flat assembler
Message board for the users of flat assembler.
Index
> Unix > Disassembled .o shows only file format, |
Author |
|
cvfirefox 17 Oct 2008, 01:55
Code: # Copyright 1998-2004 - Just For Fun Software, Inc., all rights reserved # Author: George Woltman # Email: woltman@alum.mit.edu # # ******************************************************** # ******************************************************** # ******************** FFT MACROS ********************** # ******************************************************** # ******************************************************** ## Perform a 32-element FFT. .macro fft32 LOCAL b1b, c1b ## Do a forward FFT only .if type EQ 1 call x32_fft jmp xmiddle_1 .endif ## Do a squaring .if type EQ 2 call x32_fft call xmiddle_2 ## jmp x32_finish_unfft .endif ## Do a multiply .if type EQ 3 call x32_fft call xmiddle_3 jmp x32_finish_unfft .endif ## Do a multiply with pre-FFTed inputs .if type EQ 4 call xmiddle_4 jmp x32_finish_unfft .endif ## Do FFT levels 1,2,3 ## Values 0-31 is real data. ## ## On input the 32-byte cache lines hold these data values: ## 0 16 1 17 ## 2 ... ## ... ... ## 14 ... ## On output the 32-byte cache lines hold these data values: ## 0 4 1 5 ## 2 ... ## 8 ... ## ... ## Do 4 eight_reals_first_fft macros ## distance between fft data elements is 4 .if type EQ 1 x32_fft: copy_7_words subl %eax,%eax b1b: disp eight_reals_first_fft, 4*dist1, 8*dist1, 8 leal dist1(%esi),%esi ## Next source pointer addb $256/4, %al ## Test loop counter jnc b1b ## Iterate if necessary leal esi-4*dist1,%esi ## Restore source pointer ## End common FFT code ret .endif ## Do inverse FFT levels 1,2,3 ## On input the 32-byte cache lines hold these data values: ## 0 4 1 5 ## 2 ... ## 8 ... ## ... ## On output the 32-byte cache lines hold these data values: ## 0 16 1 17 ## 2 ... ## ... ... ## 14 ... ## Do 4 eight_reals_last_unfft macros ## distance between fft data elements is 4 .if type EQ 2 x32_finish_unfft: c1b: disp eight_reals_last_unfft, 8, 4*dist1, 8*dist1 leal dist1(%esi),%esi ## Next source pointer addb $256/4, %al ## Test loop counter jnc c1b ## Iterate if necessary leal esi-4*dist1,%esi ## Restore source pointer fft_3_ret .endif .endm .end Hey guys, when I asssembled this AT&T macro file to .o and I disassembled it with objdump, it only shows the file format instead of the binary and source codes. What is wrong with the source codes here? Need to know it because it is urgent. _________________ -If you ain't first, then you're last. |
|||
17 Oct 2008, 01:55 |
|
dogman 23 Jul 2013, 14:01
Depends whether you write the recipe on rice paper or eggroll skins. But yeah
To the OP, write an assembly stub and invoke your macro! |
|||
23 Jul 2013, 14:01 |
|
bitRAKE 23 Jul 2013, 18:44
There is an updated version of these codes in MASM format, easier to convert, imho. Hm...over four years - seems OP solved his problem.
https://github.com/rudimeier/mprime/tree/master/gwnum |
|||
23 Jul 2013, 18:44 |
|
dogman 23 Jul 2013, 18:49
Haha sorry about that. I guess this forum hasn't had too much activity. I saw it in the first few posts and didn't look at the date.
|
|||
23 Jul 2013, 18:49 |
|
< Last Thread | Next Thread > |
Forum Rules:
|
Copyright © 1999-2025, Tomasz Grysztar. Also on GitHub, YouTube.
Website powered by rwasa.