Hi,
I'm spotting a memory leak when I return a tuple of arrays, and attempt to
unpack the tuple in the calling statement. An test code is below... If I
replace the var (a1,a2)= .... statement in the code below with var a1 =,
the memory leak goes away.
I'm not sure if I'm not allowed to unpack a tuple as I'm doing (but the
compiler certainly lets me do it). This is, of course, easy to work
around....
Thanks in advance!
-- Nikhil
Specifically, consider the following Chapel code :
config const N=128;
const Ntot=N**3;
proc mkarray() {
var pos : [0.. #Ntot] real;
var vel : [0.. #Ntot] real;
return (pos, vel);
}
var (a1,a2) = mkarray();
writeln("Done");
compiled as :
chpl -o tes_memleakt.x test_memleak.chpl
run as :
./test_memleak.x --memTrack=true --memStats=true --memLeaks
I get...
Done
=================
Memory Statistics
==============================================================
Current Allocated Memory 33554944
Maximum Simultaneous Allocated Memory 67110280
Total Allocated Memory 100665032
Total Freed Memory 67110088
==============================================================
=================================================================================================================
Allocated Memory (Bytes) Number Size Total Description
Address
=================================================================================================================
test_memleak.chpl:7 1 136 136
[domain(1,int(64),false)] real(64)0x00007f5c6962d280
test_memleak.chpl:7 1 136 136
[domain(1,int(64),false)] real(64)0x00007f5c6962d320
test_memleak.chpl:7 2097152 8 16777216 array elements
0x00007f5c5ee00000
test_memleak.chpl:7 2097152 8 16777216 array elements
0x00007f5c5de00000
test_memleak.chpl:5 1 96 96
domain(1,int(64),false) 0x00007f5c6962a180
test_memleak.chpl:6 1 96 96
domain(1,int(64),false) 0x00007f5c6962a1e0
test_memleak.chpl:7 1 24 24
listNode(BaseArr) 0x00007f5c69619d80
test_memleak.chpl:7 1 24 24
listNode(BaseArr) 0x00007f5c69619d60
=================================================================================================================
---------------------------------
Nikhil Padmanabhan
[email protected]
------------------------------------------------------------------------------
Transform Data into Opportunity.
Accelerate data analysis in your applications with
Intel Data Analytics Acceleration Library.
Click to learn more.
http://pubads.g.doubleclick.net/gampad/clk?id=278785471&iu=/4140
_______________________________________________
Chapel-users mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/chapel-users
