare = unsafePerformIO (getWrongWife >>= sex) http://blog.ertes.de/
participants (4)
-
Ertugrul Soeylemez
-
Jack Kennedy
-
Michael Snoyman
-
Sean Bartell