#haskell - Wed 9 May 2007 between 00:19 and 01:24

NY Lost Funds



chessguy?seen procyon112
lambdabotI saw procyon112 leaving #gentoo-haskell and #haskell 1d 8m 11s ago, and .
wyIs there a cube root function?
SamB_XP> 27 ** (1/3)
ddarius(**(1/3))
lambdabot3.0
dmwit> 27 ^^ (1/3)
lambdabotAdd a type signature
dmwit:t (^^)
lambdabotforall a b. (Integral b, Fractional a) => a -> b -> a
mainland_i'm having a hard time figuring out how to use the profiler
i have a 4-lien function that takes up about 50% of my run-time
ddariusmainland_: You've RTFM?
SamB_XPmainland_: what have you figured out so far?
did you read about all the profiling-related compiler options?
mainland_well, i turned on -auto-all
yes, i did
i added cost centers to all 4 expressions in my 4-line function
just for kicks
SamB_XPoh?
mainland_and yet 50% still sits in the "individual" time column
CaleHow many entries?
mainland_1
Calehmm, okay
mainland_space leak whose thunks all get collapsed?
Calemm, I don't know. I think it attributes time to the thing which is creating the thunks in that case.
wyHow about cube root into complex numbers?
SamB_XPhmm?
CaleWhat is the code doing?
SamB_XPyou want the *other* cube roots?
mainland_all the other profiling numbers make sense and are broken down nicely, but this one function that doesn't do anything "owns" 50% of my runtime
ddarius:t exp
lambdabotforall a. (Floating a) => a -> a
wySamB_XP: yes. Just for test.
mainland_it's a theorem prover
heavy use of LogicT
Calewy: x**(1/3) ?
dmwit> (3 :+ 4) ^ (1/3) -- ?
lambdabotAdd a type signature
dmwit:t (:+)

Page: 2 9 16 23 30 37 44 51 58 65 72 

IrcArchive

NY Lost Funds