oldk1331 wrote:
>
> I said before that the following test takes too much time to run:
> testIntegrate("sqrt(tan(x)^2 + 2*tan(x) + 2)", "x", "alg")
>
> I did some profiling, and found that 40% time is spent in
> symmetricRemainder:
>
> Callers
> Total. Function
> Count % Count % Callees
> ------------------------------------------------------------------------
> 3 0.0 |MODRING;reduce;RMod$;6| [76]
> 13 0.1 |INS-;symmetricRemainder;3S;27| [1]
> 3431 39.3 |IDPO;map;M2$;6| [5]
> 1104 12.7 3451 39.6 |INS-;symmetricRemainder;3S;27| [1]
> 13 0.1 |INS-;symmetricRemainder;3S;27| [1]
> 160 1.8 |INT;=;2$B;30| [4]
> 149 1.7 |ABELMON-;*;Pi2S;2| [14]
> 67 0.8 SB-VM::ALLOC-TRAMP [15]
> 228 2.6 |ORDSET-;>;2SB;4| [7]
> 173 2.0 |ORDSET-;<=;2SB;6| [20]
> 73 0.8 |INT;+;3$;34| [9]
> 160 1.8 |INT;*;3$;37| [6]
> 180 2.1 |ABELGRP-;*;Nni2S;3| [11]
> 366 4.2 |INT;rem;3$;44| [16]
> 369 4.2 IDENTITY [2]
> 407 4.7 |INT;<;2$B;31| [3]
>
>
> We can see that "=$INT, +$INT, *$INT, <$INT" are not inlined, and
> "*$ABELMON, *$ABELGRP, >, <=" are not defined in INT, thus impossible
> to get inlined.
>
> So I want to add inline optimization for these functions.
> (I also add inline support for "positive?, even?, ^")
>
> =========== test case
> )time on
> )lisp (require :sb-sprof)
>
> f := sqrt(tan(x)^2 + 2*tan(x) + 2)
> g := integrate(f, x)
>
> f2 := D(g, x)
>
> )lisp (sb-sprof:start-profiling)
> normalize(f-f2)
> )lisp (sb-sprof:stop-profiling)
>
> )lisp (sb-sprof:report)
> ===========
>
> Benchmark result:
> D(g, x) normalize(f-f2)
> before patch 40.3 s 87.4 s
> after patch 28.4 s 57.3 s
>
> So it's 30% faster, saves 40 seconds in testing.
OK, good. But I see some funny formatting below, which may be
due tabs. Please make sure there are no tabs when you commit.
>
> diff --git a/src/algebra/integer.spad b/src/algebra/integer.spad
> index 0e619c2f..35f673b2 100644
> --- a/src/algebra/integer.spad
> +++ b/src/algebra/integer.spad
> @@ -59,7 +59,6 @@
> ZP ==> SparseUnivariatePolynomial %
> ZZP ==> SparseUnivariatePolynomial Integer
> x, y : %
> - n : NonNegativeInteger
^^^^^^
Would be indentation error if real.
--
Waldek Hebisch
--
You received this message because you are subscribed to the Google Groups
"FriCAS - computer algebra system" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
To post to this group, send email to [email protected].
Visit this group at https://groups.google.com/group/fricas-devel.
For more options, visit https://groups.google.com/d/optout.
