tag:blogger.com,1999:blog-5246987755651065286.post6617099516599046012..comments2024-02-22T16:15:42.388-08:00Comments on cbloom rants: 01-27-09 - Oddworld Memory Memoriescbloomhttp://www.blogger.com/profile/10714564834899413045noreply@blogger.comBlogger4125tag:blogger.com,1999:blog-5246987755651065286.post-64518940736291637722009-01-28T10:34:00.000-08:002009-01-28T10:34:00.000-08:00Er, but the main part was it wasn't just all small...Er, but the main part was it wasn't just all small blocks, it was just a hard-coded list of "problem" sizes that was even sparser and more random than the SW one.Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-5246987755651065286.post-4079039912786116692009-01-28T10:33:00.000-08:002009-01-28T10:33:00.000-08:00Yep, to make Terra Nova run in 8M (or was it 4M?) ...Yep, to make Terra Nova run in 8M (or was it 4M?) I made the standard malloc detect a certain set of fixed sizes and allocate them each from their own small heap.<BR/><BR/>The actual issue was really to prevent fragmentation, though; after running for a while on a level we started thrashing from not being able to load a particular enemy model, which was 64K, and we had more than 64K free but it was fragmented to hell and back.Anonymousnoreply@blogger.comtag:blogger.com,1999:blog-5246987755651065286.post-35665999798894398442009-01-28T07:18:00.000-08:002009-01-28T07:18:00.000-08:00Funny, we ended up rebooting between levels in Mec...Funny, we ended up rebooting between levels in MechAssault 1 as well (we had similar problems with wild dynamic allocations).Noelhttps://www.blogger.com/profile/10216065035079926033noreply@blogger.comtag:blogger.com,1999:blog-5246987755651065286.post-7188848422726504472009-01-28T06:22:00.000-08:002009-01-28T06:22:00.000-08:00I remember Sean doing something very similar when ...I remember Sean doing something very similar when we were running out of memory in Terra Nova back in 1996.dfanhttps://www.blogger.com/profile/16523251716744122695noreply@blogger.com