IRC Logs

Network: freenode Channel: #csharp

	Enter your search terms Submit search form
	Web www.irclog.org

[16:34]<wyfaj`>Remi_Woler: the first time you invoke a method for example, it'll basically be compiled to native code by the JIT compiler on the fly... which is why it can be optimized for the current processor etc
[16:34]<zndy_wjfnz>or is that because not all the variables are not known at compile tome?
[16:34]<wyfaj`>Remi_Woler: there is some overhead, but only once
[16:34]<dzdyzzz-wjzc>christ, HD's have gone back to RLL?
[16:34]<fyfuym-syfggag>damieng-work: ?
[16:34]<wyfaj`>once a method is jitted, it'll run "at native speed" in subsequent calls
[16:34]<azyfm>which accounts for some of the long startup time of .NET apps
[16:34]<dzdyzzz-wjzc>Liquid-Silence, different encoding scheme.
[16:34]<zyccus>damieng-work: foldoc
[16:34]<zndy_wjfnz>err, I mean: Not all the variables are known at runtime :)
[16:34]<fyfuym-syfggag>O_0
[16:34]<zyccus>damieng-work: I have no idea what HDs use... I don't do hardware
[16:35]<fyfuym-syfggag>anyway
[16:35]<fyfuym-syfggag>irc is way to discracting
[16:35]<fyfuym-syfggag>distracting
[16:35]<wyfaj`>discracking?
[16:35]<fyfuym-syfggag>speak laters
[16:35]<fyfuym-syfggag>man
[16:35]<dzdyzzz-wjzc>the basic problem with magnetic media is you can't just map 0 to no charge and 1 to charge because you get too many charges or no-charges together. resulting in bleed.
[16:35]<fyfuym-syfggag>I have had 3 hours sleep this weekend
[16:35]<fyfuym-syfggag>tiredness
[16:36]<zndy_wjfnz>wilco`: If I say: for(int 1=0; i < 10; i++) <== How many times is that compiled? 1 time? Or 11 times?
[16:36]<wyfaj`>1
[16:36]<zyccus>damieng-work: should save in denary then ;)
[16:36]<azyfm>it's compiled once
[16:36]<azyfm>(unless, for some reason, the JIT compiler discards the compiled code from the cache)
[16:36]<zndy_wjfnz>wilco`: But how would the compiler know for which value of i it needs to compile?
[16:36]<wyfaj`>heh
[16:37]<wyfaj`>that question doesn't make sense really ;)
[16:37]<zndy_wjfnz>hmm
[16:37]<wyfaj`>Remi_Woler: after jit compilation i is still a variable like it was before...
[16:37]<wyfaj`>it's just compiled to a different level, basically
[16:37]<trc>I thought the ## channels were supposed to be topic channels that weren't owned
[16:37]<wyfaj`>gtg now - will be back tonight
[16:37]<trc>(@topic)
[16:38]<azyfm>Remi_Woler: At the x86 level, i would be stored either in a register or a memory location
[16:38]<zndy_wjfnz>well, I read somewhere that there is a differance between compiletime and runtime (duh), and that some code-compilation was stalled to runtime, because the variables are unknown at compiletime
[16:40]<dzdyzzz-wjzc>Generics are like that.
[16:40]<zndy_wjfnz>so that is where the 'how many times' question came from
[16:40]<zyccus>that's how generics are different from templates
[16:40]<zyccus>well, one way
[16:41]<dzdyzzz-wjzc>I'm surprised nobody has built compression into HD's to be honest.
[16:41]<dzdyzzz-wjzc>sure it can be done at OS level but it can't take advantage of the +- charge side.
[16:42]<zndy_wjfnz>damieng-work: patent the idea, sell it to Seagate :p
[16:42]<zyccus>surely the HD manufacturers squeeze the most use from the space they can already?
[16:42]<dzdyzzz-wjzc>Heh. you could probably start with modified huffman and adjust to ensure a maximum charge or no charge length
[16:42]<dzdyzzz-wjzc>rikkus, it would need assistance from the OS probably.
[16:43]<zyccus>oh
[16:43]<dzdyzzz-wjzc>you'd basically have to move all FS management to the disk h/w
[16:43]<dzdyzzz-wjzc>which wouldn't be a bad thing...
[16:43]<trc>that seems like a bad-ish idea
[16:43]<zyccus>hardware or firmware?
[16:44]<dzdyzzz-wjzc>the firmware could do it.
[16:44]<zyccus>Tak: quite a lot of the drive management is done by the drive itself anyway
[16:44]<trc>"Sorry, this HD only supports WinFS..."
[16:44]<dzdyzzz-wjzc>no...
[16:44]<dzdyzzz-wjzc>more like "this hd implements a filesystem HDFS"...
[16:44]<dzdyzzz-wjzc>that now works with linux, osx, windows...
[16:45]<trc>what if you don't like hdfs?
[16:45]<zyccus>you'd have to layer NTFS, reiserfs etc. on top
[16:45]<trc>hm
[16:45]<dzdyzzz-wjzc>then buy a HD without it.
[16:45]<dzdyzzz-wjzc>or switch the HD to managed mode.
[16:45]<trc>as long as there's a managed mode, that'd be fine
[16:45]<dzdyzzz-wjzc>Wonder if apple will switch from hfs+ to zfs for next release....
[16:45]<zyccus>damieng-work: so would NTFS etc. be a thin wrapper over HDFS when using a drive that did it?
[16:45]<dzdyzzz-wjzc>guess so.
[16:45]<zyccus>that could work then
[16:46]<dzdyzzz-wjzc>I'm not sure all the effort would be worth it though. especially considering the rate at which hd's increase in capacity.
[16:46]<dzdyzzz-wjzc>750GB now iirc?
[16:46]<zyccus>something like that
[16:47]<zyccus>I just got a couple of 160GB drives at work and I have no use for the space... I was on a couple of 40s before
[16:49]<zndy_wjfnz>heh, I have 2x250GB in my system for storage only, and I hardly have 10Gb left :p
[16:49]<zyccus>Remi_Woler: too much pr0n?
[16:50]<zndy_wjfnz>nah, I delete that after I watched it :p
[16:50]<dzdyzzz-wjzc>Well this 100gb in my laptop is getting a bit tight... but thats mostly because i only gave 20gb to the windows partition... doh.
[16:50]<dzdyzzz-wjzc>so wow and psychonauts live on my ipod :D
[16:50]<dzdyzzz-wjzc>portable apps indeed.
[16:50]<zyccus>the only place I take up room is with TV programmes I haven't watched yet, DVDs too big to fit on a single disk, and music... my home machine is running low on space
[16:51]<trc>ditto
[16:51]<zyccus>but even with the 2G databases I deal with at work, I can't fill 80G
[16:51]<dzdyzzz-wjzc>if i want to keep a video file i've downloaded i normally buy the dvd.
[16:51]<zyccus>and I have a couple of virtual machines
[16:52]<zyccus>I only download UK TV programmes... which you can't get on DVD, and I wouldn't pay for anyway because if I had to, I'd just remember to watch them :)
[16:52]<zyccus>or use that machine under the telly... I think it's called a vidio
[16:52]<mjzymygo>Rikkus: you'll enjoy drive space then haveing to de-fuck a 90GB database within its 220 GB transaction log...
[16:53]<zyccus>Moridin8: rather you than me
[16:53]<mjzymygo>me remember bad things from a year ago
[16:53]<mjzymygo>(doh)
[16:53]<zyccus>de-fuck in that case would be 'DBCC FIXEVERYTHING and see if it fixes it.. if not, restore the backup'
[16:54]<mjzymygo>Rikkus: nah, the DB was fucked internally. had to run everything out and rebuild.
[16:54]<mjzymygo>using the transaction log.
[16:54]<zyccus>fun fun fun
[16:54]<mjzymygo>which was difficult because the data kept comming out fucked.
[16:54]<zyccus>what server software?
[16:55]<mjzymygo>teach people to set proper admin/maint/repair schedules on the databases won't it
[16:55]<mjzymygo>2 * Win2000Advanced clustered
[16:55]<zyccus>SQL Server presumably
[16:56]<mjzymygo>Sql-Server 2000 Enterprise SP2 (because the idiots hadn't done SP3 because of its cross-chain security thing that is optional anyway)

Back to #csharp
Or go to some related logs:
#ubuntu
#python
#blender
#gentoo
#php