libera/#commonlisp - IRC Chatlog
Search
18:40:02
luis
Someone at ELS did some work on method inlining, or some sort of class sealing and I can't recall who. Does anybody have any idea what I'm talking about?
19:21:14
Bike
Is there any way to wrap quicklisp dependency resolution around a dynamic extent? So that e.g. (with-quicklisp (asdf:load-system :foo)) will try getting the system from quicklisp if it's not available. I think quicklisp internally establishes a handler on asdf:missing-dependency
19:48:06
Bike
the way quicklisp does it internally involves asdf internals and some kind of table to avoid loops
20:24:13
pjb
Bike: (defmacro with-quicklisp (load-form) `(ql:quickload ,@(rest load-form))) (with-quicklisp (asdf:load-system :foo)) #| ERROR: System "foo" not found |#
22:06:23
seok
is there a resource for clojure from the perspective of a common lisp programmer? or a reference to the differences
23:03:24
nij-
I'm doing a long computation with many intermediate result. At the end, I need to analyze that result. However, the intermediate results are too large, and sometimes it fill the heaps. I wonder if there are some data structure that is persistant and doesn't use memory at all (hopefully, it feels like using a hash table)?
23:04:36
nij-
? what's wrong? I don't mind the poset-analysis being slow, so I think it's fine if it's backended by files on disks.
23:06:56
nij-
Yeah, i'm really looking for an existing CL system that does that for me (with nice abstraction, so I feel like I'm using a hash table).
23:30:06
nij-
> LevelDB is not an SQL database. Like other NoSQL and dbm stores, it does not have a relational data model and it does not support SQL queries.
23:31:19
_death
like a hash table, basically.. (leveldb:puts db "Hello" "World") (leveldb:gets db "Hello") => "World"
23:33:04
_death
it doesn't do that out of the box.. you will need to encode them into octets or strings
23:36:39
_death
again, you'll need to do that yourself or use another system.. leveldb is rather low level
23:39:08
_death
maybe fact-base or something.. I don't remember if it stores everything in memory or not, though
5:14:42
beach
nij-: I am not talking about RAM, but about virtual memory. You are about to implement your own virtual memory it seems.
5:20:05
beach
Or SSD. That's what virtual memory does. And it seems to me that you want to use external memory anyway, to implement a specific "virtual memory" yourself, which is bound to be worse.
5:21:40
beach
Either way, if you are going to use disk memory, you are talking many orders of magnitude slower than RAM, so you are probably in for some performance problems anyway.
5:22:29
beach
With a week or so worth of salary (which is at least what it is going to take you to implement your virtual memory), you can buy a decent computer.
5:24:55
hayley
(Reminds me to get back to my FPGA stuff; my back-of-hand estimations for the space and time needed aren't pretty, and so I am looking into custom hardware.)
5:25:05
beach
I suppose that depends on your computer. But RAM is not very expensive these days. At least not compared to the effort you will have to make to get this thing to work.
5:28:21
beach
nij-: Did you check how much RAM that computer can have? It has got to be in the specifications for it.
5:28:36
hayley
Using the internal SSD as swap space should be...bearable. From memory, macOS will let you swap until you run out of disk space.
5:31:12
beach
In my book, the best approach is what makes me do the least amount of work, within a reasonable cost, of course.
5:35:20
beach
nij-: That question is way too general. It depends on the density of your data for one thing.
5:35:53
beach
If it is very dense, an array is fine. If it is sparse, there are tons of possibilities depending on the exact structure.