Differences between ASSEMBLY code assembled with Hatari and real machine

All 680x0 related coding posts in this section please.

Moderators: simonsunnyboy, Mug UK, Zorro 2, Moderator Team

Post Reply
Robbizz
Atari freak
Atari freak
Posts: 71
Joined: Wed Jul 01, 2020 7:08 pm

Differences between ASSEMBLY code assembled with Hatari and real machine

Post by Robbizz »

Generally, to preserve the hardware of my 30-year-old Falcon, I use an emulator (Hatari in this case) to program, and then, once the necessary tests are done, I do one last test on a real machine. In this case I am using the Atari Falcon LINEF instructions to be able to use a 68882 FPU, and in particular DEVPAC 3 with the following source code:

Code: Select all

		move.w	#1000,d0
		fmove.p	#0,fp5
loop
		fadd.p	#0.5,fp5
		dbf	d0,loop

		move.w	#1,-(sp)		* waits for a key
		trap	#1
		addq.l	#2,a7

		clr.w	-(sp)			* exits
		trap	#1
The code needs no explanation, I'm just looping to add 0.5 to the value in question. The weird thing is that on the emulator I actually see the value 0.5 (naturally expressed in 32-bit IEEE 754.) When it's time to add the value (see second sphoto), but when I do EXACTLY the same thing, with EXACTLY the same configuration on of a real Falcon, I get what is seen in the first photo. Anyone have any idea what's going on?

Also I would like to understand how to execute the next instruction (CTRL + Z with devpac 3 under debug), when I am in zoom mode in the registers, so as to see the registers related to the FPU as well. Currently I am forced to do CTRL + Z and then, to see the register of the FPU changed, I have to do an ALT + Z to zoom the register page and actually see that the register has changed; a somewhat boring debug ... Thanks in advance for the help.

51d689b2-e33f-442a-a08c-fe2224e23422.jpg
Schermata 2021-11-19 alle 23.58.24.png
You do not have the required permissions to view the files attached to this post.
User avatar
Cyprian
10 GOTO 10
10 GOTO 10
Posts: 2343
Joined: Fri Oct 04, 2002 11:23 am
Location: Warsaw, Poland

Re: Differences between ASSEMBLY code assembled with Hatari and real machine

Post by Cyprian »

What about using the Hatari's debugger?
https://hatari.tuxfamily.org/doc/debugger.html
Mega ST 1 / 7800 / Portfolio / Lynx II / Jaguar / TT030 / Mega STe / 800 XL / 1040 STe / Falcon030 / 65 XE / 520 STm / SM124 / SC1435
DDD HDD / AT Speed C16 / TF536 / SDrive / PAK68/3 / Lynx Multi Card / LDW Super 2000 / XCA12 / SkunkBoard / CosmosEx / SatanDisk / UltraSatan / USB Floppy Drive Emulator / Eiffel / SIO2PC / Crazy Dots / PAM Net
Hatari / Steem SSE / Aranym / Saint
http://260ste.atari.org
ThorstenOtto
Atari God
Atari God
Posts: 1737
Joined: Sun Aug 03, 2014 5:54 pm

Re: Differences between ASSEMBLY code assembled with Hatari and real machine

Post by ThorstenOtto »

Did you just transfer the assembled program to your Falcon and debug it there, or did you transfer the source code and assemble it again? I'm asking because i know that different assemblers handle floating constants differently. I'm not sure how DevPac handles them, PASM would interprete it like you used them in your source code, but for example the gnu assembler expects them to be written as hex constants. Instead of specifying the constants in the instruction, its maybe safer to do something like this:

Code: Select all

	move.w	#1000,d0
		fmove.x	zero,fp5
loop
		fadd.x	half,fp5
		dbf	d0,loop

		move.w	#1,-(sp)		* waits for a key
		trap	#1
		addq.l	#2,a7

		clr.w	-(sp)			* exits
		trap	#1

zero:        dc.l 0,0,0
half:         dc.l $3ffe0000,0,0

Also remember that fadd.p uses the packed decimal format.
User avatar
Eero Tamminen
Fuji Shaped Bastard
Fuji Shaped Bastard
Posts: 2459
Joined: Sun Jul 31, 2011 1:11 pm

Re: Differences between ASSEMBLY code assembled with Hatari and real machine

Post by Eero Tamminen »

Which Hatari FPU emulation mode you are using? Only fully SW implemented FPU emulation ("--fpu-softfloat") in Hatari Git version is 100% compatible functionality-wise.

(The faster emulation modes using host FPU instructions have some very minor differences from real 68881 / 68882 / 68040 / 68060. Those differences do not matter for 99% of the SW, but e.g. for couple of FIT's 060 demos they do.)

Another potential issue is your FPU code being affected by the machine state before your code. Are you sure you've initialized everything properly, and that this initialization itself is not dependent on some machine state that easily differs between machines (having the same HW)?
Moulinaie
Captain Atari
Captain Atari
Posts: 458
Joined: Wed Feb 01, 2012 9:34 pm

Re: Differences between ASSEMBLY code assembled with Hatari and real machine

Post by Moulinaie »

Robbizz wrote: Fri Nov 19, 2021 11:30 pm
loop
fadd.p #0.5,fp5
dbf d0,loop

Hi,

have a look at the memory window. On the first, the argument is 0000-0000-0000-0000-0000-0000. (so FPU value of zero)
and on the other, the argument is 4001-0005-0000-0000-0000-0000 (so FPU value of 0.5).
Sans titre 1.jpg
You didn't load the same binary.
Maybe the different conditions of compilation make a correct binary on one side and an incorrect on the other side.

Guillaume.
You do not have the required permissions to view the files attached to this post.
Robbizz
Atari freak
Atari freak
Posts: 71
Joined: Wed Jul 01, 2020 7:08 pm

Re: Differences between ASSEMBLY code assembled with Hatari and real machine

Post by Robbizz »

Moulinaie wrote: Sun Nov 21, 2021 4:31 pm ...
have a look at the memory window. On the first, the argument is 0000-0000-0000-0000-0000-0000. (so FPU value of zero)
and on the other, the argument is 4001-0005-0000-0000-0000-0000 (so FPU value of 0.5).
...
You didn't load the same binary.
Maybe the different conditions of compilation make a correct binary on one side and an incorrect on the other side.
Hi Guillaume, thanks for your interest. It is exactly the same source file on the same devpac 3 (I made a copy of both Devpac 3 and the source code and then transferred it to the real Falcon machine respecting the partition names as well). The only thing that can leave me in doubt is the configuration of the Falcon virtual machine on Hatari, but it is strange that on the virtual machine it works and on the real one it does not ... In the assembly phase on real Falcon I discovered this: that value is rounded, that is, if I write 0.5 it rounds it to 0 and if I write 0.6 it rounds it to 1.0 (while in Hatari the value corresponds to the decimal I write).
Last edited by Robbizz on Mon Nov 22, 2021 12:16 am, edited 1 time in total.
Robbizz
Atari freak
Atari freak
Posts: 71
Joined: Wed Jul 01, 2020 7:08 pm

Re: Differences between ASSEMBLY code assembled with Hatari and real machine

Post by Robbizz »

ThorstenOtto wrote: Sat Nov 20, 2021 7:24 am Did you just transfer the assembled program to your Falcon and debug it there, or did you transfer the source code and assemble it again? I'm asking because i know that different assemblers handle floating constants differently. I'm not sure how DevPac handles them, PASM would interprete it like you used them in your source code, but for example the gnu assembler expects them to be written as hex constants. Instead of specifying the constants in the instruction, its maybe safer to do something like this:
....
Thanks for your interest. Unfortunately even with your suggestion it doesn't work. In the assembly phase on real Falcon I discovered this: that value is rounded, that is, if I write 0.5 it rounds it to 0 and if I write 0.6 it rounds it to 1.0 (while in Hatari the value corresponds to the decimal I write).
Robbizz
Atari freak
Atari freak
Posts: 71
Joined: Wed Jul 01, 2020 7:08 pm

Re: Differences between ASSEMBLY code assembled with Hatari and real machine

Post by Robbizz »

Eero Tamminen wrote: Sun Nov 21, 2021 3:53 pm Which Hatari FPU emulation mode you are using? Only fully SW implemented FPU emulation ("--fpu-softfloat") in Hatari Git version is 100% compatible functionality-wise.

(The faster emulation modes using host FPU instructions have some very minor differences from real 68881 / 68882 / 68040 / 68060. Those differences do not matter for 99% of the SW, but e.g. for couple of FIT's 060 demos they do.)

Another potential issue is your FPU code being affected by the machine state before your code. Are you sure you've initialized everything properly, and that this initialization itself is not dependent on some machine state that easily differs between machines (having the same HW)?
I'm not very good with Hatari settings. I simply tried to get as close as possible to the real machine, that is: Atari Falcon 030, 16 Mhz, MC68882, 14 Mbyte. I have tried to modify the other parameters but without obtaining differences in the results. If you notice, in the program I correctly initialize the registers, therefore, if I had old values, they would be replaced with the new values that I am going to write. The problem is that the real falcon writes me rounded integer values, while the virtual falcon on Hatari writes me the values actually written in decimal. It seems that the real Falcon has no FPU, but it is not because it could not show me the relative registers.

I forgot. I have two Falcons, and on both of them I have this result.
ThorstenOtto
Atari God
Atari God
Posts: 1737
Joined: Sun Aug 03, 2014 5:54 pm

Re: Differences between ASSEMBLY code assembled with Hatari and real machine

Post by ThorstenOtto »

Robbizz wrote: Sun Nov 21, 2021 8:26 pm [Thanks for your interest. Unfortunately even with your suggestion it doesn't work. In the assembly phase on real Falcon I discovered this: that value is rounded, that is, if I write 0.5 it rounds it to 0 and if I write 0.6 it rounds it to 1.0 (while in Hatari the value corresponds to the decimal I write).
That rounding is really strange, especially when you are sure you are using the same devpac version with the same options. However, if you use my suggestion, you don't write "0.5" anywhere, but specify the already converted values somewhere in the source, so there is no trouble with rounding. Of course, doing the conversion yourself to find out the hex values that must be used to represent the float values is a bit troublesome.
Moulinaie
Captain Atari
Captain Atari
Posts: 458
Joined: Wed Feb 01, 2012 9:34 pm

Re: Differences between ASSEMBLY code assembled with Hatari and real machine

Post by Moulinaie »

Hi,
Does your Falcon have a FPU 68882 ???
I think that if you don't have the FPU, Devpac doesn't convert correctly the real numbers.
Try with a program like SYSINFO to read the COOKIE list and see what value exists for the _FPU cookie.

Guillaume.
Robbizz
Atari freak
Atari freak
Posts: 71
Joined: Wed Jul 01, 2020 7:08 pm

Re: Differences between ASSEMBLY code assembled with Hatari and real machine

Post by Robbizz »

Moulinaie wrote: Mon Nov 22, 2021 11:03 am Hi,
Does your Falcon have a FPU 68882 ???
I think that if you don't have the FPU, Devpac doesn't convert correctly the real numbers.
Try with a program like SYSINFO to read the COOKIE list and see what value exists for the _FPU cookie.

Guillaume.
Both of my Falcons have FPU 68882.
User avatar
Cyprian
10 GOTO 10
10 GOTO 10
Posts: 2343
Joined: Fri Oct 04, 2002 11:23 am
Location: Warsaw, Poland

Re: Differences between ASSEMBLY code assembled with Hatari and real machine

Post by Cyprian »

Robbizz wrote: Sun Nov 21, 2021 8:24 pm
Moulinaie wrote: Sun Nov 21, 2021 4:31 pm ...
have a look at the memory window. On the first, the argument is 0000-0000-0000-0000-0000-0000. (so FPU value of zero)
and on the other, the argument is 4001-0005-0000-0000-0000-0000 (so FPU value of 0.5).
...
You didn't load the same binary.
Maybe the different conditions of compilation make a correct binary on one side and an incorrect on the other side.
Hi Guillaume, thanks for your interest. It is exactly the same source file on the same devpac 3 (I made a copy of both Devpac 3 and the source code and then transferred it to the real Falcon machine respecting the partition names as well). The only thing that can leave me in doubt is the configuration of the Falcon virtual machine on Hatari, but it is strange that on the virtual machine it works and on the real one it does not ... In the assembly phase on real Falcon I discovered this: that value is rounded, that is, if I write 0.5 it rounds it to 0 and if I write 0.6 it rounds it to 1.0 (while in Hatari the value corresponds to the decimal I write).
as @Moulinaie pointed out, the MON (devpac debugger) shows that those binaries are not the same - "4001-0005" vs "0000-0000". Maybe they were assembled with a different Devpac's conditions/flags.
Would be possible to share both of them (the one compiled under the Falcon and the other one - under the Hatari)?
Mega ST 1 / 7800 / Portfolio / Lynx II / Jaguar / TT030 / Mega STe / 800 XL / 1040 STe / Falcon030 / 65 XE / 520 STm / SM124 / SC1435
DDD HDD / AT Speed C16 / TF536 / SDrive / PAK68/3 / Lynx Multi Card / LDW Super 2000 / XCA12 / SkunkBoard / CosmosEx / SatanDisk / UltraSatan / USB Floppy Drive Emulator / Eiffel / SIO2PC / Crazy Dots / PAM Net
Hatari / Steem SSE / Aranym / Saint
http://260ste.atari.org
Robbizz
Atari freak
Atari freak
Posts: 71
Joined: Wed Jul 01, 2020 7:08 pm

Re: Differences between ASSEMBLY code assembled with Hatari and real machine

Post by Robbizz »

Cyprian wrote: Mon Nov 22, 2021 5:39 pm as @Moulinaie pointed out, the MON (devpac debugger) shows that those binaries are not the same - "4001-0005" vs "0000-0000". Maybe they were assembled with a different Devpac's conditions/flags.
Would be possible to share both of them (the one compiled under the Falcon and the other one - under the Hatari)?
Hi Cyprian, here are the binaries: FLOAT1.PRG produced with Hatari, while FLOAT2.PRG produced with real Falcon. On both runs the same version of DEVPAC 3 and with the exact same configurations. I am also attaching the program because I have added a couple of things.
FLOAT.zip
You do not have the required permissions to view the files attached to this post.
User avatar
Eero Tamminen
Fuji Shaped Bastard
Fuji Shaped Bastard
Posts: 2459
Joined: Sun Jul 31, 2011 1:11 pm

Re: Differences between ASSEMBLY code assembled with Hatari and real machine

Post by Eero Tamminen »

Robbizz wrote: Sun Nov 21, 2021 10:14 pm
Eero Tamminen wrote: Sun Nov 21, 2021 3:53 pm Which Hatari FPU emulation mode you are using? Only fully SW implemented FPU emulation ("--fpu-softfloat") in Hatari Git version is 100% compatible functionality-wise.
I'm not very good with Hatari settings. I simply tried to get as close as possible to the real machine, that is: Atari Falcon 030, 16 Mhz, MC68882, 14 Mbyte. I have tried to modify the other parameters but without obtaining differences in the results.
Have you tried in Hatari both of these (on separate test runs):
* Enabling "Full software FPU emulation"
* Disabling FPU completely [1]
?

[1] it sounds that when you assemble your program in Devpac on the real Falcon, it might not find/detect the FPU and rounds values to integers as result.
User avatar
Cyprian
10 GOTO 10
10 GOTO 10
Posts: 2343
Joined: Fri Oct 04, 2002 11:23 am
Location: Warsaw, Poland

Re: Differences between ASSEMBLY code assembled with Hatari and real machine

Post by Cyprian »

I just compiled that with VASM "vasmm68k_mot.exe -nocase -maxerrors=20 -devpac -m68030 -m68881 -no-opt -nosym -Ftos -o " and with the Devpac (GEN030, "Options": checked 68881/2 and "Processor: 68030") under Hatari and:
- Devpac output is the same as FLOAT1.PRG
- VASM output is the same as FLOAT2.PRG

I could ask about that issue the VASM maintainer.

Anyone with the Falcon could check attached Devpac with Robbizz's FLOAT.zip?
DEVPAC31.030.zip
You do not have the required permissions to view the files attached to this post.
Mega ST 1 / 7800 / Portfolio / Lynx II / Jaguar / TT030 / Mega STe / 800 XL / 1040 STe / Falcon030 / 65 XE / 520 STm / SM124 / SC1435
DDD HDD / AT Speed C16 / TF536 / SDrive / PAK68/3 / Lynx Multi Card / LDW Super 2000 / XCA12 / SkunkBoard / CosmosEx / SatanDisk / UltraSatan / USB Floppy Drive Emulator / Eiffel / SIO2PC / Crazy Dots / PAM Net
Hatari / Steem SSE / Aranym / Saint
http://260ste.atari.org
Robbizz
Atari freak
Atari freak
Posts: 71
Joined: Wed Jul 01, 2020 7:08 pm

Re: Differences between ASSEMBLY code assembled with Hatari and real machine

Post by Robbizz »

Eero Tamminen wrote: Mon Nov 22, 2021 10:18 pm Have you tried in Hatari both of these (on separate test runs):
* Enabling "Full software FPU emulation"
* Disabling FPU completely [1]
?
[1] it sounds that when you assemble your program in Devpac on the real Falcon, it might not find/detect the FPU and rounds values to integers as result.
Enabling "Full FPU emulation software" I have no idea where it is. I use Hatari 2.3.1 with Mac (attached you will find my current working configuration of Falcon on Hatari).

I tried in Hatari to disable the FPU and when assembling the code I get a:
Error: maths co-processor required at line 5 ....

Of course on real Falcon, if the FPU doesn't work properly (or is missing), I would get the exact same message.
I have two Falcon that behave in the same way (that is, they round the value to me), with MC68882 installed and definitely working because with other programs that use them they work correctly.

I tried to put the value directly in IEEE 754 (I also tried in exponential form), but the situation does not change, the number is always rounded to the real Falcon ...
Schermata 2021-11-23 alle 00.19.24.png
You do not have the required permissions to view the files attached to this post.
Robbizz
Atari freak
Atari freak
Posts: 71
Joined: Wed Jul 01, 2020 7:08 pm

Re: Differences between ASSEMBLY code assembled with Hatari and real machine

Post by Robbizz »

Cyprian wrote: Mon Nov 22, 2021 10:42 pm I just compiled that with VASM "vasmm68k_mot.exe -nocase -maxerrors=20 -devpac -m68030 -m68881 -no-opt -nosym -Ftos -o " and with the Devpac (GEN030, "Options": checked 68881/2 and "Processor: 68030") under Hatari and:
- Devpac output is the same as FLOAT1.PRG
- VASM output is the same as FLOAT2.PRG

I could ask about that issue the VASM maintainer.

Anyone with the Falcon could check attached Devpac with Robbizz's FLOAT.zip?
DEVPAC31.030.zip
I tried to assemble with VASM on real Falcon using your parameters and I get as in FLOAT2.PRG.

In Hatari assembling with VASM I obviously get the same result as you: FLOAT2.PRG.
User avatar
Cyprian
10 GOTO 10
10 GOTO 10
Posts: 2343
Joined: Fri Oct 04, 2002 11:23 am
Location: Warsaw, Poland

Re: Differences between ASSEMBLY code assembled with Hatari and real machine

Post by Cyprian »

Frank Wille (VASM maintainer) just confirmed to me that "Packed Decimal floating point constants are not yet supported" (".p") in VASM: http://sun.hasenbraten.de/vasm/release/vasm_19.html. Therefore also FLOAT2.PRG is wrong.
Mega ST 1 / 7800 / Portfolio / Lynx II / Jaguar / TT030 / Mega STe / 800 XL / 1040 STe / Falcon030 / 65 XE / 520 STm / SM124 / SC1435
DDD HDD / AT Speed C16 / TF536 / SDrive / PAK68/3 / Lynx Multi Card / LDW Super 2000 / XCA12 / SkunkBoard / CosmosEx / SatanDisk / UltraSatan / USB Floppy Drive Emulator / Eiffel / SIO2PC / Crazy Dots / PAM Net
Hatari / Steem SSE / Aranym / Saint
http://260ste.atari.org
User avatar
Cyprian
10 GOTO 10
10 GOTO 10
Posts: 2343
Joined: Fri Oct 04, 2002 11:23 am
Location: Warsaw, Poland

Re: Differences between ASSEMBLY code assembled with Hatari and real machine

Post by Cyprian »

and the question is whether the opcode for "fadd.p #0.5,fp5" is correct in FLOAT1.PRG
Mega ST 1 / 7800 / Portfolio / Lynx II / Jaguar / TT030 / Mega STe / 800 XL / 1040 STe / Falcon030 / 65 XE / 520 STm / SM124 / SC1435
DDD HDD / AT Speed C16 / TF536 / SDrive / PAK68/3 / Lynx Multi Card / LDW Super 2000 / XCA12 / SkunkBoard / CosmosEx / SatanDisk / UltraSatan / USB Floppy Drive Emulator / Eiffel / SIO2PC / Crazy Dots / PAM Net
Hatari / Steem SSE / Aranym / Saint
http://260ste.atari.org
ThorstenOtto
Atari God
Atari God
Posts: 1737
Joined: Sun Aug 03, 2014 5:54 pm

Re: Differences between ASSEMBLY code assembled with Hatari and real machine

Post by ThorstenOtto »

The question is there a reason to use packed decimal format at all? It is a format that isn't supported on 040+ and has to be emulated.
User avatar
Eero Tamminen
Fuji Shaped Bastard
Fuji Shaped Bastard
Posts: 2459
Joined: Sun Jul 31, 2011 1:11 pm

Re: Differences between ASSEMBLY code assembled with Hatari and real machine

Post by Eero Tamminen »

Robbizz wrote: Mon Nov 22, 2021 11:37 pm Enabling "Full FPU emulation software" I have no idea where it is. I use Hatari 2.3.1 with Mac (attached you will find my current working configuration of Falcon on Hatari).
Apparently Mac UI does not have that option (Mac UI is not maintained by Hatari developers, but by outside contributors). Just use the SDL GUI, that should be accessible also from the Mac GUI. That option is in the SDL GUI CPU dialog, see:
https://hatari.tuxfamily.org/doc/manual ... CPU_Dialog
User avatar
Cyprian
10 GOTO 10
10 GOTO 10
Posts: 2343
Joined: Fri Oct 04, 2002 11:23 am
Location: Warsaw, Poland

Re: Differences between ASSEMBLY code assembled with Hatari and real machine

Post by Cyprian »

Cyprian wrote: Mon Nov 22, 2021 10:42 pm I just compiled that with VASM "vasmm68k_mot.exe -nocase -maxerrors=20 -devpac -m68030 -m68881 -no-opt -nosym -Ftos -o " and with the Devpac (GEN030, "Options": checked 68881/2 and "Processor: 68030") under Hatari and:
- Devpac output is the same as FLOAT1.PRG
- VASM output is the same as FLOAT2.PRG

I could ask about that issue the VASM maintainer.

Anyone with the Falcon could check attached Devpac with Robbizz's FLOAT.zip?
DEVPAC31.030.zip
thanks to Frank Wille, Vasm (daily build: http://sun.hasenbraten.de/vasm/index.php?view=source) is now fixed, and the output is the same as my Devpac's output and FLOAT1.PRG.
Mega ST 1 / 7800 / Portfolio / Lynx II / Jaguar / TT030 / Mega STe / 800 XL / 1040 STe / Falcon030 / 65 XE / 520 STm / SM124 / SC1435
DDD HDD / AT Speed C16 / TF536 / SDrive / PAK68/3 / Lynx Multi Card / LDW Super 2000 / XCA12 / SkunkBoard / CosmosEx / SatanDisk / UltraSatan / USB Floppy Drive Emulator / Eiffel / SIO2PC / Crazy Dots / PAM Net
Hatari / Steem SSE / Aranym / Saint
http://260ste.atari.org
Robbizz
Atari freak
Atari freak
Posts: 71
Joined: Wed Jul 01, 2020 7:08 pm

Re: Differences between ASSEMBLY code assembled with Hatari and real machine

Post by Robbizz »

Cyprian wrote: Thu Nov 25, 2021 10:09 pm thanks to Frank Wille, Vasm (daily build: http://sun.hasenbraten.de/vasm/index.php?view=source) is now fixed, and the output is the same as my Devpac's output and FLOAT1.PRG.
Great Cyprian, I discovered VASM thanks to you and it looks very nice. I've been using it since last night. I would like to bring a graphics library that I made last month in devpac 3 environment in VASM (I think there is no modification to do given the compatibility with TOS), to understand if there are optimizations differences between the two assemblers. I am a beginner in Assembly (I have been studying it, time permitting, for a few months), and I wanted to try my hand and put myself in the shoes of a developer, when he has to write a graphics library, having only the basic assembly instructions available. I wrote this library for the Atari ST / E and it simply allows me to define and set a pixel, with the desired color, on the screen starting from the bit planes of that resolution. Then I expanded and wrote some code to draw a line with the Bresenham Algorithm; first for simplicity I wrote everything in C and then I passed in Assembly. I've noticed that this way of writing Assembly code is much faster and saves me from getting lost in debugging. And finally I created a little demo of a color-changing line bouncing between the edges of the screen.
Anyway thanks to you Cyprian for your interest and of course to all the others who participated.
Post Reply

Return to “680x0”