Dear Arnaldo, IMHO you should post the code (or a link where one can get your code) that demonstrates this problem. Otherwise it's very hard to tell. Best, Dmitrii
2008/7/16 Arnaldo Mandel <[EMAIL PROTECTED]>: > Hi everybody, > > After struggling with a very slow running function, I was met with a > mystery. Here is profiling data: > > gap> ProfileOperationsAndMethods(true); > gap> ProfileGlobalFunctions( true ); > gap> DisplayProfile(); > count self/ms chld/ms function > 16902 0 0 ADD_LIST > 10 TOTAL > gap> x4:=Complete(BFST[4]);; > gap> time; > 2250 > gap> DisplayProfile(); > count self/ms chld/ms function > 13694 0 0 Size: for a list that is a collection > 23967 0 0 ADD_LIST > 4229 2220 0 Order: for a group > 2260 TOTAL > gap> x5:=Complete(BFST[5]);; > gap> time; > 1323640 > gap> DisplayProfile(); > count self/ms chld/ms function > 13737 10 0 WriteAll: output text file > 18332 30 10 WriteAll > 12942 20 20 Concatenation > 43325 40 10 Enumerator: for a collection that is a list > 6471 30 50 String: for an object, and a positive integer > 12942 0 90 String > 162543 90 10 APPEND_LIST > 316529 80 30 Size: for a list that is a collection > 167708 130 0 Remove: two arguments, fast > 209192 240 10 SHALLOW_COPY_OBJ: for a presentation in default * > 2 50 200 Perform > 43325 260 0 SSortedList: for a plist > 167708 270 130 Remove > 226094 490 -10 ADD_LIST > 43325 240 300 AsSSortedList: for a plist > 43325 180 550 Enumerator: for a collection with known `AsSSort* > 94236 1322830 80 Order: for a group > 1325010 TOTAL > > As you can see, Order takes up all the time. However, Complete does > not involve any group. Its argument is a list of small matrices, and > the return value is of the same type. In between, several new > matrices are created; each new matrix is a StructuralCopy of an > existing one, followed by some simple modifications. Functions that are > explicitly called are: Size(list), StructuralCopy, Append, Add, Remove, > AsSet(small list of integers), Print, RemInt, String(integer,integer). > > So, who is calling Order? > > am > > -- > Arnaldo Mandel > Departamento de Ciência da Computação - Computer Science Department > Universidade de São Paulo, Bra[sz]il > [EMAIL PROTECTED] > Talvez você seja um Bright http://the-brights.net Maybe you are a Bright. > > _______________________________________________ > Forum mailing list > [email protected] > http://mail.gap-system.org/mailman/listinfo/forum > _______________________________________________ Forum mailing list [email protected] http://mail.gap-system.org/mailman/listinfo/forum
