Ask HN: What are some examples of beautiful x86 assembly code?

Post by msl » Fri, Apr 27, 2018, 11:25 AM UTC

chatmasta wrote: post #17

kulu2002 wrote: post #15

This is not x86 but certainly an optimised code https://www.pagetable.com/?p=774

Pretty incredible that such world changing code fits on one page. It’s ineresting that when tools are new, like ASM in 1978, they give high leverage to the first to use them. Microsoft was able to leverage a small amount of code into a world changing platform. Now it would be nearly impossible to do the same with a team the same size. But in 2018, the nascent state of ML tools looks similar to the nascent state of pr…

Post by neilsimp1 » Fri, Apr 27, 2018, 11:49 AM UTC

jbaiter wrote: post #12

I bought a nice little booklet of small x86 gems a few years back[1][2] and this one is my favorite: .loop: xadd rax,rdx loop .loop Fibonacci with two instructions. [1] https://www.amazon.com/dp/1502958082 [2] https://www.xorpd.net/pages/xchg_rax/snip_00.html

Post by giancarlostoro » Fri, Apr 27, 2018, 11:59 AM UTC

jbaiter wrote: post #12

I bought a nice little booklet of small x86 gems a few years back[1][2] and this one is my favorite: .loop: xadd rax,rdx loop .loop Fibonacci with two instructions. [1] https://www.amazon.com/dp/1502958082 [2] https://www.xorpd.net/pages/xchg_rax/snip_00.html

Post by linuxlizard » Fri, Apr 27, 2018, 12:47 PM UTC

mschaef wrote: post #19

msravi wrote: post #7

IBM PC x86 BIOS code: https://github.com/kaneton/appendix-bios OT: This brings back memories of tinkering with the MS-DOS boot process. Back then, the BIOS would read the MBR and copy its contents to 0x7C00 and start execution from there. So you could assemble your own code (using no less than MS-DOS debug) and plonk it into the MBR. I remember doing things like fooling the boot loader into thinking there's less ram…

What was the advantage of this over 'Terminate and Stay Resident'?

Post by vardump » Fri, Apr 27, 2018, 1:17 PM UTC

varjag wrote: post #18

vardump wrote: post #16

Earlier quoted context omitted.

That's remarkably close to 68k str-functions. Page 133 http://www.atarimania.com/documents/Asm_Lang_Prog_68K_Family... * STRLEN - RETURNS LENGTH OP NULL TERMINATED STRING IN D0 * A0 -> STRING STRLEN: MOVE.L A0,-(SP) SAVE REG CLR.L D0 INITIALIZE STRLENI:TST.B (A0)+ NULL? BEQ STRLENR YES, RETURN ADDQ.L #1, D0 BUMB COUNT BRA STRLENI LOOP STRLENR:MOVE.L (SP)+,A0 RESTORE REG RTS We might also want to copy a string: * STRC…

Wow it's close indeed, even stack manipulation looks the same. Would be MOV R0, -(SP) .... MOV (SP)+, R0 for the PDP code.

  > cat > test.asm
  strlen:   movem.l %a0-%a1,-(%sp)
            movl    %a0, %a1     ;# copy a0 to a1
  slenloop: tst.b   (%a0)+
            bne.b   slenloop
            addq.l  #1, %a1
            sub.l   %a0, %a1
            move.l  %a1, %d0
            movem.l (%sp)+,%a0-%a1
            rts
  ^D

  > m68k-linux-gnu-as test.asm && m68k-linux-gnu-objdump -d a.out
  a.out:     file format elf32-m68k
  
  Disassembly of section .text:
  
  00000000 :
     0:	48e7 00c0      	moveml %a0-%a1,%sp@-
     4:	2248           	moveal %a0,%a1
  
  00000006 :
     6:	4a18           	tstb %a0@+
     8:	66fc           	bnes 6 
     a:	5289           	addql #1,%a1
     c:	93c8           	subal %a0,%a1
     e:	2009           	movel %a1,%d0
    10:	4cdf 0300      	moveml %sp@+,%a0-%a1
    14:	4e75           	rts