Re: register allocation (was: Spilling problems)

2004-08-07 Thread Dan Sugalski
At 9:51 AM +0200 8/6/04, Leopold Toetsch wrote: Dan Sugalski wrote: ... extraordinarily large (60-80k lines) ... ... dying with an out-of-memory error. How many symbols does the sub have? Please set a breakpoint at imcc/reg_alloc.c:468 and inspect: n_symbols (or just create a debug print

register allocation (was: Spilling problems)

2004-08-06 Thread Leopold Toetsch
Dan Sugalski wrote: ... extraordinarily large (60-80k lines) ... ... dying with an out-of-memory error. How many symbols does the sub have? Please set a breakpoint at imcc/reg_alloc.c:468 and inspect: n_symbols (or just create a debug print statement there) The interference_graph size is

Re: Spilling problems

2004-08-05 Thread Leopold Toetsch
Dan Sugalski wrote: ... In this case I'm hitting the double spill error, but this is, I expect, tied in with the infinite loop the register spiller hits on some code. Should be fixed now. Hopefully. - There were 2 bugs in the code WRT calculating life range of spilled regs and the ordering of

Re: Spilling problems

2004-08-05 Thread Dan Sugalski
At 9:44 AM -0400 8/5/04, Dan Sugalski wrote: At 11:47 AM +0200 8/5/04, Leopold Toetsch wrote: Dan Sugalski wrote: ... In this case I'm hitting the double spill error, but this is, I expect, tied in with the infinite loop the register spiller hits on some code. Should be fixed now. Hopefully. -

RE: Spilling problems

2004-08-05 Thread Dan Sugalski
-of-memory error. (Which is also a potential compromise/DOS attack on parrot, so it needs fixing for security reasons too) -Original Message- From: Dan Sugalski [mailto:[EMAIL PROTECTED] Sent: Thursday, August 05, 2004 9:44 AM To: Leopold Toetsch Cc: [EMAIL PROTECTED] Subject: Re: Spilling

RE: Spilling problems

2004-08-05 Thread Dan Sugalski
: Dan Sugalski [mailto:[EMAIL PROTECTED] Sent: Thursday, August 05, 2004 11:15 AM To: Butler, Gerald Cc: [EMAIL PROTECTED] Subject: RE: Spilling problems [Cc'd back to the list, since it's of general interest] At 9:51 AM -0400 8/5/04, Butler, Gerald wrote: I hate to intrude on this discussion, but, I

FW: Spilling problems

2004-08-05 Thread Butler, Gerald
Oops. Sorry, meant to post this to the list. -Original Message- From: Butler, Gerald Sent: Thursday, August 05, 2004 9:52 AM To: 'Dan Sugalski' Subject: RE: Spilling problems I hate to intrude on this discussion, but, I was wondering if anyone could give a brief explanation (or point

RE: Spilling problems

2004-08-05 Thread Butler, Gerald
- From: Dan Sugalski [mailto:[EMAIL PROTECTED] Sent: Thursday, August 05, 2004 11:15 AM To: Butler, Gerald Cc: [EMAIL PROTECTED] Subject: RE: Spilling problems [Cc'd back to the list, since it's of general interest] At 9:51 AM -0400 8/5/04, Butler, Gerald wrote: I hate to intrude

Spilling problems

2004-08-04 Thread Dan Sugalski
Well, I thought this was only an issue with massively large programs (of which I have *far* too many), but I'm making IMCC fall over and die with subs of less than 6k-lines of PIR. (5945 lines, counting the sub start and end lines, with old style (.pcc_begin/.pcc_end) function calls) In this