Yay!!!<div><br></div><div>I made a small change in Types.chs and got my original cairo-binding-based program to be just as blazing fast. The only problem I have with this is that I used multiparameter type classes.</div><div>
<br></div><div>Dear gtk2hs team! Is it possible to incorporate my changes? I'm pretty sure people will be happy by an order-of-magnitude speedup. Probably the stuff could be wrapped in #define's for those who aren't using GHC and can't use multiparameter type classes?</div>
<div><br></div><div>I am pretty sure I could have done the same with rewrite rules, but I tried for a while and to no avail.</div><div><br></div><div>FAILED SOLUTION: rewrite rules</div><div><div>cFloatConv :: (RealFloat a, RealFloat b) => a -> b</div>
<div>cFloatConv = realToFrac</div><div>{-# NOINLINE cFloatConv #-}</div><div>{-# RULES "cFloatConv/float2Double" cFloatConv = float2Double #-}</div><div>{-# RULES "cFloatConv/double2Float" cFloatConv = double2Float #-}</div>
<div>{-# RULES "cFloatConv/self" cFloatConv = id #-}</div></div><div><br></div><div>For some reason, the rules don't fire. Anyone got an idea why?</div><div><br></div><div>SUCCEEDED SOLUTION: multiparameter type classes</div>
<div><br></div><div>I rewrote cFloatConv like this:</div><div><br></div><div><div>import GHC.Float</div><div>class (RealFloat a, RealFloat b) => CFloatConv a b where</div><div> cFloatConv :: a -> b</div><div> cFloatConv = realToFrac</div>
<div><br></div><div>instance CFloatConv Double Double where cFloatConv = id</div><div>instance CFloatConv Double CDouble</div><div>instance CFloatConv CDouble Double</div><div>instance CFloatConv Float Float where cFloatConv = id</div>
<div>instance CFloatConv Float Double where cFloatConv = float2Double</div><div>instance CFloatConv Double Float where cFloatConv = double2Float</div></div><div><br></div><div>and replaced a couple of constraints in functions below by usage of CFloatConv.</div>
<div><br></div><div><br><div class="gmail_quote">On Wed, Nov 2, 2011 at 2:25 PM, Felipe Almeida Lessa <span dir="ltr"><<a href="mailto:felipe.lessa@gmail.com">felipe.lessa@gmail.com</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex;">
+gtk2hs-devel<br>
<div class="HOEnZb"><div class="h5"><br>
On Wed, Nov 2, 2011 at 8:15 AM, Eugene Kirpichov <<a href="mailto:ekirpichov@gmail.com">ekirpichov@gmail.com</a>> wrote:<br>
> Any idea how to debug why all the GMP calls?<br>
> I'm looking at even the auto-generated source for cairo bindings, but I<br>
> don't see anything at all that could lead to *thousands* of them.<br>
<br>
Found them. Look at the Types module and you'll see<br>
<br>
cFloatConv :: (RealFloat a, RealFloat b) => a -> b<br>
cFloatConv = realToFrac<br>
<br>
This function (or its cousins peekFloatConv, withFloatConv...) are<br>
used *everywhere*.<br>
<br>
Looking at this module with ghc-core we see that GHC compiled a<br>
generic version of cFloatConv:<br>
<br>
Graphics.Rendering.Cairo.Types.$wcFloatConv<br>
:: forall a_a3TN b_a3TO.<br>
(RealFloat a_a3TN, RealFrac b_a3TO) =><br>
a_a3TN -> b_a3TO<br>
[GblId,<br>
Arity=3,<br>
<br>
Unf=Unf{Src=<vanilla>, TopLvl=True, Arity=3, Value=True,<br>
ConLike=True, Cheap=True, Expandable=True,<br>
Guidance=IF_ARGS [3 3 0] 12 0}]<br>
Graphics.Rendering.Cairo.Types.$wcFloatConv =<br>
\ (@ a_a3TN)<br>
(@ b_a3TO)<br>
(w_s5zg :: RealFloat a_a3TN)<br>
(ww_s5zj :: RealFrac b_a3TO)<br>
(w1_s5zA :: a_a3TN) -><br>
fromRational<br>
@ b_a3TO<br>
($p2RealFrac @ b_a3TO ww_s5zj)<br>
(toRational<br>
@ a_a3TN<br>
($p1RealFrac<br>
@ a_a3TN ($p1RealFloat @ a_a3TN w_s5zg))<br>
w1_s5zA)<br>
<br>
Note that this is basically cFloatConv = fromRational . toRational.<br>
<br>
*However*, GHC also compiled a Double -> Double specialization:<br>
<br>
Graphics.Rendering.Cairo.Types.cFloatConv1<br>
:: Double -> Double<br>
[GblId,<br>
Arity=1,<br>
<br>
Unf=Unf{Src=InlineStable, TopLvl=True, Arity=1, Value=True,<br>
ConLike=True, Cheap=True, Expandable=True,<br>
Guidance=ALWAYS_IF(unsat_ok=True,boring_ok=False)<br>
Tmpl= \ (eta_B1 [Occ=Once!] :: Double) -><br>
case eta_B1 of _ { D# ww_a5v3 [Occ=Once] -><br>
case $w$ctoRational ww_a5v3<br>
of _ { (# ww2_a5v8 [Occ=Once], ww3_a5v9 [Occ=Once] #) -><br>
$wfromRat ww2_a5v8 ww3_a5v9<br>
}<br>
}}]<br>
Graphics.Rendering.Cairo.Types.cFloatConv1 =<br>
\ (eta_B1 :: Double) -><br>
case eta_B1 of _ { D# ww_a5v3 -><br>
case $w$ctoRational ww_a5v3<br>
of _ { (# ww2_a5v8, ww3_a5v9 #) -><br>
$wfromRat ww2_a5v8 ww3_a5v9<br>
}<br>
}<br>
<br>
...which is also equivalent to fromRational . toRational however with<br>
the type class inlined! Oh, god...<br>
<br>
Cheers,<br>
<br>
--<br>
Felipe.<br>
<br>
</div></div><div class="HOEnZb"><div class="h5">_______________________________________________<br>
Haskell-Cafe mailing list<br>
<a href="mailto:Haskell-Cafe@haskell.org">Haskell-Cafe@haskell.org</a><br>
<a href="http://www.haskell.org/mailman/listinfo/haskell-cafe" target="_blank">http://www.haskell.org/mailman/listinfo/haskell-cafe</a><br>
</div></div></blockquote></div><br><br clear="all"><div><br></div>-- <br>Eugene Kirpichov<br>Principal Engineer, Mirantis Inc. <a href="http://www.mirantis.com/" target="_blank">http://www.mirantis.com/</a><br>Editor, <a href="http://fprog.ru/" target="_blank">http://fprog.ru/</a><br>
</div>