Click Here
home features news forums classifieds faqs links search
6076 members 
Amiga Q&A /  Free for All /  Emulation /  Gaming / (Latest Posts)
Login

Nickname

Password

Lost Password?

Don't have an account yet?
Register now!

Support Amigaworld.net
Your support is needed and is appreciated as Amigaworld.net is primarily dependent upon the support of its users.
Donate

Menu
Main sections
» Home
» Features
» News
» Forums
» Classifieds
» Links
» Downloads
Extras
» OS4 Zone
» IRC Network
» AmigaWorld Radio
» Newsfeed
» Top Members
» Amiga Dealers
Information
» About Us
» FAQs
» Advertise
» Polls
» Terms of Service
» Search

IRC Channel
Server: irc.amigaworld.net
Ports: 1024,5555, 6665-6669
SSL port: 6697
Channel: #Amigaworld
Channel Policy and Guidelines

Who's Online
75 crawler(s) on-line.
 13 guest(s) on-line.
 0 member(s) on-line.



You are an anonymous user.
Register Now!
 dreamlandfantasy:  5 mins ago
 DiscreetFX:  15 mins ago
 Rob:  20 mins ago
 JKD:  25 mins ago
 BigD:  39 mins ago
 jacadcaps:  44 mins ago
 Karlos:  1 hr 12 mins ago
 miggymac:  1 hr 15 mins ago
 Comi:  2 hrs 1 min ago
 zipper:  2 hrs 3 mins ago

/  Forum Index
   /  General Technology (No Console Threads)
      /  Is 68k a proprietary instruction set?
Register To Post

Goto page ( Previous Page 1 | 2 )
PosterThread
Karlos 
Re: Is 68k a proprietary instruction set?
Posted on 6-May-2022 22:43:42
#21 ]
Elite Member
Joined: 24-Aug-2003
Posts: 2478
From: As-sassin-aaate! As-sassin-aaate! Ooh! We forgot the ammunition!

@Jose

I don't know why anyone considering themselves an Amiga enthusiast would "hate" the classics. The classic was everything that subsequent incarnations have not been: revolutionary.

_________________
Doing stupid things for fun...

 Status: Offline
Profile     Report this post  
Hypex 
Re: Is 68k a proprietary instruction set?
Posted on 7-May-2022 16:56:25
#22 ]
Elite Member
Joined: 6-May-2007
Posts: 10571
From: Greensborough, Australia

@agami

Quote:
The company producing the fictional 68080 ASIC could then potentially refer to the copyright act in the specific jurisdiction and see what degree of redesign of an original work renders said copyright claim null and void, and may prepare their own brief back to NXP elaborating their position.


It would obviously be a derivative work. And the 080 doesn't exactly follow the standard 68K model with MMU and extra features. An 040 based ASIC would a better model as it's just about the best of the 68K series without chopping features back and a standard 68K core. But, this idea is nothing new. Plenty of Chinese companies create cheaper copies of other designs in industry as part of their business model.

Quote:
If for some reason these two organizations can't come to some sort of workable agreement, one of them may turn to the courts to pursue their desired outcome.


It almost seems worth the risk but producing it would require some substantial outlay.

 Status: Offline
Profile     Report this post  
Hypex 
Re: Is 68k a proprietary instruction set?
Posted on 7-May-2022 17:09:47
#23 ]
Elite Member
Joined: 6-May-2007
Posts: 10571
From: Greensborough, Australia

@matthey

Quote:
The InnovASIC Fido ASIC using the 68k CPU32 ISA was produced and is still available on the market.


Just a few days ago a friend interested in producing new Amiga hardware told me about this Fido. A new 68K available at 66Mhz for a reasonable price? Wolf wolf!

It has some interesting features. Endian reversing on the fly with top bit set. Ethernet. Context banks. I don't know how compatible it is to 68K pure code nor what ISA it is based on. The data bus looks to be only 16 bits so could as cripling as a Falcon. But can address 2GB of memory which is the max AmigaOS can do.

Of course, we need Ghz speeds now, but as a fast 66Mhz 68060 alternative it's a cost effective replacement.

Quote:
The 68060 may fail to operate correctly if cooled too far out of the operating range. Maybe he meant 100MHz. I have caught myself saying and typing gigabytes of memory when I meant megabytes of memory as it applies to the Amiga too. Technology and bloat has increased by several orders of magnitude.


I wasn't given many details. It was for some specific industrial application. And yes I was actually told in the 1,000 Mhz clock region.

Last edited by Hypex on 08-May-2022 at 02:41 PM.

 Status: Offline
Profile     Report this post  
matthey 
Re: Is 68k a proprietary instruction set?
Posted on 7-May-2022 22:21:27
#24 ]
Super Member
Joined: 14-Mar-2007
Posts: 1476
From: Kansas

Hypex Quote:

It would obviously be a derivative work. And the 080 doesn't exactly follow the standard 68K model with MMU and extra features. An 040 based ASIC would a better model as it's just about the best of the 68K series without chopping features back and a standard 68K core. But, this idea is nothing new. Plenty of Chinese companies create cheaper copies of other designs in industry as part of their business model.


The 68040 chopped the 6888x FPU in a big way and not in a good way. The 68040 FPU added FSop and FDop instructions which were good ideas that simplified IEEE single and double precision floating point consistency but at the same time removed FINT and FINTRZ instructions for rounding to integers. FINTRZ is used for C floating point rounding often and it was a major mistake to leave it out as demonstrated by the 68060 adding FINT/FINTRZ back. The FSop and FDop additions instead of Fop were good because they removed the need to change the FPCR for the precision of an operation which is expensive. FINT uses the FPCR settings for rounding and FINTRZ specifies the instruction rounding without changing the FPCR. The encoding is open to specify the other 3 rounding modes with FINT variations like FINTRZ which should be extremely cheap to implement even though the other rounding modes are not used as often.

The 68040 was more popular than the 68060 as it was used at the height of 68k popularity. It was not as popular on the Amiga as it was on the Mac due to CBM being cheap and it running so hot. The 68060 is a much better design even though there was more integer mode chopping which unfortunately included the 64 bit multiply instructions already used effectively by compilers. The 68060 FPU was an improvement as a few rarely used instructions were dropped while the common FINT/FINTRZ instructions were added back. The chopping back then was due to transistor/gate/area constraints like the Apollo core has in an affordable FPGA today. This is likely why the Apollo Core FPU is chopped compared to even the 68060 FPU. The area for all instructions is trivial today in an ASIC. x86-64 CPUs still have fully implemented FPUs in hardware even though it is deprecated and the SIMD unit has replaced it.

Hypex Quote:

Just a few days ago a friend interested in producing new Amiga hardware told me about this Fido. A new 68K available at 66Mhz for a reasonable price? Wolf wolf!

It has some interesting features. Endian reversing on the fly with top bit set. Ethernet. Context banks. I don't know how compatible it is to 68K pure code nor what ISA it is based on. The data bus looks to be only 16 bits so could as cripling as a Falcon. But can address 2GB of memory which is the max AmigaOS can do.

Of, we need Ghz speeds now, but as a fast 66Mhz 68060 alternative it's a cost effective replacement.


CPU32 lacks bit field instructions and the rarely used double memory indirect addressing modes of the 68020+. The 68k CPU32 ISA is much more complete than trying to use ColdFire for 68k and is significantly enhanced over the 68000 ISA. The Fido supervisor mode is not like the Motorola supervisor mode which is a bigger hurdle to using the Fido for an Amiga than the CPU32 ISA.

Hypex Quote:

I wasn't given many details. It was for some specific industrial application. And yes I was actually told in the 1,000 Mhz clock region.


Embedded processors are more likely to be underclocked for reliability than overclocked. 1GHz would be a huge overclock too. The later v5 ColdFire did not clock to 1GHz and the deeper superpiped v6 was not planned to either.

1994 68060 50-66MHz 500nm (420nm for Rev6)
2002 ColdFire v5 300-366MHz 130nm
2003 ColdFire v6 615-800MHz 100nm (estimated)

ColdFire v5, which closely resembles the superscalar 68060 design, was never sold as an off the shelf part and v6 was cancelled as far as I know. ColdFire v5 was fully synthesizable for flexibility so likely little effort was put into further optimizing the core for higher clock speeds. Reaching 1GHz was challenging in the '90s. For comparison, CPUs which first reached 1GHz were the AMD Athlon in mid 1999, Pentium III in late 1999 and the Alpha 21264C in 2001.

 Status: Offline
Profile     Report this post  
billt 
Re: Is 68k a proprietary instruction set?
Posted on 7-May-2022 23:59:55
#25 ]
Elite Member
Joined: 24-Oct-2003
Posts: 3199
From: Maryland, USA

@kolla

Quote:
CPU cores implemented on FPGAs are not in direct competition with their ASIC cousins, neither in performance nor physically. So they are “safe”.


I dont think it really works that way. If they wanted to sue, I don't believe fpga is a valid legal shield from that.

Last edited by billt on 08-May-2022 at 12:02 AM.

_________________
All glory to the Hypnotoad!

 Status: Offline
Profile     Report this post  
Goto page ( Previous Page 1 | 2 )

[ home ][ about us ][ privacy ] [ forums ][ classifieds ] [ links ][ news archive ] [ link to us ][ user account ]
Copyright (C) 2000 - 2019 Amigaworld.net.
Amigaworld.net was originally founded by David Doyle