Hi,
I did a test: load a set of objects, delete and load again
results: ace35b and the "ace4 at the end of the workshop" stabilize
after 3 interations at three times the initial database size.
ace950620, however, keeps constant size (see lines *.dbsize below).
ace950620 requires the same amount of main memory as the ace35b.
the test was run with 400 KB .ace file of Sequence data for the IGD model.
Nice to see this bug fixed.
best wishes
Detlef
------------------------------------------------------------------
the script:
load: dbinit wspec.igd
t1: tace
do parse igd.create.ace.gz
exit
d1: dbsize
t2: tace
do parse igd.create.ace.gz
exit
d2: dbsize
t3: tace
do parse igd.create.ace.gz
exit
d3: dbsize
t4: tace
do parse igd.create.ace.gz
exit
d4: dbsize
credel: dbinit wspec.igd
t1: tace
do @igd.credel.cmd
mem
exit // t1
d1: dbsize
t2: tace
do @igd.credel.cmd
mem
exit
d2: dbsize
t3: tace
do @igd.credel.cmd
mem
exit
d3: dbsize
t4: tace
do @igd.credel.cmd
mem
exit
d4: dbsize
where igd.credel.cmd is:
parse igd.create.ace.gz
find Sequence
parse igd.deltag.ace.gz
find Sequence
parse igd.create.ace.gz
find Sequence
parse igd.delobj.ace.gz
where find Sequence
classes
----------------- aquila 1.4 comparison --------------------------------
reference version (old): ace35b
Sat Jun 24 16:49:38 MDT 1995
IRIX mbp-sgi5 5.2 02282015 IP19 mips
database on local file system
under test (new): atmp
Sat Jun 24 17:02:41 MDT 1995
IRIX mbp-sgi5 5.2 02282015 IP19 mips
ace950620
database files on local file system
compiled with SGI_4_NEW
label/command type old new comment/ratio
-----------------------------------------------------------------------
NOTE: some lines delete by hand
load.dbinit #msgs 2 6 problematic
load.t1.exit real 7 8 sec, ratio 1.14
load.t1.exit cpu 3.30 3.80 sec, ratio 1.15
load.d1.dbsize KB 4995 same ratio 1.00
load.t2.exit real 4 9 sec, ratio 2.25
load.t2.exit cpu 2.80 same sec, ratio 1.00
load.d2.dbsize KB 4995 same ratio 1.00
load.t3.exit real 8 7 sec, ratio 0.88
load.t3.exit cpu 2.70 2.80 sec, ratio 1.04
load.d3.dbsize KB 4996 4995 ratio 1.00
load.t4.exit real 4 7 sec, ratio 1.75
load.t4.exit cpu 2.60 2.70 sec, ratio 1.04
load.d4.dbsize KB 4996 4995 ratio 1.00
credel.dbinit #msgs 2 6 problematic
credel.t1._._.mem sz 7136 7200 ratio 1.01
credel.t1._._.mem rss 6388 6436 ratio 1.01
credel.t1.exit real 22 26 sec, ratio 1.18
credel.t1.exit cpu 8.10 9.40 sec, ratio 1.16
credel.d1.dbsize KB 4996 4995 ratio 1.00
credel.t2._._.mem sz 7104 7152 ratio 1.01
credel.t2._._.mem rss 6356 6388 ratio 1.01
credel.t2.exit real 30 26 sec, ratio 0.87
credel.t2.exit cpu 8.40 9.40 sec, ratio 1.12
credel.d2.dbsize KB 9988 4995 ratio 0.50
credel.t3._._.mem sz 7104 7152 ratio 1.01
credel.t3._._.mem rss 6368 6388 ratio 1.00
credel.t3.exit real 30 26 sec, ratio 0.87
credel.t3.exit cpu 8.20 9.00 sec, ratio 1.10
credel.d3.dbsize KB 14980 4996 ratio 0.33
credel.t4._._.mem sz 7072 7152 ratio 1.01
credel.t4._._.mem rss 6324 6388 ratio 1.01
credel.t4.exit real 21 24 sec, ratio 1.14
credel.t4.exit cpu 8.20 9.70 sec, ratio 1.18
credel.d4.dbsize KB 14981 4996 ratio 0.33
-----------------------------------------------------------------------
