Re: [Monetdb-developers] [Monetdb-checkins] MonetDB5/src/modules/kernel batcalc.mx, , 1.145, 1.146 batifthen.mx, , 1.10, 1.11 group.mx, , 1.99, 1.100 kprelude.mx, , 1.7, 1.8
On Tue, Dec 11, 2007 at 05:32:04PM +0000, Niels Nes wrote:
Update of /cvsroot/monetdb/MonetDB5/src/modules/kernel In directory sc8-pr-cvs16.sourceforge.net:/tmp/cvs-serv28347/src/modules/kernel
Modified Files: batcalc.mx batifthen.mx group.mx kprelude.mx Log Message: batifthen.mx: use (fast) direct assignments instead of bunfastins for ifthenelse (not for ifthen as it can be small as the input bat) group.mx: return read only bats.
Intel's icc says: ======== icc -no-gcc -DHAVE_CONFIG_H -I. -I/scratch/monet.Int.64.64.d.23892/MonetDB5/src/modules/kernel -I../../.. -I/scratch/monet.Int.64.64.d.23892/MonetDB5/src/modules/kernel -I../atoms -I/scratch/monet.Int.64.64.d.23892/MonetDB5/src/modules/kernel/../ atoms -I../../mal -I/scratch/monet.Int.64.64.d.23892/MonetDB5/src/modules/kernel/../../mal -I/scratch/monet.Int.64.64.d.23892/MonetDB/.Int.64.64.d-Fedora6/include/MonetDB -I/scratch/monet.Int.64.64.d.23892/MonetDB/.Int.64.64.d-Fedora6/include/Mone tDB/common -I/scratch/monet.Int.64.64.d.23892/MonetDB/.Int.64.64.d-Fedora6/include/MonetDB/gdk -DLIBBATIFTHEN -we140 -wd1418 -Wall -w2 -c99 -mp1 -O3 -restrict -unroll -tpp7 -axWP -we266 -Werror -wd1418,1419,279,981,810,193,111,1357,1572,1599 -D_RE ENTRANT -c batifthen.c -KPIC -DPIC -o .libs/lib_batifthen_la-batifthen.o /scratch/monet.Int.64.64.d.23892/MonetDB5/src/modules/kernel/batifthen.mx(430): error #593: variable "bi" was set but never used BATiter bi, tbi; ^
/scratch/monet.Int.64.64.d.23892/MonetDB5/src/modules/kernel/batifthen.mx(430): error #593: variable "bi" was set but never used BATiter bi, tbi; ^
compilation aborted for batifthen.c (code 2) make[7]: *** [lib_batifthen_la-batifthen.lo] Error 1 ========
basic questions: is it correct that the input's heads are no longer used in CMDifThenElseCst2_@1(), while they are (still) used in CMDifThenElseCst_@1(), CMDifThenElse_@1() & CMDifThenElseCst1_@1() ?
Stefan
[...]
Index: batifthen.mx
RCS file: /cvsroot/monetdb/MonetDB5/src/modules/kernel/batifthen.mx,v retrieving revision 1.10 retrieving revision 1.11 diff -u -d -r1.10 -r1.11 --- batifthen.mx 14 Nov 2007 12:52:15 -0000 1.10 +++ batifthen.mx 11 Dec 2007 17:32:02 -0000 1.11
[...]
@@ -289,38 +436,31 @@ @:getBATdescriptor(bid,b,"batcalc.ifThenElse")@ @:getBATdescriptor(tid,tb,"batcalc.ifThenElse",BBPreleaseref(b->batCacheid);)@ @:chkSize(b,tb,ifThenElse)@
- @:resBAT(@1,"batcalc.ifThenElse")@
- @:resultBAT(@1,"batcalc.ifThenElse")@
- bn->tsorted = FALSE;
BATkey(BATmirror(bn), FALSE);
if (ATOMvarsized(TYPE_@1))
val = *(@1**)val;bi = bat_iterator(b); tbi = bat_iterator(tb);
t = (bit*) Tloc(b,BUNfirst(b));
- if (ATOMvarsized(TYPE_@1))
val = *(@1**)val;
BUNfastins(bn, BUNhead(bi,p), (ptr) &nilval);
BUNfastins(bn, NULL, (ptr) &nilval);
BUNfastins(bn, BUNhead(bi,p), BUNtail(tbi, p));
BUNfastins(bn, NULL, BUNtail(tbi, p));
BUNfastins(bn, BUNhead(bi,p), val);
BUNfastins(bn, NULL, val);} +@:ifthenImpl(@1)@ @c -#include "mal_config.h" -#include "batifthen.h"
-@:ifthencstImpl(int)@ -@:ifthencstImpl(sht)@ -@:ifthencstImpl(lng)@ -@:ifthencstImpl(oid)@ -@:ifthencstImpl(flt)@ -@:ifthencstImpl(dbl)@ -@:ifthencstImpl(str)@ -@:ifthencstImpl(chr)@ -@:ifthencstImpl(bit)@ +@:ifthenelseDefault(oid)@ +@:ifthenelseDefault(str)@ @}
[...]
On Wed, Dec 12, 2007 at 12:40:09PM +0100, Stefan Manegold wrote:
On Tue, Dec 11, 2007 at 05:32:04PM +0000, Niels Nes wrote:
Update of /cvsroot/monetdb/MonetDB5/src/modules/kernel In directory sc8-pr-cvs16.sourceforge.net:/tmp/cvs-serv28347/src/modules/kernel
Modified Files: batcalc.mx batifthen.mx group.mx kprelude.mx Log Message: batifthen.mx: use (fast) direct assignments instead of bunfastins for ifthenelse (not for ifthen as it can be small as the input bat) group.mx: return read only bats.
Intel's icc says:
icc -no-gcc -DHAVE_CONFIG_H -I. -I/scratch/monet.Int.64.64.d.23892/MonetDB5/src/modules/kernel -I../../.. -I/scratch/monet.Int.64.64.d.23892/MonetDB5/src/modules/kernel -I../atoms -I/scratch/monet.Int.64.64.d.23892/MonetDB5/src/modules/kernel/../ atoms -I../../mal -I/scratch/monet.Int.64.64.d.23892/MonetDB5/src/modules/kernel/../../mal -I/scratch/monet.Int.64.64.d.23892/MonetDB/.Int.64.64.d-Fedora6/include/MonetDB -I/scratch/monet.Int.64.64.d.23892/MonetDB/.Int.64.64.d-Fedora6/include/Mone tDB/common -I/scratch/monet.Int.64.64.d.23892/MonetDB/.Int.64.64.d-Fedora6/include/MonetDB/gdk -DLIBBATIFTHEN -we140 -wd1418 -Wall -w2 -c99 -mp1 -O3 -restrict -unroll -tpp7 -axWP -we266 -Werror -wd1418,1419,279,981,810,193,111,1357,1572,1599 -D_RE ENTRANT -c batifthen.c -KPIC -DPIC -o .libs/lib_batifthen_la-batifthen.o /scratch/monet.Int.64.64.d.23892/MonetDB5/src/modules/kernel/batifthen.mx(430): error #593: variable "bi" was set but never used BATiter bi, tbi; ^
/scratch/monet.Int.64.64.d.23892/MonetDB5/src/modules/kernel/batifthen.mx(430): error #593: variable "bi" was set but never used BATiter bi, tbi; ^
compilation aborted for batifthen.c (code 2) make[7]: *** [lib_batifthen_la-batifthen.lo] Error 1 ========
basic questions: is it correct that the input's heads are no longer used in CMDifThenElseCst2_@1(), while they are (still) used in CMDifThenElseCst_@1(), CMDifThenElse_@1() & CMDifThenElseCst1_@1()
No in this fall back case all should still use the head. I'll check in a fix shortly.
Niels
?
Stefan
[...]
Index: batifthen.mx
RCS file: /cvsroot/monetdb/MonetDB5/src/modules/kernel/batifthen.mx,v retrieving revision 1.10 retrieving revision 1.11 diff -u -d -r1.10 -r1.11 --- batifthen.mx 14 Nov 2007 12:52:15 -0000 1.10 +++ batifthen.mx 11 Dec 2007 17:32:02 -0000 1.11
[...]
@@ -289,38 +436,31 @@ @:getBATdescriptor(bid,b,"batcalc.ifThenElse")@ @:getBATdescriptor(tid,tb,"batcalc.ifThenElse",BBPreleaseref(b->batCacheid);)@ @:chkSize(b,tb,ifThenElse)@
- @:resBAT(@1,"batcalc.ifThenElse")@
- @:resultBAT(@1,"batcalc.ifThenElse")@
- bn->tsorted = FALSE;
BATkey(BATmirror(bn), FALSE);
if (ATOMvarsized(TYPE_@1))
val = *(@1**)val;bi = bat_iterator(b); tbi = bat_iterator(tb);
t = (bit*) Tloc(b,BUNfirst(b));
- if (ATOMvarsized(TYPE_@1))
val = *(@1**)val;
BUNfastins(bn, BUNhead(bi,p), (ptr) &nilval);
BUNfastins(bn, NULL, (ptr) &nilval);
BUNfastins(bn, BUNhead(bi,p), BUNtail(tbi, p));
BUNfastins(bn, NULL, BUNtail(tbi, p));
BUNfastins(bn, BUNhead(bi,p), val);
BUNfastins(bn, NULL, val);} +@:ifthenImpl(@1)@ @c -#include "mal_config.h" -#include "batifthen.h"
-@:ifthencstImpl(int)@ -@:ifthencstImpl(sht)@ -@:ifthencstImpl(lng)@ -@:ifthencstImpl(oid)@ -@:ifthencstImpl(flt)@ -@:ifthencstImpl(dbl)@ -@:ifthencstImpl(str)@ -@:ifthencstImpl(chr)@ -@:ifthencstImpl(bit)@ +@:ifthenelseDefault(oid)@ +@:ifthenelseDefault(str)@ @}
[...]
-- | Dr. Stefan Manegold | mailto:Stefan.Manegold@cwi.nl | | CWI, P.O.Box 94079 | http://www.cwi.nl/~manegold/ | | 1090 GB Amsterdam | Tel.: +31 (20) 592-4212 | | The Netherlands | Fax : +31 (20) 592-4312 |
SF.Net email is sponsored by: Check out the new SourceForge.net Marketplace. It's the best place to buy or sell services for just about anything Open Source. http://sourceforge.net/services/buy/index.php _______________________________________________ Monetdb-developers mailing list Monetdb-developers@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/monetdb-developers
participants (2)
-
Niels Nes -
Stefan Manegold