68k details - Page 48

Don_Adan · 10 December 2018, 18:37

Quote:

Originally Posted by Hedeon

What?! Close the longest troll.thread?? Nooo...

Right, dont close this thread. I always want to hear about advantages x86 code over 68k code. And maybe litwr will be soon very good 68k coder. Learning 68k coding is easy enough.

ross · 10 December 2018, 18:53

Quote:

Does it matter though? Yes, especially if there is a vulnerability that can be exploited in the host hypervisor via the guest.

Megol · 11 December 2018, 15:46

Quote:

Originally Posted by ross

It means not understanding how server farms work today and what it means to have a virtual machine run by a complete stranger.
Detect and abuse are one step away..

Intel created the UMIP or User mode Instruction Prevention extension to stop abuse of the SMSW instruction (and some others) _outside_ the virtualization problem.
Sure one could applaud their determination of keeping backwards compatibility with crazy design decisions for the 80286, but couldn't they have added this extension with the 80386 instead of waiting 34 years(!) to patch a known problem?!?

It's just another of those things with x86 that doesn't make sense.

Quote:

Heard by chance of Meltdown and Spectre and their consequences?

Spectre and Meltdown use completely unrelated mechanisms?

ross · 11 December 2018, 17:15

Quote:

Originally Posted by Megol

Spectre and Meltdown use completely unrelated mechanisms?

Right, Meltdown seems to affect only Intel processors and *fully* (HW support) virtualized machines are not affected.
More problematic are Spectre vulnerabilities (bounds check bypass and branch target injection) that can cross between VMs..

alkis · 11 December 2018, 18:23

Ok, how about this: https://projecteuler.net/problem=25

And here is my implementation (noob on asm)

Code:

	;; vasmm68k_mot -Fhunkexe -o euler25 euler25.asm
DIGITS = 1000
NUMSTORAGE = DIGITS/2
        opt x+
	lea.l	-NUMSTORAGE(sp),sp
	move.l	sp,a1
	lea.l	-NUMSTORAGE(sp),sp
	move.l	sp,a2
	lea.l	-NUMSTORAGE(sp),sp
	move.l	sp,a3

	;; clear the numbers
	move.l	a3,a0
	moveq	#0,d0
	move.w	#(3*NUMSTORAGE)/4-1,d1
.clrloop
	move.l	d0,(a0)+
	dbra	d1,.clrloop

	;; init *a1 and *a2 to 1
	moveq	#1,d0
	move.b	d0,NUMSTORAGE-1(a1)
	move.b	d0,NUMSTORAGE-1(a2)
	moveq	#2,d0	; d0 will hold the term-number we are now calculating
	moveq	#1-1,d7		;number of pair of digits, normalised for dbra

	;; n1 ->
	;; n2 ->
	;; n3 ->
	;; n1 ->
	;; n2 ->
	;; n3 ->
	
	;; addfib n1,n2,n3  n3 = n1 + n2
addfib:	MACRO
	addq	#1,d0		;increase the term-number
	;; move num2 to num3
	moveq	#NUMSTORAGE/4-1,d1
	move.l	\2,a4
	move.l	\3,a5
mvloop\@
	move.l	(a4)+,(a5)+
	dbra	d1,mvloop\@

	lea.l	NUMSTORAGE(\1),a4
	lea.l	NUMSTORAGE(\3),a5
	move.w	d7,d1
addloop\@
	abcd	-(a4),-(a5)
	dbra	d1,addloop\@
	bcc.s	nocarry\@
	move.b	#1,-(a5)	
	addq	#1,d7
nocarry\@
	cmp.b	#9,(\3)
	bgt.s	done
	ENDM

loop:	
	addfib a1,a2,a3
	addfib a2,a3,a1
	addfib a3,a1,a2
	bra	loop
	
done:	
	lea.l	3*NUMSTORAGE(sp),sp
	rts

224 bytes.
After you run it, type set in shell and see the answer in RC.

ross · 11 December 2018, 21:58

I've just finished a 44 bytes version (no bss space, only code and stack usage).
I publish it or wait a bit for someone different result? (I love this little competitions

)

alkis · 11 December 2018, 22:46

Quote:

Originally Posted by ross

I've just finished a 44 bytes version (no bss space, only code and stack usage).
I publish it or wait a bit for someone different result? (I love this little competitions

)

lol, 44 bytes?

Man, I knew I was noob. I didn't knew I was that much of a noob though

Well played, sir!

ross · 11 December 2018, 22:53

Quote:

Originally Posted by alkis

lol, 44 bytes?

Man, I knew I was noob. I didn't knew I was that much of a noob though

Well played, sir!

Uh sorry! I've not counted rts!
So actually 46

Just curious, i've not assembled your source.
What's your result?

Mine is 4782.

alkis · 11 December 2018, 22:58

Quote:

Originally Posted by ross

Uh sorry! I've not counted rts!
So actually 46

Just curious, i've not assembled your source.
What's your result?

Mine is 4782.

4782 indeed.

alkis · 11 December 2018, 23:26

Quote:

Originally Posted by ross

I've just finished a 44 bytes version (no bss space, only code and stack usage).

The no bss space got me wondering

Umm, what's the final executable size?

ross · 11 December 2018, 23:35

Quote:

Originally Posted by alkis

The no bss space got me wondering

Umm, what's the final executable size?

84: header(36)+code(46)+pad(2)

alkis · 12 December 2018, 00:16

code 94, final exe 132

Code:

	;; vasmm68k_mot -Fhunkexe -o euler25 euler25.asm
DIGITS = 1000
NUMSTORAGE = DIGITS/2
	opt x+
	
	lea.l	-NUMSTORAGE(sp),sp
	move.l	sp,a1
	lea.l	-NUMSTORAGE(sp),sp
	move.l	sp,a2
	lea.l	-NUMSTORAGE(sp),sp
	move.l	sp,a3


	;; init *a1 and *a2 to 1
	moveq	#1,d0
	move.b	d0,NUMSTORAGE-1(a1)
	move.b	d0,NUMSTORAGE-1(a2)
	move.b	d0,(a2)
	moveq	#2,d0	; d0 will hold the term-number we are now calculating
	moveq	#1-1,d7		;number of pair of digits, normalised for dbra

loop:
	addq	#1,d0
	
	lea.l	NUMSTORAGE(a2),a6
	lea.l	NUMSTORAGE(a1),a4
	lea.l	NUMSTORAGE(a3),a5

	move.w	d7,d1
	
digloop:
	move.b	-(a6),-1(a5)
	abcd	-(a4),-(a5)
	scc	d2
	move.b	(a6),(a4)
	move.b	(a5),(a6)
	dbra	d1,digloop
	tst.b	d2
	bne.s	nocarry
	move.b	#1,-(a6)
	move.b	#0,-(a4)
	addq	#1,d7
nocarry:
	cmp.b	#9,(a2)
	bgt.s	done
	bra.s	loop
done:	
	lea.l	3*NUMSTORAGE(sp),sp
	rts

ross · 12 December 2018, 00:37

Quote:

Originally Posted by alkis

code 94, final exe 132

Much better.

If you only look at the formula you see that at step n you need to sum n-1 to n-2.
But you have to actually think about what does it represent the Fibonacci sequence: a recurrence relation.
So if you use n-2 as the current n..

a/b · 12 December 2018, 13:02

At 46 right now, gonna keep trying. Out of curiosity, I have 18 instructions.
Awaiting meynaf's 44 ;P...

ross · 12 December 2018, 13:12

Quote:

Originally Posted by a/b

At 46 right now, gonna keep trying. Out of curiosity, I have 18 instructions.
Awaiting meynaf's 44 ;P...

There is a 44 bytes possibility but is forbidden with scheduler active

The 44 bytes code is 16 instructions.

My 46 bytes version is 18 instructions, like yours.
But probably there are more possibilities to do it in 46 bytes.

And yes as usual when meynaf will see my code i'm expecting a 2 bytes strip

ross · 12 December 2018, 13:21

Quote:

Originally Posted by a/b

Out of curiosity

I have a curiosity too.

Since I wanted to find a way to use the link instruction (I committed myself to using it at all costs!), did you accidentally use it too?

a/b · 12 December 2018, 13:51

Yeah, I'm using link.

meynaf · 12 December 2018, 14:14

Quote:

Originally Posted by a/b

Awaiting meynaf's 44 ;P...

44 bytes for what ?
Does it include displaying the number ?

Quote:

Originally Posted by ross

And yes as usual when meynaf will see my code i'm expecting a 2 bytes strip

Maybe, maybe not. I have bigger things on the fire currently.

ross · 12 December 2018, 15:29

Quote:

Originally Posted by meynaf

44 bytes for what ?
Does it include displaying the number ?

Nope, a simple return code displayed on CLI (so a value in d0.l at exit).

Quote:

Maybe, maybe not. I have bigger things on the fire currently.

Yep, don't distract yourself

meynaf · 12 December 2018, 15:47

Quote:

Originally Posted by ross

Nope, a simple return code displayed on CLI (so a value in d0.l at exit).

Ok then, let's go.

Code:

 move.l #4782,d0
 rts

(What ? It gives the right result, no ?)

Similar Threads
Thread	Thread Starter	Forum	Replies	Last Post
Any software to see technical OS details?	necronom	support.Other	3	02 April 2016 12:05
2-star rarity details?	stet	HOL suggestions and feedback	0	14 December 2015 05:24
EAB's FTP details...	Basquemactee1	project.Amiga File Server	2	30 October 2013 22:54
req details for sdl	turrican3	request.Other	0	20 April 2008 22:06
Forum Details	BippyM	request.Other	0	15 May 2006 00:56

11 December 2018, 21:58	#946
ross Defendit numerus Join Date: Mar 2017 Location: Crossing the Rubicon Age: 53 Posts: 4,468	I've just finished a 44 bytes version (no bss space, only code and stack usage). I publish it or wait a bit for someone different result? (I love this little competitions )

12 December 2018, 13:02	#954
a/b Registered User Join Date: Jun 2016 Location: europe Posts: 1,039	At 46 right now, gonna keep trying. Out of curiosity, I have 18 instructions. Awaiting meynaf's 44 ;P...

12 December 2018, 13:51	#957
a/b Registered User Join Date: Jun 2016 Location: europe Posts: 1,039	Yeah, I'm using link.

Currently Active Users Viewing This Thread: 1 (0 members and 1 guests)