Large Screen BOB problem

BooBoo · 02 May, 2011, 02:51 PM

With a Large Screen Screen 1008x1023 positioning a BOB around Y521 and beyond the BOB becomes distorted
The two images below -The First image shows the BOB at Y520 and the Second at 521

Sprites seem ok - What causes this and is there anyway around it?

BooBoo · 05 May, 2011, 04:37 PM

PowerBobs seems to fix this - But I dont realy like using the PowerBobs version of AMAL as it seems very bugged.

Anyone know of a contact for Manuel Andre?

Hungry Horace · 07 May, 2011, 07:13 PM

Are you using PowerBobs with the reg screen, out of interest?

BooBoo · 08 May, 2011, 10:08 AM

Yep just the Demo version with nag screen

Lonewolf10 · 05 Aug, 2011, 07:13 PM

That's a strange one. I have never seen that before, but I haven't worked with extra large screens in AMOS.

Did you get it fixed in the end?

Regards,
Lonewolf10

judeEZT · 13 Jan, 2015, 09:10 PM

Damn! I'm having the exact same problem.

I'm coding a 8-way scroller and need a large screen, but bitplanes seems get disjointed.

Anyone nows the best way to solve this?

bruceuncle · 16 Jan, 2015, 09:55 PM

If it's working okay in the PowerBobs extension then it looks like an AMOS bug. So I've added it to the bugs register.
Anyone know of a likely cause?

Sent from my Lumia 800 using Tapatalk

SamuraiCrow · 20 Jan, 2015, 07:15 AM

Hmmm... Maybe it's a hardware glitch. Sometimes there would be a problem with flaws in the design of the chipset and it would get worked around in the Kickstart routines but not in the hardware-banging code of the day. That's one reason that using the Kickstart versions of the routines might be preferred.

bruceuncle · 24 Jan, 2015, 02:30 PM

Thanks SamuraiCrow, but I think this one's more likely an AMOS bug as BooBoo mentions that PowerBobs appears to fix it. For now, it's just added to the bugs list. I haven't given it a high priority so I'll find out when I get there, so to speak...

SamuraiCrow · 24 Jan, 2015, 03:12 PM

I'm not sure. Remember that PowerBobs uses the CPU to blit instead of the blitter.

bruceuncle · 25 Jan, 2015, 05:43 AM

Quote from: SamuraiCrow on 24 Jan, 2015, 03:12 PM
I'm not sure. Remember that PowerBobs uses the CPU to blit instead of the blitter.

Ah, now I see what you're getting at SamuraiCrow. Sorry for being a bit dense sometimes

.

Do you think an emulator (I use WINUAE) would be accurate enough to test that sort of problem?

SamuraiCrow · 25 Jan, 2015, 03:42 PM

If you use cycle exact timing for an A500 it should be accurate enough. Also disable the JIT on the CPU for the most accurate tests.

Hungry Horace · 29 Jan, 2015, 08:35 AM

on this topic, did anyone ever manage to reverse engineer PowerBobs to get rid of the nag screen and/or potentially include it as a 'bob' command alternative in new versions of AMOS?

SamuraiCrow · 29 Jan, 2015, 03:43 PM

Sorry, no. But I have been commissioned by the Apollo team to write software blitter routines for an FPGA accelerator so using the Kickstart routines would be wiser.

Lonewolf10 · 14 Feb, 2015, 07:54 PM

I haven't tried doing anything like that with any extensions.

I haven't used it for ages, but at a guess the 'nag screen' will likely be part of the Cold start "C_Lib -> L0" routine. A well placed jmp instruction should enable it to be skipped without causing the rest of it to fall over.
Which version of Powerbobs are you using? If I have it I can give it a try, but no promises...