01:55:07 Jey [~Jey@114.245.187.10] has joined #ccl 01:57:21 Jey_ [~Jey@114.245.195.251] has joined #ccl 02:00:23 -!- Jey [~Jey@114.245.187.10] has quit [Ping timeout: 252 seconds] 14:18:07 -!- Jey_ [~Jey@114.245.195.251] has quit [Quit: Jey_] 15:34:35 say I have... 27 GB lisp-heap-gc-threshold, 20 GB currently allocated with 10 GB in one big hash table and 8 GB of garbage that could be collected, and a (setf (gethash ...)) makes the HT grow to 15 GB 15:35:00 there's no reason to believe that the GC will fire before the new 15 GB allocation, is there? 15:35:19 i.e., I may end up with 35 GB allocated for a short time 15:49:42 sounds like a gbyers question. Gary? 15:51:48 or maybe rme knows. Matt? 16:49:35 Potentially any allocation can cause a gc. If we can satisfy an allocation without needing to extend the heap, we'll do that (though we might decide to egc first). 16:50:06 -!- sellout [~Adium@c-98-245-162-253.hsd1.co.comcast.net] has quit [Quit: Leaving.] 16:56:21 sure -- but I don't think you expect the GC to happen before the allocation, or do you? 16:56:43 that is, the heap limit is more like a suggestion of when to gc, rather than a hard limit that you are trying to stay under if at all possible 17:05:02 If allocating an object would cross the limit, we'll do a full gc before allocating the object. That way, we might not have to extend the heap, since the gc may have freed enough space. (I hope I'm understanding what you're asking.) 17:10:06 ah okay, hrmn... 17:10:44 so, i have been seeing behavior where i ask for a certain limit, say 110 GB, and the "starting full GC" message says 122 GB are allocated 17:11:41 i was imagining that maybe this big hash table getting resized, and that it was being allocated before the gc was triggered 17:12:13 but it sounds like that's not how CCL works, and that a big allocation like that should instead result in a GC before the allocation attempt 17:29:46 ddp_ [~ddp@216.243.111.165] has joined #ccl 17:32:15 -!- ddp [~ddp@93.182.131.75] has quit [Ping timeout: 252 seconds] 17:32:21 -!- ddp_ is now known as ddp 18:08:41 sellout [~Adium@c-98-245-162-253.hsd1.co.comcast.net] has joined #ccl 18:16:29 -!- dsp_ [~tt@lebesgue.cowpig.ca] has quit [Read error: Connection reset by peer] 18:16:41 dsp_ [~tt@lebesgue.cowpig.ca] has joined #ccl 18:18:42 -!- dsp_ [~tt@lebesgue.cowpig.ca] has quit [Read error: Connection reset by peer] 18:21:41 dsp_ [~tt@lebesgue.cowpig.ca] has joined #ccl 18:24:05 -!- dsp_ [~tt@lebesgue.cowpig.ca] has quit [Read error: Connection reset by peer] 18:26:43 dsp_ [~tt@lebesgue.cowpig.ca] has joined #ccl 18:27:23 -!- dsp_ [~tt@lebesgue.cowpig.ca] has quit [Read error: Connection reset by peer] 18:31:44 dsp_ [~tt@lebesgue.cowpig.ca] has joined #ccl 20:15:59 -!- ddp [~ddp@216.243.111.165] has quit [Remote host closed the connection] 20:16:16 ddp [~ddp@anon-157-36.relakks.com] has joined #ccl