| |
| |
| |
|
Page: 1 2 3 4 5 6 7 8 9 10 11
Comments:
<0> sentox
<1> _fury: I tried that in the past, with vc++ 7.1 and it was indeed very slow
<2> yeah
<2> streams are typically slower
<3> what?
<3> php isn't compiled.
<2> http://rafb.net/paste/results/GeL7Rk24.html
<2> that code is 2 seconds
<2> my istream_iterator code is 28 seconds
<2> and php using regex ( +) replacement is 8 seconds
<2> on the same test data
<2> so what have we learned ;)
<4> where's your C++ codfe?
<2> http://rafb.net/paste/results/puaG6b65.html
<2> uses remove_if with a mutating predicate, and istream_iterator/ostream_iterator
<4> hm
<1> Your first code is not using streams ;)
<2> well stream iterators
<2> I suspect plain streams might be slower
<2> here's a cleaner version http://rafb.net/paste/results/XYwZsT38.html
<2> so what I did, for when I hit ctrl+F5 when mIRC has focus, because I do that *all* the time when I go to build something, is:
<2> /cf5 say I am a ****ing idiot.
<2> try it.
<5> so, AIM opened up their code API or something today?
<6> heh
<6> something remotely along those lines.
<7> Alright, I'm confused. IT seems Boost won't build the Thread library because it's missing "windows.h"... Is that supposed to be part of the platform SDK?
<1> The windows platform sdk, yes
<8> hi :)
<7> But I have it installed... *grumbles*
<8> anyone here using the vc++ 2005 express?
<1> Yeah, but perhaps you got to set a path somewhere
<7> It's in the VC++ Include directories.
<8> It's pretty nice.. but if anyone's using it.. could you tell me what you use to make your resources (dialogs etc) for a win32 native app?
<8> it seems to only want you to design forms for CLR apps :)
<7> Well, whatever. Now it works, and I haven't changed a damned thing.
<7> That's why I hate all thigns related to building and compilation... Bleh.
<2> very interesting, this code is considerably faster when compiled with VC71 compared with gcc4 on linux
<2> using stream iterators
<9> did you build with optimizations?
<2> using fopen(), fread() and fwrite(), it is about the same
<2> yes
<2> -O3
<2> I mean its more than double the time. the same operation averages 18 seconds slower on linux
<2> CPU is probably the bottleneck though
<2> this cpu is faster
<5> same hardware?
<5> oh heh... well duh.
<2> but I would think it would not matter since the disk should be the bottleneck
<5> yes, but you cannot comapre apples to apples unless you use apples
<2> yeah
<2> but again
<2> The C is the same speed
<2> fopen() etc
<2> its only the iostreams that are slower
<2> the faster machine is a Athlon64 3200+, and the slower unix machine is a P3 933
<2> the C is the same speed
<5> thats a damn big difference
<2> yeah
<5> and if your in the multiple second range, there is tons of CPU and disk involved.
<5> why is the C the same speeD?
<2> but again the interesting thing is that the C is the same speed on both CPUs
<2> why? I have no idea
<5> oh.
<2> here is the code, needs boost to compile: http://rafb.net/paste/results/4qElpq97.html
<8> dunno if that would be fair to even call it comparing apples to oranges.. more like apples to planets :D
<5> Well, the Disk I/O is also related to memory and CPU performance as well as hard drive controller, hard drive, etc
<2> I just got carried away benchmarking php regex with c/c++
<2> and now I' mjust wasting time
<2> Twister the faster machine is a SATA WD Raptor 10K rpm drive
<2> the slower one is IDE 7200RPM
<5> well geezuz crmainy.
<5> Of course its gonna be faster
<5> what the hell.
<2> and again, still the same speed with C
<2> it doesn't bother me that the C++ is faster, it bothers me that the C is the same
<5> Well, obviously the other machine has a huge advantage
<5> The C doing I/O ?
<2> yes
<2> I suppose the C++ streams are more CPU intensive
<2> while the C code is very basic
<5> well... hmm
<5> get a GCC port fer winders
<5> then test, then draw conclusions that are actually valid :)
<2> listen
<5> I hear ya.
<2> its a valid conclusion that the C takes the same amount of time on a much slower processor and disk
<5> I get it.
<5> But you have no idea why.
<2> while the C++ with stream iterators is significantly slower on the slower machine
<5> Maybe the GCC compiler is poorer at c++
<2> I just find that interesting at all
<2> at=is
<5> C compilers are often on an even plane
<5> or near even
<5> C++ compilers seem to fluctuate.
<5> Especially when you start optimizing
<2> I mean, the C stuff would probably compiler with a C compiler, but by C code I mean that it uses C I/O instead of C++
<5> All I am saying is compare the generated ASM if you want to really figure it out
<2> yeah
<2> I'm intrigued, but not that intrigued ;)
<2> besides I don't know how to get gcc's ASM output
<2> I know how with VC
<5> haha
<2> haha?
<2> there's nothing funnya bout that.
<5> well, I am outahere. Interesting observation... but not interesting enough to lose (more) sleep over :)
<5> the haha was for the intrigued part
<2> oh
<2> alright goodnight
<3> alright isn't a word.
<3> all right is what you're after.
<9> you'd be surprised how many times I've had people ask me why I say "all right" instead of "alright" haha
<9> mostly on AIM
<10> in english it's a word
<9> no, it's not
<10> prescriptive grammar enthusiasts could have made an argument ten years ago
<10> but descriptively it's usable and in every dictionary
<9> and it's listed as "non-standard"
<10> so was "email"
<10> languages change
<10> welcome to 2006
<9> heh
<10> you wouldn't use an old compiler would you? :)
<11> hey i need help in C
<11> how can i rewrite this : if(x==a) x=b; else x=a; using logic operator... andi ineed to use or ( ^ )
<12> Good evening guys!
<0> Hi PM2
<12> What's up JBlitzen?
<0> Not much
<0> Just discovered that my DVD burner is FAST
<12> lol
<12> Well that's good I guess ;)
<0> 1.5 gigabytes in 2 minutes
<12> Wooo... very good indeed! :)
<0> I thought it had errored out in the preparation phase when it ejected the DVD, but it had actually finished
<3> so?
<0> I should pay more attention to technology
<3> my memory is faster than that.
<0> Your memory is gay
<12> lol
<12> Memory is supposed to be faster ... :P
<12> If it's not, it's probably time to put that 486 away ... :)
<12> Anyways
<0> And this lightscribe stuff is pretty fancy
<0> That part took 10 or 15 minutes
<12> Don't know what it is yet :)
<0> Which, I don't mind
<0> Lightscribe is a technology that uses the same lasers involved in the DVD burning to impart an image to the reverse side of special DVD's and CD's.
<12> Would anyone be willing to tell me what's wrong with this code : http://www.noidea128.org/sourcefiles/16018.html
<3> 486 memory is still faster.
Return to
#c++
or
Go to some related
logs:
Bo Peep to a 19th century cockney street vendor
hotass
#javascript
msr 206
What is the smallest continent?
#networking
#AllNiteCafe
The _____ Revolution began in 1905 on 'Bloody Sunday' when troops fired at demon
peter parka
#linux
|
|