Click Here
home features news forums classifieds faqs links search
6071 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
26 crawler(s) on-line.
 58 guest(s) on-line.
 0 member(s) on-line.



You are an anonymous user.
Register Now!
 edwardsjethro:  9 mins ago
 joeyunderwood:  10 mins ago
 Sikharubel:  13 mins ago
 Musashi5150:  36 mins ago
 MagicSN:  1 hr 4 mins ago
 Gunnar:  1 hr 11 mins ago
 pixie:  1 hr 22 mins ago
 kolla:  1 hr 43 mins ago
 retrofaza:  1 hr 45 mins ago
 VooDoo:  2 hrs 21 mins ago

/  Forum Index
   /  Amiga General Chat
      /  68k Developement
Register To Post

Goto page ( 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 | 13 | 14 | 15 | 16 | 17 | 18 | 19 | 20 | 21 | 22 | 23 | 24 | 25 | 26 Next Page )
PosterThread
g01df1sh 
68k Developement
Posted on 17-Jul-2018 9:17:55
#1 ]
Super Member
Joined: 16-Apr-2009
Posts: 1777
From: UK

Hi

I was just wondering why 68k Development stopped and Motorola chose PPC. If money was not a issue Could the 68k be advanced to the point that it is faster than PPC when using Asic. Vampire is proving that 68k still has room for advancement.

_________________
A1200 ACA1232 128MB Indivison MkIICr
Elbox empty Power Tower
RPi3 Emulating C64 ZX Atari PS BBC
Wii with Amiga emulation
Vampire v4 SA

 Status: Offline
Profile     Report this post  
rzookol 
Re: 68k Developement
Posted on 17-Jul-2018 12:08:51
#2 ]
Regular Member
Joined: 4-Oct-2005
Posts: 318
From: Poland, Lublin

@g01df1sh

Is it? My 18-years-old Blizzard PPC is still faster than Vampire.

 Status: Offline
Profile     Report this post  
g01df1sh 
Re: 68k Developement
Posted on 17-Jul-2018 12:23:20
#3 ]
Super Member
Joined: 16-Apr-2009
Posts: 1777
From: UK

@rzookol

Vampire is fpga not Asic yet

_________________
A1200 ACA1232 128MB Indivison MkIICr
Elbox empty Power Tower
RPi3 Emulating C64 ZX Atari PS BBC
Wii with Amiga emulation
Vampire v4 SA

 Status: Offline
Profile     Report this post  
wawa 
Re: 68k Developement
Posted on 17-Jul-2018 13:14:47
#4 ]
Elite Member
Joined: 21-Jan-2008
Posts: 6259
From: Unknown

@rzookol

besides your ppc is certainly clocked higher than any 68k including apollo core in an fpga. counting clock by clock ppcs might not get away as glorious in comparison.

 Status: Offline
Profile     Report this post  
kolla 
Re: 68k Developement
Posted on 17-Jul-2018 13:29:57
#5 ]
Elite Member
Joined: 20-Aug-2003
Posts: 2859
From: Trondheim, Norway

68k development did not stop, it moved to ColdFire. Motorola did 68k CISC and 88k RISC, then moved 68k to ColdFire and dropped 88k to join IBM and Apple for PowerPC.

_________________
B5D6A1D019D5D45BCC56F4782AC220D8B3E2A6CC

 Status: Offline
Profile     Report this post  
Hypex 
Re: 68k Developement
Posted on 17-Jul-2018 13:57:25
#6 ]
Elite Member
Joined: 6-May-2007
Posts: 11180
From: Greensborough, Australia

@g01df1sh

The reason is Apple and IBM. Motorola were working on a followup. The 88000, or 88K. A RISC CPU. It failed to get martket penetration on a major scale and when PowerPC came along it disappered. Read about it here.

https://en.wikipedia.org/wiki/Motorola_88000

We can note here that the 68K survives in the 68K ColdFire. Which is more like a cut down 68K since they would be less compatible than a 6502 is against a 6510. Another Motorola, the DragonBall, is also based on the 68K.

A touch of irony, the ColdFire is said to run up to 300Mhz, a speed the 68080 still cannot touch. In it's WIP FPGA state.

Regarding money, I suppose Apple sales just didn't count for enough, since the 68K was being used in Apple Macs and Intel had deeper market penetration. There was just so much more money going into Intel, the way I see it. Yes the Amiga contributed, but Commodore helped it to contribute less.

As to being advanced faster than PPC, a PPC what? PPC back then or PPC now? In any case sure I think 68K could have matched anything for PPC or x86. The 68K is like an 80x86. They are both CISC designs. The 68K was a better and more modern design than the x86 ever was, so it could certainly compete with it. We only have to look at how they hacked and refactored the old x86 design over the centuries to become the monster it is today. So, with that in mind, I certainly think the 68K could have been engineered to the same point wth mutl cores, 64-bit data and address bus, 64* registers, 512 bit vectors and Ghz speeds!

* By the time x86 was 64-bit it matched the 16 register count of the 68K, up from 8. So why not double what x86 is doing and then double again!

Last edited by Hypex on 17-Jul-2018 at 02:03 PM.

 Status: Offline
Profile     Report this post  
pavlor 
Re: 68k Developement
Posted on 17-Jul-2018 14:45:02
#7 ]
Elite Member
Joined: 10-Jul-2005
Posts: 9578
From: Unknown

@g01df1sh

68k was no longer competitive in the early 90s.

In 1993, Intel had 486DX/2 66 MHz and Pentium... Motorola 68040 33 MHz. And no, there is no magic that will make 68040 faster than these Intel CPUs, far from that.

 Status: Offline
Profile     Report this post  
ErikBauer 
Re: 68k Developement
Posted on 20-Jul-2018 14:51:09
#8 ]
Super Member
Joined: 25-Feb-2004
Posts: 1141
From: Italy

@pavlor

In 1993 68040 could run at 40Mhz, with a comparable speed to the 486/66Mhz, too bad it was way more expensive.

By 1994 68060 came out, a brilliant competitor to the Intel Pentium, but it was too late and could not be clocked at high enough clockrate to compete with it's Intel counterpart.

No, magic could not have saved 68K family... a better engineering that could allow a better performance/cost ratio maybe could have, maybe not: Microsoft and Intel were already growing titans by then.




_________________
God created Paula so that Allister Brimble and Dave Whittaker could do music

Check my Amiga gameplays (ITA)!

 Status: Offline
Profile     Report this post  
pavlor 
Re: 68k Developement
Posted on 20-Jul-2018 15:37:37
#9 ]
Elite Member
Joined: 10-Jul-2005
Posts: 9578
From: Unknown

@ErikBauer

Quote:
In 1993 68040 could run at 40Mhz, with a comparable speed to the 486/66Mhz, too bad it was way more expensive.


Comparable? Somewhat yes. 68040 will be as fast in FPU operations, faster in bus related work, but most integer code will run faster on 486DX/2.

 Status: Offline
Profile     Report this post  
ErikBauer 
Re: 68k Developement
Posted on 20-Jul-2018 15:53:00
#10 ]
Super Member
Joined: 25-Feb-2004
Posts: 1141
From: Italy

@pavlor

AFAIK 68040 was far more efficient than the i486 reguarding Integer Operations, making that comparison a good toe to toe situation.

But the real problems of 68k line were overheating and production costs

_________________
God created Paula so that Allister Brimble and Dave Whittaker could do music

Check my Amiga gameplays (ITA)!

 Status: Offline
Profile     Report this post  
pavlor 
Re: 68k Developement
Posted on 20-Jul-2018 16:24:29
#11 ]
Elite Member
Joined: 10-Jul-2005
Posts: 9578
From: Unknown

@ErikBauer

Quote:
AFAIK 68040 was far more efficient than the i486 reguarding Integer Operations, making that comparison a good toe to toe situation.


Far more? Few % maybe, but certainly not much more. However, benchmarking different CPU architectures was not that easy back then - too few application benchmarks across the platforms and no "synthetic" benchmarks one can really believe (well there are at least Spec CPU92 numbers for both 68040 and 80486DX/2).

 Status: Offline
Profile     Report this post  
Korni 
Re: 68k Developement
Posted on 20-Jul-2018 16:35:41
#12 ]
Member
Joined: 9-Jan-2007
Posts: 97
From: Poland

I have a machine with 486 DX/4 100. Amiga with 040/40 is way more usable and fun to use, even with no magic involved.

 Status: Offline
Profile     Report this post  
JimIgou 
Re: 68k Developement
Posted on 20-Jul-2018 17:44:01
#13 ]
Regular Member
Joined: 30-May-2018
Posts: 114
From: Unknown

@Korni

Agreed. I didn't see the point in adopting X86 until it neared 200 MHz.
And even then I started with Cyrix and then moved to AMD.

Intel...still the enemy.

Last edited by JimIgou on 20-Jul-2018 at 05:45 PM.

 Status: Offline
Profile     Report this post  
matthey 
Re: 68k Developement
Posted on 20-Jul-2018 18:49:24
#14 ]
Super Member
Joined: 14-Mar-2007
Posts: 1968
From: Kansas

Quote:

Hypex wrote:
@g01df1sh

The reason is Apple and IBM. Motorola were working on a followup. The 88000, or 88K. A RISC CPU. It failed to get market penetration on a major scale and when PowerPC came along it disappered. Read about it here.

https://en.wikipedia.org/wiki/Motorola_88000


My understanding from Mitch Alsup (one of the 88k chip architects who is active on comp.arch) is that the decision to drop the 88k and go with PPC was mostly politics. There were many competitive and similar RISC architectures to choose from at the time and PPC was an attempt to standardize on a single RISC architecture (probably pushed by computer manufacturers like Apple to have multiple CPU sources). As far as I know, the 88k was only developed by Motorola. Mitch stated that it was developed as an independent project in parallel with 68060 development.

Quote:

We can note here that the 68K survives in the 68K ColdFire. Which is more like a cut down 68K since they would be less compatible than a 6502 is against a 6510. Another Motorola, the DragonBall, is also based on the 68K.


The ColdFire is a separate architecture thus *not* 68k. There have been many successful 68k based (non-ColdFire) embedded CPUs. The Dragonball was early based on the 68000 ISA while most later 68k embedded CPUs used the CPU32 ISA.

Quote:

A touch of irony, the ColdFire is said to run up to 300Mhz, a speed the 68080 still cannot touch. In it's WIP FPGA state.


You are comparing an FPGA CPU to an ASIC. An ASIC can more easily achieve higher clock speeds and should have better performance/clock. The FPGA 68080 (and even older fab 68060) have better performance/clock than most if not all ColdFire ASICs though. Performance has probably been a low priority for the ColdFire CPU designs. ColdFire ISA decisions also prioritize power and area over performance.

Quote:

Regarding money, I suppose Apple sales just didn't count for enough, since the 68K was being used in Apple Macs and Intel had deeper market penetration. There was just so much more money going into Intel, the way I see it. Yes the Amiga contributed, but Commodore helped it to contribute less.

As to being advanced faster than PPC, a PPC what? PPC back then or PPC now? In any case sure I think 68K could have matched anything for PPC or x86. The 68K is like an 80x86. They are both CISC designs. The 68K was a better and more modern design than the x86 ever was, so it could certainly compete with it. We only have to look at how they hacked and refactored the old x86 design over the centuries to become the monster it is today. So, with that in mind, I certainly think the 68K could have been engineered to the same point with multiple cores, 64-bit data and address bus, 64* registers, 512 bit vectors and Ghz speeds!


Yes, x86 won because of market forces at that time. Not only had x86 CPUs become commodities for PCs but 3D games placed emphasis on performance over other characteristics of the CPU. The 68060 was one of the best overall CPUs for a PC in 1994.

Pentium@75MHz 80502, 3.3V, 0.6um, 3.2 million transistors, 9.5W max
68060@75MHz 3.3V, 0.6um, 2.5 million transistors, ~5.5W max *1
PPC 601@75MHz 3.3V, 0.6um, 2.8 million transistors, ~7.5W max *2

*1 estimate based on 68060@50MHz 3.9W max, 68060@66MHz 4.9W max
*2 estimate based on 601@66MHz 7W max, 601@80MHz 8W max

The 68060 is the clear winner in PPA (Power, Performance and Area) often used to evaluate embedded CPUs today. The 68060 is 42% more energy efficient and is using 21% fewer transistors compared to the most comparable in order Pentium while giving similar performance (and better performance/MHz than the OoO PPC 601). The Performance is similar between these CPUs with the Pentium only having an advantage when it was clocked up due to mass production (for embedded uses lower clock speeds and better performance/clock is more reliable and cheaper).

Yes, the 68k ISA could be improved and made 64 bit like the x86 to x86_64. The 68k has several advantages like more free encoding space. One of the reasons why the 68060 performed so well compared to the Pentium and PPC is that it has good code density (probably about 10% better than the x86). The lack of free encoding space in the x86 ISA made x86_64 code really fat. A 64 bit 68k ISA can have a larger advantage in code density over x86_64 than the 68k did over the x86. A 68k 64 bit ISA can more easily and cleanly support more powerful addressing modes used by 64 bit software as well.

Quote:

* By the time x86 was 64-bit it matched the 16 register count of the 68K, up from 8. So why not double what x86 is doing and then double again!


The 68k can *not* easily encode more integer registers without fattening code, adding complexity to CPU designs and making compiler support more difficult. The x86_64 really needed more than 8 GP registers as 8 requires many more accesses of memory/caches but 16 is actually a good number for most algorithms. The extra x86_64 integer registers are not free either as the instructions are bigger unlike using all 16 68k registers. It would be possible for the 68k to free up the A4-A6 registers by using PC relative addressing more and accessing the frame pointer data from the SP (more efficient use of registers). It would be easier to double the number of 68k FPU registers from 8 to 16 while maintaining good compatibility. The 68k could have a SIMD unit or other units with more registers.

Quote:

pavlor wrote:
@g01df1sh

68k was no longer competitive in the early 90s.

In 1993, Intel had 486DX/2 66 MHz and Pentium... Motorola 68040 33 MHz. And no, there is no magic that will make 68040 faster than these Intel CPUs, far from that.



The 68k was more competitive with x86 in the early '90s than PPC is with x86_64 today. The 68040 had better performance per clock than the 486 but it was *not* a good design.

Quote:

ErikBauer wrote:
@pavlor

In 1993 68040 could run at 40Mhz, with a comparable speed to the 486/66Mhz, too bad it was way more expensive.

By 1994 68060 came out, a brilliant competitor to the Intel Pentium, but it was too late and could not be clocked at high enough clockrate to compete with it's Intel counterpart.

No, magic could not have saved 68K family... a better engineering that could allow a better performance/cost ratio maybe could have, maybe not: Microsoft and Intel were already growing titans by then.


The 68040 could not easily be clocked up because of heat which also added to the cost of manufacturing. The 68060 was a new improved design which solved the problem and others. It really is a great start of what should have been many CPUs based on this design (the 68k could have stayed in order longer than the Pentium which moved quickly to OoO). I believe it could have been clocked up with minor changes and faster die shrinks (there are 50MHz rated 68060s running at 100MHz with passive cooling). Motorola didn't want the 68060 competing with PPC on the desktop where they chose the semi-standardized commodity PPC to put an end to the commodity x86. There was a lot of RISC hype at that time about limitless clock speeds and simple CPU designs with ultra sophisticated compilers.

 Status: Offline
Profile     Report this post  
BigD 
Re: 68k Developement
Posted on 20-Jul-2018 19:41:07
#15 ]
Elite Member
Joined: 11-Aug-2005
Posts: 7307
From: UK

@g01df1sh

This seems a very similar thread to this one!


https://amigaworld.net//modules/newbb/viewtopic.php?topic_id=40773&forum=14&start=420&viewmode=flat&order=1

_________________
"Art challenges technology. Technology inspires the art."
John Lasseter, Co-Founder of Pixar Animation Studios

 Status: Offline
Profile     Report this post  
OneTimer1 
Re: 68k Developement
Posted on 20-Jul-2018 21:44:33
#16 ]
Cult Member
Joined: 3-Aug-2015
Posts: 962
From: Unknown

Quote:

g01df1sh wrote:

I was just wondering why 68k Development stopped and Motorola chose PPC.


Because there was no big customer left in the PC / Workstation market using this CPU.

The 68k got further developments for the embedded market, known products where Coldfire or Dragonball

https://en.wikipedia.org/wiki/NXP_ColdFire
https://en.wikipedia.org/wiki/Freescale_DragonBall

Quote:

g01df1sh wrote:

If money was not a issue


Money is always an issue and Freescale didn't want to waste it on a dead market.

Quote:

Could the 68k be advanced to the point that it is faster than PPC when using Asic. Vampire is proving that 68k still has room for advancement.


Yes, it could you only need a big development team like they used for the new pentium families and a chip factory with the most state of the art technology.
And you will get a high performance multi core chip no one would need an no one would buy.

 Status: Offline
Profile     Report this post  
matthey 
Re: 68k Developement
Posted on 20-Jul-2018 23:22:48
#17 ]
Super Member
Joined: 14-Mar-2007
Posts: 1968
From: Kansas

Quote:

OneTimer1 wrote:
Quote:

g01df1sh wrote:

I was just wondering why 68k Development stopped and Motorola chose PPC.


Because there was no big customer left in the PC / Workstation market using this CPU.


It is true that the Workstation market abandoned the 68k with the RISC hype. This was a high margin market and more profitable than the likes of Apple sales. Mitch Alsup said, 'When I was at Moto, we had a saying "Apple paid for the FAB, but you made no profit on them"'

Quote:

Quote:

g01df1sh wrote:

If money was not a issue


Money is always an issue and Freescale didn't want to waste it on a dead market.


It was much more expensive to bring a CPU to market back then. Only large companies with their own fabs did this while today we have many fabless semi-conductor small businesses. There would have been some risk for Motorola/Freescale to have continued the 68k line on the same scale but the 68060 was a nice CPU for embedded use as 32 bit CPUs starting being used more for this purpose. Sadly, Motorola/Freescale marketed and pushed the PPC for embedded use instead despite the inferior code density and unfriendlyness of the PPC. Motorola/Freescale lost most of the huge embedded market to ARM by shoving PPC down customers throats and not upgrading the much preferred 68k to be higher performance.

Quote:

Quote:

Could the 68k be advanced to the point that it is faster than PPC when using Asic. Vampire is proving that 68k still has room for advancement.


Yes, it could you only need a big development team like they used for the new pentium families and a chip factory with the most state of the art technology.
And you will get a high performance multi core chip no one would need an no one would buy.


You don't need as big of a team or as much money to make a CPU today. The Vampire guys have done most of the work required to make an ASIC CPU and they are backyard amateurs. They could probably have a cheap ASIC made for under $1 million and perhaps less if certain people liked their design and it met certain requirements. The CPUs could be mass produced for a few dollars using older under-utilized fab technology (I know a CEO who wants to mass produce embedded CPUs for IoT sensors and is a fan of the 68k but the Vampire does not meet his requirements). We are talking about a CPU more along the lines of an ARM Coretex-A53 in a Raspberry Pi which is also in order for better energy efficiency while lacking good single core performance. The 68060 design has already shown that the 68k can be competitive here.

It is unlikely anyone can compete with Intel for the high end PC CPU market. These CPUs cost billions of dollars to create and manufacture and have more transistors used for caches than any 68k CPU used for the whole CPU. Mitch Alsup went on to work for AMD designing an Atom like CPU which was never sold after his 88k work. He said, "A New design team should not target faster than 1/3rd of Intel's fastest design as a first effort. Only after the team learns the idiosyncrasies of the tools and FAB can they target 1/2 Intel frequencies."

 Status: Offline
Profile     Report this post  
retro 
Re: 68k Developement
Posted on 21-Jul-2018 15:46:47
#18 ]
Super Member
Joined: 16-Dec-2003
Posts: 1049
From: Unknown

@matthey

(I know a CEO who wants to mass produce embedded CPUs for IoT sensors and is a fan of the 68k but the Vampire does not meet his requirements)

way does it not meet his requirements ???.....what does it need ??
MMU/FPU more mhz or less power what does lack since it can't be used ?

 Status: Offline
Profile     Report this post  
Hypex 
Re: 68k Developement
Posted on 21-Jul-2018 16:35:37
#19 ]
Elite Member
Joined: 6-May-2007
Posts: 11180
From: Greensborough, Australia

@matthey

Quote:
My understanding from Mitch Alsup (one of the 88k chip architects who is active on comp.arch) is that the decision to drop the 88k and go with PPC was mostly politics. There were many competitive and similar RISC architectures to choose from at the time and PPC was an attempt to standardize on a single RISC architecture (probably pushed by computer manufacturers like Apple to have multiple CPU sources). As far as I know, the 88k was only developed by Motorola. Mitch stated that it was developed as an independent project in parallel with 68060 development.


Politics. How typical. And yet, it didn't help with PPC in the end, when it died on the desktop. Even if it survived in the other markets. Which ARM has taken over just about.

Quote:
The ColdFire is a separate architecture thus *not* 68k. There have been many successful 68k based (non-ColdFire) embedded CPUs. The Dragonball was early based on the 68000 ISA while most later 68k embedded CPUs used the CPU32 ISA.


The ColdFire is referred to as ColdFire/68k in the outside world so if it is totally separate and not at all opcode compatible in anyway then a fairlytale has been spread. It was also considered as a replacement CPU to speed up the Amiga. As below. In any case I didn't say
they were compatible.

http://www.vesalia.de/e_dragon.htm

Quote:
You are comparing an FPGA CPU to an ASIC.


Thus why I said the 68080 was in a WIP state.

Quote:
Yes, x86 won because of market forces at that time. Not only had x86 CPUs become commodities for PCs but 3D games placed emphasis on performance over other characteristics of the CPU. The 68060 was one of the best overall CPUs for a PC in 1994.


I wasn't aware the 68060 could be used in a PC back then.

A Mac, yes, but a PC?

Quote:
The 68060 is the clear winner in PPA (Power, Performance and Area) often used to evaluate embedded CPUs today.


If only that was to last.

Quote:
One of the reasons why the 68060 performed so well compared to the Pentium and PPC is that it has good code density (probably about 10% better than the x86).


It could do 8, 16 or 32 operations without addon hacks. And 64-bit math by combining registers, though that's a work around, for not having 64-bit wide registers. Though it did have 80-bit in FPU which was unheard of or matched today?

I almost laughed when I found x86 had byte sized instructions. Yes I know it came from 8-bit world but still. I suppose it can have alignment restrictions lifted.

Quote:
The 68k can *not* easily encode more integer registers without fattening code, adding complexity to CPU designs and making compiler support more difficult.


I, of course, purposely blew up the register count out of humour.

Quote:
The 68k could have a SIMD unit or other units with more registers.


Which is what they are putting in the Vampire 68K. As well as a 64-bit wide register file as I understand it. Though it disturbs me they are basing SIMD on Intel MMX. And they think themselves true to the 68K.

I know the game. Trying to turn the Amiga into another PC. Turning the 68K a K86!

Last edited by Hypex on 21-Jul-2018 at 04:39 PM.

 Status: Offline
Profile     Report this post  
matthey 
Re: 68k Developement
Posted on 21-Jul-2018 17:24:06
#20 ]
Super Member
Joined: 14-Mar-2007
Posts: 1968
From: Kansas

Quote:

retro wrote:
@matthey

(I know a CEO who wants to mass produce embedded CPUs for IoT sensors and is a fan of the 68k but the Vampire does not meet his requirements)

way does it not meet his requirements ???.....what does it need ??
MMU/FPU more mhz or less power what does lack since it can't be used ?


He needs the core to be open. This means the complete HDL source is available to him, can be modified and is patent free. All patents have expired on the 68k so it is part way there. He would also like the code and build system to be portable and verified for an ASIC where the Vampire is highly optimized for a particular type of FPGA and the logic elements inside. Here the philosophies are far apart.

He does *not* need an MMU. He does need some DSP/SIMD work done. Power needs to be fairly low but it needs to be balanced with performance. He wants a small footprint and highly values code density. He would like to eventually have simpler low end (low power 32 bit CPU with DSP) and advanced high end (higher performance 64 bit CPU) sensors but most architectures and CPU designs don't scale well. He was looking at a more modern SuperH (most patents expired) design but the SuperH has severe limitations and some of his information on code density was skewed which I pointed out (the 68k has better code density and performance traits when optimizing for code size).

It may already be too late. Amiga related companies have failed to license their technology and make partners which could have helped them survive, especially for embedded applications. The Amiga world seems to be a small little planet isolated from the rest of the universe.

 Status: Offline
Profile     Report this post  
Goto page ( 1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | 11 | 12 | 13 | 14 | 15 | 16 | 17 | 18 | 19 | 20 | 21 | 22 | 23 | 24 | 25 | 26 Next Page )

[ 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