Welcome, Guest. Please login or register.

Author Topic: 68060 speedup patches + more  (Read 12091 times)

Description:

0 Members and 1 Guest are viewing this topic.

Offline utri007Topic starter

Re: 68060 speedup patches + more
« Reply #59 from previous page: January 15, 2016, 07:41:25 PM »
Quote from: Thomas Richter;802200
This is probably side-tracking this thread, but just for curiousity: What is exactly this interrupt problem? I do have a GVP spectrum, with P96, and that's working just fine. So I wonder what the problem is.


TCP/IP stack and USB stack makes WHDLoad games  unstable. Thats why there is commented out options to disable most TCP/IP stacks and USB on s:whdloadstartup file. Same with latest versions of  Picasso96 and some cards. You don't play?

http://whdload.de/docs/en/bugs.html

Using Picasso96 versions equal or greater than release 1.36 also the graphics card Spectrum creates such interrupts. To avoid this the Picasso96 software must be reverted to a pre 1.36 release or the gfxcard driver must be disabled.
ACube Sam 440ep Flex 800mhz, 1gb ram and 240gb hd and OS4.1FE
A1200 Micronic tower, OS3.9, Apollo 060 66mhz, xPert Merlin, Delfina Lite and Micronic Scandy, 500Gb hd, 66mb ram, DVD-burner and WLAN.
A1200 desktop, OS3.9, Blizzard 060 66mhz, 66mb ram, Ide Fix Express with 160Gb HD and WLAN
A500 OS2.1, GVP+HD8 with 4mb ram, 1mb chip ram and 4gb HD
Commodore CDTV KS3.1, 1mb chip, 4mb fast ram and IDE HD
 

Offline Oldsmobile_Mike

Re: 68060 speedup patches + more
« Reply #60 on: January 15, 2016, 07:50:47 PM »
I haven't had any problem with Poseidon running in the background when I start WHDLoad games.  Obviously USB mice and keyboards aren't supported by games, but I haven't had to shut anything down.

My Spectrum card, on the other hand, some games will just start with a black screen unless I switch to a non-P96 screenmode prior to starting them.  That /bugs link looks interesting, will have to research further.  Thanks!  :)
Amiga 500: 2MB Chip|16MB Fast|30MHz 68030+68882|3.9|Indivision ECS|GVP A500HD+|Mechware card reader + 8GB CF|Cocolino|SCSI DVD-RAM
Amiga 2000: 2MB Chip|136MB Fast|50MHz 68060|3.9|Indivision ECS + GVP Spectrum|Mechware card reader + 8GB CF|AD516|X-Surf 100|RapidRoad|Cocolino|SCSI CD-RW
 Amiga videos and other misc. stuff at https://www.youtube.com/CompTechMike/videos
 

Offline nicholas

Re: 68060 speedup patches + more
« Reply #61 on: January 15, 2016, 07:57:08 PM »
Quote from: utri007;802204
TCP/IP stack and USB stack makes WHDLoad games  unstable. Thats why there is commented out options to disable most TCP/IP stacks and USB on s:whdloadstartup file. Same with latest versions of  Picasso96 and some cards. You don't play?

http://whdload.de/docs/en/bugs.html

Using Picasso96 versions equal or greater than release 1.36 also the graphics card Spectrum creates such interrupts. To avoid this the Picasso96 software must be reverted to a pre 1.36 release or the gfxcard driver must be disabled.


Would this patch help perhaps?

http://aminet.net/package/util/boot/BK_IntAckFix
“Een rezhim-i eshghalgar-i Quds bayad az sahneh-i ruzgar mahv shaved.” - Imam Ayatollah Sayyed  Ruhollah Khomeini
 

Offline Acill

Re: 68060 speedup patches + more
« Reply #62 on: January 15, 2016, 08:17:31 PM »
Quote from: Thomas Richter;802202
So, did my above hint help?


Havent had a chance to try installing again with the installer this time :) I did get the linked MMUlib package this morning. I plan to try again when I get home from work. Stupid work getting in the way of my fun!!
Proud Retired Navy Chief!

A4000T - CSPPC - Mediator
Powerbook G4 15", 17"
Powermac G5 2GHZ
AmigaOne X5000
Need Amiga recap or other services in the US? Visit my website at http://www.acill.com and take a look or on facebook at http://facebook.com/acillclassics
 

Offline utri007Topic starter

Re: 68060 speedup patches + more
« Reply #63 on: January 15, 2016, 08:27:07 PM »
Quote from: Oldsmobile_Mike;802206
I haven't had any problem with Poseidon running in the background when I start WHDLoad games.  Obviously USB mice and keyboards aren't supported by games, but I haven't had to shut anything down.

My Spectrum card, on the other hand, some games will just start with a black screen unless I switch to a non-P96 screenmode prior to starting them.  That /bugs link looks interesting, will have to research further.  Thanks!  :)


It is possible to but required parts of poseidon to ROM and then mouse and keyboard would be supported.

I had Spectrum and I had that problem with Picasso96 and Whdload, but not with my xPert Merlin.

By the way I made some speed test and Merlin is quite much faster. Got some 2fps more to Quake with Merlin. That is interesting because it is possible to get xPert Merlin for very economy price. It has it problems, but most of them are cured with fixes. Karl Werner Riedel still supports card. Unsolvable problem is with 24bit screens, but compared to Spectrum it is OK have 16 bit workbench, as Spectrum can't do better and 16bit works just fine with Merlin.
« Last Edit: January 15, 2016, 08:36:56 PM by utri007 »
ACube Sam 440ep Flex 800mhz, 1gb ram and 240gb hd and OS4.1FE
A1200 Micronic tower, OS3.9, Apollo 060 66mhz, xPert Merlin, Delfina Lite and Micronic Scandy, 500Gb hd, 66mb ram, DVD-burner and WLAN.
A1200 desktop, OS3.9, Blizzard 060 66mhz, 66mb ram, Ide Fix Express with 160Gb HD and WLAN
A500 OS2.1, GVP+HD8 with 4mb ram, 1mb chip ram and 4gb HD
Commodore CDTV KS3.1, 1mb chip, 4mb fast ram and IDE HD
 

guest11527

  • Guest
Re: 68060 speedup patches + more
« Reply #64 on: January 15, 2016, 09:09:10 PM »
Quote from: Oldsmobile_Mike;802206
My Spectrum card, on the other hand, some games will just start with a black screen unless I switch to a non-P96 screenmode prior to starting them.  That /bugs link looks interesting, will have to research further.

If that's a bug, it's the bug of the game(s) not using the Os.

Let me explain: The GVP spectrum has a video switch on its board. This switch can either switch the output from the Cirrus chip on the GVP board to the VGA output, or the VGA input comming from the pass-through cable from the chipset. The video switch is driven by a register on the GVP board, and the P96 triggers this switch whenever intuition switches between a native and a GVP (Cirrus) screen.

Now, if a program bypasses the Os and writes into the Custom chip addresses directly without using Intution (or Graphics, to be precise), then the video output switch remains in the position that redirects the Cirrus output to the VGA out.

So it's not at all related to P96 (which is not used by these games) nor Graphics (which is neither used by the games) nor by the hardware (the video switch is very much necessary for a video pass-through, and how else than by software should it be triggered?).

It is the fault of lazy programmers not using graphics and intuition - where P96 would be able to redirect video output properly.
 

guest11527

  • Guest
Re: 68060 speedup patches + more
« Reply #65 on: January 15, 2016, 09:35:36 PM »
Quote from: utri007;802204
TCP/IP stack and USB stack makes WHDLoad games  unstable.
While I don't know the details, it doesn't surprise me. A "WHDLoad" game is something that bypasses the Os, and that expects pretty much a bare-bone system without anything unexpected in it. Now you add hardware on top that requires regular servicing through the Os, or rather, their corresponding device drivers. A "WHDLoad game" will usually disable all the Os services, including interrupts. No surprise the hardware runs amok.

In other words: Bad programming. Such games program a complex system (the Amiga) as if the hardware is a C64, in the tradition of the 8-bit systems where you just went directly to the hardware. (Un)fortunately, the Amiga system is an expansible system to which hardware resources can be added, but it has an Os that can be told to service these devices. If you don't use the Os... well, bad luck.

Quote from: utri007;802204
Thats why there is commented out options to disable most TCP/IP stacks and USB on s:whdloadstartup file. Same with latest versions of  Picasso96 and some cards.
Not surprising, and not at all the fault of the TCP/IP stack, nor Picasso96, nor the USB stack. "WHDLoad" tries to shut down the system safely, but you cannot simply expect to shut down some parts of the system while keeping others working. The Os keeps the whole thing working together - you shut it down - the system falls apart.


Quote from: utri007;802204
You don't play?
My games on the amiga are called "DevPac" and "SAS/C". Oh, you mean, games? Well, not anymore these days. Back then, I was (and still am) into adventures. Never had problems with extensions like P96. If I wanted to play a game: Reboot the system, boot directly from the disk.

What you're saying is not surprising: It says mainly "WHLoad does not work". Certainly not - it cannot. For making WHLoad to work, it would need to know all possible extensions that could be connected to the Amiga, and how to shut them down or restore them safely. If this is possible at all.
WHDLoad is a hack. It works "mostly", but of course not if it hits something unexpected and something that has not been built into WHDLoad.

Quote from: utri007;802204
Using Picasso96 versions equal or greater than release 1.36 also the graphics card Spectrum creates such interrupts. To avoid this the Picasso96 software must be reverted to a pre 1.36 release or the gfxcard driver must be disabled.
P96 has no interface to disable board interrupts. Mostly because the Os does not require such a feature. It is not P96 which "breaks the interface". You rather expect that WHDload should reconfigure the graphics card bypassing P96. Since there are many graphics cards, this type of solution does not scale - essentially WHDLoad would need to know all available graphics cards. And all TCP/IP stacks. And all USB stacks.

I hope this explains a bit what is so terribly wrong about this program. You cannot hold together the system while shutting only parts down, and you cannot scale it up to the list of possible hardware extensions because every possible extension requires manual care in WHDLoad. You can only boot up the system without ever having initialized the extensions so they cannot interfere with such ancient software.
 

Offline klx300r

  • Amiga 1000+AmigaOne X1000
  • Hero Member
  • *****
  • Join Date: Sep 2007
  • Posts: 3261
  • Country: ca
  • Thanked: 20 times
  • Gender: Male
    • Show only replies by klx300r
    • http://mancave-ramblings.blogspot.ca/
Re: 68060 speedup patches + more
« Reply #66 on: January 15, 2016, 10:07:53 PM »
Quote from: nicholas;802207
Would this patch help perhaps?

http://aminet.net/package/util/boot/BK_IntAckFix

thanks for the link! interested to test out if my 060 has the interrupt issue Harry talks about
____________________________________________________________________
c64-dual sids, A1000, A1200-060@50, A4000-CSMKIII
Indivision AGA & Catweasel MK4+= Amazing
! My Master Miggies-Amiga 1000 & AmigaOne X1000 !
--- www.mancave-ramblings.blogspot.ca ---
  -AspireOS.com & Amikit- Amiga for your netbook-
***X1000- I BELIEVE *** :angel:
 

Offline Oldsmobile_Mike

Re: 68060 speedup patches + more
« Reply #67 on: January 16, 2016, 03:24:31 AM »
Quote from: Thomas Richter;802212
I hope this explains a bit what is so terribly wrong about this program.

Hey now, it does work perfectly a majority of the time.  The number of games that have been successfully updated is a testament to the WHDLoad team.  And I'm sure all the programmers of those horrible, hardware-banging, non-system-friendly games certainly didn't foresee a bunch of crazy fanatics still wanting to play them, 30 years later.  :laughing:
Amiga 500: 2MB Chip|16MB Fast|30MHz 68030+68882|3.9|Indivision ECS|GVP A500HD+|Mechware card reader + 8GB CF|Cocolino|SCSI DVD-RAM
Amiga 2000: 2MB Chip|136MB Fast|50MHz 68060|3.9|Indivision ECS + GVP Spectrum|Mechware card reader + 8GB CF|AD516|X-Surf 100|RapidRoad|Cocolino|SCSI CD-RW
 Amiga videos and other misc. stuff at https://www.youtube.com/CompTechMike/videos
 

Offline Oldsmobile_Mike

Re: 68060 speedup patches + more
« Reply #68 on: January 16, 2016, 03:26:04 AM »
Quote from: Thomas Richter;802211
Let me explain: The GVP spectrum has a video switch on its board. This switch can either switch the output from the Cirrus chip on the GVP board to the VGA output, or the VGA input comming from the pass-through cable from the chipset. The video switch is driven by a register on the GVP board, and the P96 triggers this switch whenever intuition switches between a native and a GVP (Cirrus) screen.

Now, if a program bypasses the Os and writes into the Custom chip addresses directly without using Intution (or Graphics, to be precise), then the video output switch remains in the position that redirects the Cirrus output to the VGA out.

So it's not at all related to P96 (which is not used by these games) nor Graphics (which is neither used by the games) nor by the hardware (the video switch is very much necessary for a video pass-through, and how else than by software should it be triggered?).

It is the fault of lazy programmers not using graphics and intuition - where P96 would be able to redirect video output properly.

Perfect explanation!  Thanks!  :pint:
Amiga 500: 2MB Chip|16MB Fast|30MHz 68030+68882|3.9|Indivision ECS|GVP A500HD+|Mechware card reader + 8GB CF|Cocolino|SCSI DVD-RAM
Amiga 2000: 2MB Chip|136MB Fast|50MHz 68060|3.9|Indivision ECS + GVP Spectrum|Mechware card reader + 8GB CF|AD516|X-Surf 100|RapidRoad|Cocolino|SCSI CD-RW
 Amiga videos and other misc. stuff at https://www.youtube.com/CompTechMike/videos
 

Offline Acill

Re: 68060 speedup patches + more
« Reply #69 on: January 16, 2016, 03:32:14 AM »
Well using the installer for MMULib was a disaster! I am glad I backup my boot partition before I make any big install like this! After going through all the stuff using expert mode it completly killed my system. I realy dont think it like being installed to a machine with a Mediator and the RTG system it uses. I ended up having to format my boot partition and restore it back to pre MMULib.
Proud Retired Navy Chief!

A4000T - CSPPC - Mediator
Powerbook G4 15", 17"
Powermac G5 2GHZ
AmigaOne X5000
Need Amiga recap or other services in the US? Visit my website at http://www.acill.com and take a look or on facebook at http://facebook.com/acillclassics
 

guest11527

  • Guest
Re: 68060 speedup patches + more
« Reply #70 on: January 16, 2016, 06:30:08 AM »
Quote from: Acill;802236
Well using the installer for MMULib was a disaster! I am glad I backup my boot partition before I make any big install like this! After going through all the stuff using expert mode it completly killed my system. I realy dont think it like being installed to a machine with a Mediator and the RTG system it uses. I ended up having to format my boot partition and restore it back to pre MMULib.

The installer keeps a backup of all your files... No need to format anything to restore, you just need to copy them back to place.
 

Offline stefcep2

  • Hero Member
  • *****
  • Join Date: Sep 2007
  • Posts: 1467
    • Show only replies by stefcep2
Re: 68060 speedup patches + more
« Reply #71 on: January 16, 2016, 06:51:56 AM »
@thomas. My memory of history is that Amiga Inc *wanted* to release a 1 MB ROM with 3.5 but it was too costly for them to have the chips manufactured, not that it was technically difficult.  Hence the "into fast ram instead" solution users got stuck with.

Personally what i loved about Amiga was the fast boot.  OS 3.5 took that feature away with the reset on cold boot.  This reset only existed because Amiga Inc didn't provide the physical 1 MB ROMS.  

I see nothing immoral if someone goes the expense, time and effort of making a 1 MB ROM.  My argument would be that it was what the owners of 3.5 and 3.9 should have done anyway.

The fact people will go to the trouble of doing all this tells us they would gladly pay for an official 1 MB.  If the copyright owner won't shut up and take their money, who is to blame.
 

guest11527

  • Guest
Re: 68060 speedup patches + more
« Reply #72 on: January 16, 2016, 11:03:01 AM »
(Double post, removed)
 

guest11527

  • Guest
Re: 68060 speedup patches + more
« Reply #73 on: January 16, 2016, 11:03:43 AM »
Quote from: stefcep2;802245
@thomas. My memory of history is that Amiga Inc *wanted* to release a 1 MB ROM with 3.5 but it was too costly for them to have the chips manufactured, not that it was technically difficult.  Hence the "into fast ram instead" solution users got stuck with.
Well, that is not entirely correct. The problem is that it is not ony too costy, it is also too risky. Is the software really stable enough to be put in ROM, once and forever? Once it sits there, it is very hard to fix bugs. Looking at the history of 3.5 and also 3.9, it showed that it is not. It took four BoingBags to get the entire system approximately stable, and I'm myself still updating the shell from time to time.

A software large enough contains at least one bug...

So, the conclusion is that this ROM-image of the kickstart is really a nightmare in terms of flexibility. You cannot update. There are reasons by PCs run with a Bios - and UEFI is approximately as bad as an idea as the Kickstart was, even though it is much less complex. Look at all the UEFI bugs that were discovered in the last years, including some bugs that could brick your entire system...

So, ROMs are really bad solutions, and if they exist, they should only include a minimal image that is barely enough to boot up the system, and get everything else done by software that is easily replaced and upgraded.

Quote from: stefcep2;802245
Personally what i loved about Amiga was the fast boot.  OS 3.5 took that feature away with the reset on cold boot.  This reset only existed because Amiga Inc didn't provide the physical 1 MB ROMS.  

I see nothing immoral if someone goes the expense, time and effort of making a 1 MB ROM.  My argument would be that it was what the owners of 3.5 and 3.9 should have done anyway.
I don't agree with this argument. Actually, what should *really* have been done is to replace all the ROM cruft with a minimal boot rom, just large enough to support the hard disk firmware, and to initiate a soft-boot from the RDB which could include the necessary ROM modules. This would have avoided the second reset itself (because it would not have been necessary to replace anything that is not in the ROM itself), and it would  have only taken approximately two additional seconds - or something in this magnitude.

How long does it take to load one MB from the harddisk? I have not measured, but I would guess about this time.

With any ROM, it would have been impossible or very expensive to provide boing bags.

Concerning the "morality" of the story: The question is not whether it is "morally OK", the question is whether it is covered by the license you got from Amiga(?) whomever(?) with the ROM. While I'm not a lawyer, I would typically expect that it disallows disecting the entire work and reproduce it.

Quote from: stefcep2;802245
The fact people will go to the trouble of doing all this tells us they would gladly pay for an official 1 MB.  If the copyright owner won't shut up and take their money, who is to blame.

I actually come to quite the opposide conclusion. People in Amiga land want everything for free, and moan if it is not for free. In specific, a ROM solution would have meant that the boing bags would not have been for free. And I can clearly imagine the %&$#?@!%&$#?@!%&$#?@!%&$#?@!-storm that would have breaken loose if people would have to pay for an upgrade ROM for correcting other's people's faults. Including my own errors.

I like to roll out new releases to users *that have already paid* with no additional costs implied. It was my fault, and I correct it for you, and you don't have to pay to get it fixed because you already paid. Having had a ROM would have made this impossible (or at least, very hard).
 

Offline Acill

Re: 68060 speedup patches + more
« Reply #74 on: January 16, 2016, 01:35:46 PM »
It did make the old drawer, but the problem was with the assigns it made for all the tools it added. When I rebooted it didnt like my picasso96 screens any more, couldnt find ENV and asked for it, so restoring my files was  pain in the butt. It was easier to just format and reboot with my backup drive. I plan to try again and see if I selected something in the install I shouldnt have.
Proud Retired Navy Chief!

A4000T - CSPPC - Mediator
Powerbook G4 15", 17"
Powermac G5 2GHZ
AmigaOne X5000
Need Amiga recap or other services in the US? Visit my website at http://www.acill.com and take a look or on facebook at http://facebook.com/acillclassics