| |
| |
| |
|
Page: 1 2 3 4
Comments:
<0> how is everyone?
<1> dpends, you ?
<1> err
<1> depends*
<1> woah, it's the same in french but without the accent.
<2> Hello, does anybody know of an ***embly project which rewrites some of the C string functions, like strlen and strcpy, etc.. ?
<2> Hello?
<3> no
<4> Hmm. Does anyone know of a technique where I can use the floating-point registers to manipulate 64-bit integer data without loss of precision?
<5> no
<5> why not just write a set of 64bit math routines
<4> Hmm. I'm using code automatically written by g++, but it's going slowly. I benchmarked it.
<4> Basically, I'm trying to get binary Euclid's algorithm done for long long integers.
<1> Zemyla: in asm ?
<5> what did you benchmark? are you really going to be dividing/shifting enough to see a perf hit because of it?
<4> It's for reducing the numerator and denominator of a fraction.
<4> It's going to be called on every single fraction operation.
<5> ah
<5> vick: the amd docs have an optimized memcpy, that's all i can recall
<6> mornign
<2> Hello
<2> Does anybody know a good asm page or something that teaches handling strings ?
<7> vick: see /topic of this channel
<7> are interrupts a software creation or a hardware thing?
<8> hardware
<8> and usually the kernel is the only software dealing with them
<7> oggis_: yea, cpu provides commands (e.g. INT) to cause interrupts but someone needs to install interrupt event handler routines to deal with 'em right?
<7> oggis_: handlers usually can only be the kernel or the BIOS right?
<8> yeppp
<7> oggis_: k, thanks
<8> moreover, there are different types of interrupts... hardware generated (by hardware devices, for example keyboard), software generated (by int instruction) and exceptions (like #gp0 etc)
<9> I have a question
<9> on x86 processors,
<9> if you run a mfence instruction, is the address to put a store guarenteed to be evaluated at the time of the mfence?
<9> or does it possible that the cpu precomputes the store memory address and sfence only forces it to be committed
<10> hello
<10> can anyone tell me how could I do if I want this to work:
<10> movl (%eax), 0x20(%ebp)
<10> this is I want to move the adrres pointed to by %eax into the stack
<11> hmm.
<11> pupy, I thought both addresses cannot be indirect.
<11> or registers.
<12> memory to memory moves are not allowed
<10> not they can't, that's why I'm asking how to do this
<12> reg to reg I believe is allowed though
<11> pupy
<11> what about:
<10> yes it is
<11> mov (%eax), %eax
<11> mov %eax, 0x20(%ebp)
<10> uhm, didn't try, and then moving it
<10> nice one, let me try
<11> :)
<11> I'm super newbie in ***embly programming, but I thought that could be the way.
<11> deductional thinking based on study of intel instruction format.
<11> Thought.
<11> there may be instructions which allow it straight.
<12> Faster than a speeding compiler error, able to cause a gpf in a single line, its super newbie
<11> gpf?
<12> General Protection Fault
<11> and what is a speeding compiler error?
<12> No idea
<10> no, it doesn't work, though I should in theory
<11> pupy, weird. then I must hear why it doesn't.
<11> if you find out, could you tell me?
<10> well, I got this: eax 0x804912a
<10> and then, a movl (%eax), %ebx gives: ebx 0x6564796d
<12> that would be the value at (%eax)
<10> so it's pasing "my_defvalue" to ebx
<11> what is movl? NASM manual doesn't tell.
<10> this is gnu gas, same as mov long
<12> that would be at&t syntax
<12> I use fasm
<11> it may be the reason why my thing doesn't work, I thought the l is something else behind it.
<11> try:
<12> do for me its mov dword ptr
<11> mov (%eax), %eax
<11> mov %eax, 0x20(%ebp)
<11> argh
<11> movl*
<11> movl (%eax), %eax etc.
<10> that's the same as before
<10> it just p***es the value but not the memory address
<11> wait, you said it is p***ing my_defvalue to ebx?
<11> insides of it or pointer?
<12> heres one, push 0x0; ret
<10> yes, 0x6564796d
<12> lol
<10> no Cheery, let me explain
<10> 0x0804912a mydef_value
<11> thank you a lot beforehand.
<10> that's a string which holds "ny_def_value"
<10> uy
<11> aah.
<10> "mydef_value", then 0x0804912a is a pointer to it's first character on memory, where it is located
<11> what do you want to do for it?
<10> to p*** 0x0804912a to 0x20(%ebp)
<10> the adress, not it's value
<11> aah.
<12> I am soo glad I learned C before I tried ***embly language
<11> if that address is in %eax, you could just do: movl %eax, 0x20(%ebp)
<11> __mikem: why? I'm glad I left C behind. :)
<10> well, that the first I tried, though it doesn't work
<12> Because I could learn about pointers from the safety of a high level language, before I had to learn about it the hard way
<11> oh. I loved to get the final catch from forth.
<12> Cheery, what do you use if you don't use C
<11> __mikem: there's lisp, python, forth, etc.
<12> Hey cheery, I think I have just the compiler for you, hold on
<11> I'm not feeling confident in C environment.
<11> really? intresting.
<11> (do not propose brain****)
<11> I've seen it already.
<12> http://www.freebasic.net/ This is a basic compiler, but it has support for pointers, and can produce code that is every bit as fast as C
<12> It even supports inline asm
<11> heh, that sounds intresting.
<11> but where can I find some code samples, and does it work in linux?
<12> Yes, it works on linux, and code samples are available, let me just find them
<11> I'm interested to see if there's something worth watching or is it just an another programming language. :)
<12> It is probably the best compiler I used that wasn't a C compiler
<12> It has support for a larg veriety of libraries, you can get the entire list at http://www.freebasic.net/index.php/link
<11> kewl, but I'm looking for innovations, something I can use already without changing the language I'm working on.
<12> I figured that if you didn't like C, a basic compiler that can write high speed code would be a dream come true
<11> I don't know.
<11> In my dreams there are somekind of forth variant.
<12> Talk about last century
<11> Hmm. why do you think so?
<11> I remember C being older than forth.
<12> YEs, but people still use C more often than they use forth
<11> There are problems in C design I do not like about. Even it has a good compiler.
<11> first, C has no interactive interpreter.
<12> Last I checked forth was specifically designed for AI
<12> or was that smalltalk I am thinking of
<11> lisp?
<12> maybe
<11> C has no capabilities to modify code, only very limited capabilities to handle automatic code generation.
<12> Heres a question, have you ever tried coding in C
<12> Or C++ for that matter
<11> yes, I used it three years.
<12> Just wondering
<11> (C++)
<11> what do you wonder?
<12> Its just sometimes, you get people who tried C, couldn't make sence of it (or get their code to compile) and just dismissed it as a bad language inorder to save pride
<12> Same goes for C++
<11> aah. I'm not belonging to that group.
<12> For me learning C and C++ was a long hard battle, but I always thought it was worth the trouble
<11> I actually managed to work with C, but I liked more about other languages.
<11> And understood that they had things I couldn't ever have in C.
<12> You remind me of Randyll Hyde
<11> Randyll Hyde?
<12> He said the same exact thing about C
<12> He is a very well respected expert in ***embly Language Programming
Return to
#asm
or
Go to some related
logs:
#css
ps2pdf centos rpm
grub loading error 17 ubuntu
public key 5ebd2744
wusb11 fc4
lighttpd cannot load mysql extension
#perl
#php
+ubuntu +build_essential
corrupt symc8xx.sys
|
|