Re: [fricas-devel] Re: Compiling from source with sbcl 1.3.6 - heap exhausted

2016-06-14 Thread oldk1331
Thanks Ralf, patch is updated: https://github.com/oldk1331/fricas/commit/03d432c7c49648903fc5fa35a570f2a425feea40 Now texmacs.spad only takes 680MB RAM and 10s to compile. -- You received this message because you are subscribed to the Google Groups "FriCAS - computer algebra system" group. To u

Re: [fricas-devel] Re: Compiling from source with sbcl 1.3.6 - heap exhausted

2016-06-14 Thread Ralf Hemmecke
On 06/14/2016 08:53 AM, oldk1331 wrote: > Modify Cork to be a nullary function that returns constant makes > compilation much faster and memory usage much smaller: > > Cork():List Record(corkcode:String,unicode:SI) == [ Maybe better ... CorkFun(): HashTable ... == Cork: HashTable := C

[fricas-devel] Re: Compiling from source with sbcl 1.3.6 - heap exhausted

2016-06-13 Thread oldk1331
I just tested with sbcl 1.3.6, turns out heap of 3000MB is not enough, 3600MB is OK. The reason that sbcl requires so much memory and takes a long time to compile texmacs.spad, is the constant Cork which has 800 elements. Maybe sbcl tries to optimize this big variable, but in fact Cork is a consta

Re: [fricas-devel] Re: Compiling from source with sbcl 1.3.6 - heap exhausted

2016-06-13 Thread Alasdair McAndrew
Thank you! - in the end I used 8000 and it worked fine. 'Tis always the way: post a question, and then minutes later find the solution myself. -Alasdair On Mon, Jun 13, 2016 at 10:07 PM, oldk1331 wrote: > From last discussion, I think you should use at least 3GB heap, > try "--dynamic-space-si

[fricas-devel] Re: Compiling from source with sbcl 1.3.6 - heap exhausted

2016-06-13 Thread oldk1331
>From last discussion, I think you should use at least 3GB heap, try "--dynamic-space-size 3000" or more. -- You received this message because you are subscribed to the Google Groups "FriCAS - computer algebra system" group. To unsubscribe from this group and stop receiving emails from it, send