Greetings! In the forthcoming gcl release, we will be able to efficiently make use of >2Gb or physical ram.
As we compile and load code produced by gcc, possibly arbitrarily at the users command, this runs afoul of the default memory model used by gcc, which assumes code refers to addresses at most 2Gb away. There is a large memory model free from this restriction, but in the recent past I have found it to have bugs, which is not surprising given its likely lack of testing relatively speaking. To this end, GCL has a variable si::*code-block-reserve* which can reserve low memory for code early on and use this until exhausted. This is of course a heuristic workaround. Nevertheless a 10Mb block appears quite effective. The alternative is to limit the heap to 2Gb, which can be done with an environment variable. What would most users like to see? Take care, -- Camm Maguire [email protected] ========================================================================== "The earth is but one country, and mankind its citizens." -- Baha'u'llah _______________________________________________ Gcl-devel mailing list [email protected] https://lists.gnu.org/mailman/listinfo/gcl-devel
