05 January 2024, 15:03 | #61 | |
Moderator
Join Date: Dec 2010
Location: Wisconsin USA
Age: 60
Posts: 841
|
Quote:
Motorola documented the various errata conditions so system programmers could understand how to properly use Move16 (and other instructions) to avoid the crash. This means CPU library developers should not have played roulette either. There are no 060 based Turbo Boards which support Burst cycles to the Zorro2 bus so I'm afraid I can't recommend one which would make you a satisfied customer. Last edited by SpeedGeek; 05 January 2024 at 15:30. Reason: Typo correction |
|
05 January 2024, 15:06 | #62 |
Registered User
Join Date: Jan 2005
Location: UmeƄ
Age: 43
Posts: 924
|
Can burst to the Zorro2 bus with a 060 card be implemented safely?
Last edited by SpeedGeek; 05 January 2024 at 15:29. Reason: Typo correction |
05 January 2024, 15:20 | #63 | ||
Moderator
Join Date: Dec 2010
Location: Wisconsin USA
Age: 60
Posts: 841
|
Quote:
If and when I can find some time, I will try to make a safer version of CMQ&B040 which plays a little better with the rules. It won't be as fast and it won't be 100% safe either. Motorola and Commodore provided choices for Amiga users, but they did not provide perfection. Quote:
BTW, The Zorro2 bus is a 68000 protocol bus and the 68000 CPU absolutely does NOT support Burst. Last edited by SpeedGeek; 08 January 2024 at 16:29. |
||
07 January 2024, 16:52 | #64 |
Registered User
Join Date: Sep 2022
Location: Switzerland
Posts: 115
|
phew, the effect of Move16 is very drastic!
Did some own tests => original CopyMem does ~13.7 MB/s, my own copy mem ~14,5 MB/s - that's the same as a 030 patch delivers, but a 060 copymem patch reaches ~18,9 MB/s !! |
10 January 2024, 09:56 | #65 | ||
Registered User
Join Date: Mar 2018
Location: Hastings, New Zealand
Posts: 2,581
|
Quote:
Quote:
|
||
10 January 2024, 13:49 | #66 |
Registered User
Join Date: Jan 2019
Location: Germany
Posts: 3,233
|
The RKRMs predate the support of the 68040 in AmigaOs, so it is not a miracle that nothing is said about MOVE16 or 68040 (leave alone 68060) specifics. They are essentially stuck at the state of Os 2.1.
|
10 January 2024, 15:11 | #67 | |
Moderator
Join Date: Dec 2010
Location: Wisconsin USA
Age: 60
Posts: 841
|
There is actually more than 1 page, but you can start here:
http://amigadev.elowar.com/read/ADCD.../node001B.html http://amigadev.elowar.com/read/ADCD.../node033C.html http://amigadev.elowar.com/read/ADCD.../node033B.html Quote:
<sigh> |
|
10 January 2024, 15:37 | #68 |
Registered User
Join Date: Jan 2008
Location: Warsaw/Poland
Age: 55
Posts: 1,975
|
I think that You are wrong. I wrote with Michael Sinz a few emails, when i was alive. He was "Senior Amiga Systems Engineer/OS Designer for Commodore-Amiga". He worked for Commodore from Amiga kick 2.0 to 3.1. Later he worked for Scala.
http://www.sinz.org/Michael.Sinz/Amiga/index.html |
10 January 2024, 16:22 | #69 | |
Moderator
Join Date: Dec 2010
Location: Wisconsin USA
Age: 60
Posts: 841
|
Quote:
"Bryce Nesbitt managed the core OS subgroup and did an amazing job at making a very disciplined release." You can read more about this topic here: https://www.gregdonner.org/workbench...planation.html BTW, Dave Haynie was the "Senior Hardware Engineer" at Commodore too, and if you read his "Definitive Buster" article you will see that he needed permission from management just to do the Rev. 11. |
|
10 January 2024, 17:40 | #70 |
Registered User
Join Date: Jan 2008
Location: Warsaw/Poland
Age: 55
Posts: 1,975
|
Easy, if You have access to leaked Amiga OS 3.1 sources, then You can check logs for every Amiga OS module. You will be known which changes was done by every Amiga OS programmer.
|
10 January 2024, 18:21 | #71 | |
Moderator
Join Date: Dec 2010
Location: Wisconsin USA
Age: 60
Posts: 841
|
Quote:
Now, can we please get back on topic? |
|
10 January 2024, 21:15 | #72 |
Alien Bleed
Join Date: Aug 2022
Location: UK
Posts: 4,165
|
|
Currently Active Users Viewing This Thread: 1 (0 members and 1 guests) | |
Thread Tools | |
|
|