Re: limit for # of items created with new ?

2002-10-11 Thread Ljubomir Josifovski
On Tue, Sep 24, 2002 at 10:09:59AM -0700, Hans Horn wrote: List, I am experiencing a problem when I need to allocate a large number of small objects with new. (cygwin 1.3.12.4, gcc 3.2). Up to about 160 000 objects I'm doing fine, but above the process receives a sigterm and croaks. On

Re: limit for # of items created with new ?

2002-09-24 Thread Eric M. Monsler
Hans Horn wrote: When this happens, the bash shell complains : *** mount version mismatch detected - 0xA820/0x1B. You have multiple copies of cygwin1.dll on your system. Search for cygwin1.dll using the Windows Start-Find/Search facility and delete all but the most recent version. The

Re: limit for # of items created with new ?

2002-09-24 Thread Hans Horn
Yes, Sir! there's just one cywin1.dll, and lives exactly where it's meant to live thx anyways! H. Eric M. Monsler [EMAIL PROTECTED] wrote in message [EMAIL PROTECTED]">news:[EMAIL PROTECTED]... Hans Horn wrote: When this happens, the bash shell complains : *** mount version mismatch

Re: limit for # of items created with new ?

2002-09-24 Thread Hans Horn
Ooops, I just realized that the number of objects I've allocated is about 1,600,000, NOT 160,000! But still, each object (nodes of a singly linked list) weighs only 8 byte. H. Hans Horn [EMAIL PROTECTED] wrote in message amq618$9kr$[EMAIL PROTECTED]">news:amq618$9kr$[EMAIL PROTECTED]... List,

Re: limit for # of items created with new ?

2002-09-24 Thread Igor Pechtchanski
Don't forget the object headers - depending on which options you use, they can be 8 bytes per object. It would really help to see the source of the crashing program, at least the snippet with the allocation and dereferencing. Igor On Tue, 24 Sep 2002, Hans Horn wrote: Ooops, I just

Re: limit for # of items created with new ?

2002-09-24 Thread Hans Horn
Igor, Igor Pechtchanski [EMAIL PROTECTED] wrote in message [EMAIL PROTECTED]">news:[EMAIL PROTECTED]... Don't forget the object headers - depending on which options you use, they can be 8 bytes per object. It would really help to see the source of the crashing program, at least the snippet

Re: limit for # of items created with new ?

2002-09-24 Thread Randall R Schulz
Hans, That looks like over 80 megabytes just for the dynamically allocated node instances: 85000 * 30 * (8 + 8) * 2 If I understand correctly, all Cygwin app memory comes out of a single pool (hence the Cygwin heap size registry entry), so a proper accounting of available memory must take

Re: limit for # of items created with new ?

2002-09-24 Thread Christopher Faylor
On Tue, Sep 24, 2002 at 07:06:22PM -0700, Randall R Schulz wrote: Hans, That looks like over 80 megabytes just for the dynamically allocated node instances: 85000 * 30 * (8 + 8) * 2 If I understand correctly, all Cygwin app memory comes out of a single pool (hence the Cygwin heap size

Re: limit for # of items created with new ?

2002-09-24 Thread Igor Pechtchanski
Hans, I don't really have any solution, but I can confirm the below. On my system, Win2k SP2, the message is mount table size version mismatch detected - 0x11/0x4068, but otherwise identical to yours. This doesn't seem to be an out-of-memory error. Some further details: Any attempt to start a

RE: limit for # of items created with new ?

2002-09-24 Thread Ross Smith
From: Hans Horn [mailto:[EMAIL PROTECTED]] // an element of a linked list typedef struct node { node (int _v, node* _n) : v(_v), next(_n) {} int v; node* next; }; I'm surprised that compiled; the typedef shouldn't be there. int main (int argc, char** argv) { // allocate

Re: limit for # of items created with new ?

2002-09-24 Thread Igor Pechtchanski
On Tue, 24 Sep 2002, Hans Horn wrote: Igor, Igor Pechtchanski [EMAIL PROTECTED] wrote in message [EMAIL PROTECTED]">news:[EMAIL PROTECTED]... Don't forget the object headers - depending on which options you use, they can be 8 bytes per object. [snip] Are there options that cost

Re: limit for # of items created with new ?

2002-09-24 Thread Emil Briggs
On Tuesday 24 September 2002 10:28 pm, Ross Smith wrote: From: Hans Horn [mailto:[EMAIL PROTECTED]] // an element of a linked list typedef struct node { node (int _v, node* _n) : v(_v), next(_n) {} int v; node* next; }; I'm surprised that compiled; the typedef shouldn't be

Re: limit for # of items created with new ?

2002-09-24 Thread Christopher Faylor
On Tue, Sep 24, 2002 at 07:33:03PM -0700, Randall R Schulz wrote: At 19:10 2002-09-24, you wrote: On Tue, Sep 24, 2002 at 07:06:22PM -0700, Randall R Schulz wrote: Hans, ... If I understand correctly, all Cygwin app memory comes out of a single pool (hence the Cygwin heap size registry

Re: limit for # of items created with new ?

2002-09-24 Thread Hans Horn
List, I re-did the same test (test.cpp), this time using entirely malloc instead of new. The death occurs at the exact same place as with new. Some internal gcc/bash/cygwin limit I'm hitting there? H. -- Unsubscribe info: http://cygwin.com/ml/#unsubscribe-simple Bug reporting:

RE: limit for # of items created with new ?

2002-09-24 Thread Ross Smith
From: Emil Briggs [mailto:[EMAIL PROTECTED]] On Tuesday 24 September 2002 10:28 pm, Ross Smith wrote: From: Hans Horn [mailto:[EMAIL PROTECTED]] // an element of a linked list typedef struct node { node (int _v, node* _n) : v(_v), next(_n) {} int v; node* next;