Boot loader....w/o the...boot loader?

flat assembler
Message board for the users of flat assembler.

Index > OS Construction > Boot loader....w/o the...boot loader?

Goto page Previous 1, 2

Author

Thread

Night Rider

Joined: 28 Jul 2005
Posts: 72

Night Rider

tom tobias, i think that in this code (string output) no difference between OR or TEST (test is boolean operator too). They both put ZF...

04 Aug 2005, 13:39

tom tobias

Joined: 09 Sep 2003
Posts: 1320
Location: usa

tom tobias

smiddy wrote:

In my opinion this reads pretty straight forward to me.

Yup, excellent in my opinion. I absolutely agree. Your solution is perfect, my suggestion, to use TEST, as Night Rider pointed out, is silly.
CMP is the correct instruction to use. Very clear writing Smiddy!

04 Aug 2005, 17:39

LocoDelAssembly
Your code has a bug

Joined: 06 May 2005
Posts: 4633
Location: Argentina

LocoDelAssembly

Code:

PrintMsg:
        mov ah, 0x0E                                 ; teletype output
        mov bh, 0x00                                 ; we are writing on page number 0
        jmp .loadChar

.printChar:
        int 0x10                                     ; call bios video function

.loadChar:
        lodsb                                        ; load [si] to al and increment si
        test al, al                                  ; set zero flag if al = 0
        jnz .printChar                               ; if zero flag is not set, jump to printChar
        ret                                          ; return

It's only a silly modification I know but I think it could be better. If the Opcodes help of MASM (by hutch) is right "test reg, reg" is one clock faster than "cmp reg, immed" and "or reg, reg" on a 386.

04 Aug 2005, 19:43

tom tobias

Joined: 09 Sep 2003
Posts: 1320
Location: usa

tom tobias

Thanks locodelassembly, Well Done! Smile

05 Aug 2005, 15:08

Night Rider

Joined: 28 Jul 2005
Posts: 72

Night Rider

What the difference - use test/or + jf or cmp... or is just nice solution.

05 Aug 2005, 18:37

LocoDelAssembly
Your code has a bug

Joined: 06 May 2005
Posts: 4633
Location: Argentina

LocoDelAssembly

The difference is this loop is one instruction less than the original code and "test reg, reg" takes only one clock on 386. "cmp reg, immed" and "or reg, immed" takes two clocks on 386.

05 Aug 2005, 19:20

LocoDelAssembly
Your code has a bug

Joined: 06 May 2005
Posts: 4633
Location: Argentina

LocoDelAssembly

See attachment for timings, I took this from MASM by hutch, good Docs, bad assembler Razz

Description:	Opcodes help	Download
Filename:	OPCODES.zip
Filesize:	65.4 KB
Downloaded:	291 Time(s)

05 Aug 2005, 23:48

Night Rider

Joined: 28 Jul 2005
Posts: 72

Night Rider

Ok, i have been searching for such document for a long time...

06 Aug 2005, 19:11

Tomasz Grysztar

Joined: 16 Jun 2003
Posts: 7845
Location: Kraków, Poland

Tomasz Grysztar

According to Intel's 80386 Manual from 1986 (I once posted a copy here) "test reg,reg" takes two clocks, too.

06 Aug 2005, 21:22

LocoDelAssembly
Your code has a bug

Joined: 06 May 2005
Posts: 4633
Location: Argentina

LocoDelAssembly

Tomasz: I thought it's not an error of the docs because "test" don't save the result but I was wrong, the doc has a mistake. Thank you.

Night Rider: The new code now it's just "a nice solution" with one instruction less in the loop.

06 Aug 2005, 23:53

Goto page Previous 1, 2

< Last Thread | Next Thread >

Forum Rules:

You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum
You cannot attach files in this forum
You can download files in this forum