What was the best microprocessor ever built?
« Sir Clive Sinclair - Inventor Of Cool | Main | UK To Have An Hour Of Colour TV Transmission A Day »
What was the best microprocessor ever built?
TrackBack URL for this entry:
http://www.electronicsweekly.com/cgi-bin/mt/mt-tb.cgi/53246
This page contains a single entry from the blog posted on May 25, 2009 2:26 AM.
The previous post in this blog was Sir Clive Sinclair - Inventor Of Cool.
The next post in this blog is UK To Have An Hour Of Colour TV Transmission A Day.
Many more can be found on the main index page or by looking through the archives.
Comments (12)
Wot? No 6809?
Posted by Peter Seed | June 2, 2009 12:16 PM
Posted on June 2, 2009 12:16
Well Yes, Peter, strictly speaking you're right, but the 6809 didn't catch on like the 6800 and 6502 -though it was technically superior to them. 'Best' is a sloppy old word and probably includes things like take-up.
Posted by David Manners
|
June 2, 2009 1:21 PM
Posted on June 2, 2009 13:21
I think the best microprocessor was the Am2901. This was a 4 bit cascadable device from AMD and was the basis of many minicomputers and device controllers in the 70's and 80's. It allowed great flexibility in hardware design and spawned a whole family of supporting devices. You could design integer or double precision floating point hardware with it. You could do really clever micorcode for it if the hardware was designed well. The Am2903 which came next included hardware multiply but it couldn't outperform the 2901 at simpler operations so it wasn't so successful. Architecturally, the IDT49C402 was superior to the AM2901 being 16 bit and having additional internal buses. But this came much later and didn't have the take up of the Am2901.
My 2c.
Chris Watts
Posted by Chris Watts | June 10, 2009 6:18 PM
Posted on June 10, 2009 18:18
Its a no brainer the best microprocessor ever made was the Zilog z80. Zilog took the best features of the Intel 8080 (later to become the 8086 grandfather of the PC), added some important new instructions which increased code and memory efficiency while retaining object code compatibility with the 8080.
Zilog lost the plot in the 1990's but they still exist but have lost their hip glow of the 70's. We still base our products on the Z80 architecture and some of our code can be traced back to our 1970's micro.
Zilog briefly came back with the truly excellent z280 which was a 16-bit version of the z80 with some more code enhancements, including an instruction cache, and lightening fast interrupt switching. We based our top line product on it but Zilog pulled the plug on the product.
How the mighty fall - all of the names that were poster children of the 60-70's have lost touch with their innovation roots.
FPGAs now allow anyone who can write a bit of Verilog to create their own microprocessor architectures and even clone their old favourites.
Posted by Don Clarke | June 11, 2009 2:47 PM
Posted on June 11, 2009 14:47
I think you're right Don, the Z80 swept the world
Posted by David Manners
|
June 11, 2009 2:59 PM
Posted on June 11, 2009 14:59
I voted for the 68020, as this was the best of the 68k series, however the real breakthrough - the first "mainframe on a chip" was really the 68000. Your system would not let me vote again.
The reality was that this microprocessor was the first with fully 32-bit internals even if only 16 data and 24 address lines came out.
Lanuched in the UK on 4th June 1979 at the IEE - I was there!
R. Mark Clayton
PS The nice lunch at the Savoy didn't influence me at all :------¬}
Posted by R. Mark Clayton | July 16, 2009 6:59 PM
Posted on July 16, 2009 18:59
Thanks Mark, a lot of people would agree with you. So much more weird the decision to supersede the 68000 with PowerPC.
Posted by David Manners
|
July 16, 2009 7:58 PM
Posted on July 16, 2009 19:58
Has to be the 68000. I did some machine code stuff with it in the 80's and it was a joy to use. When I looked at the Intel 80486 later, I couldn't believe how rubbish it was. The World would be a lot better place now if IBM had adopted the 68000 series in their PC's from day one. Let's not even talk about DOS...
Posted by Paul Sage | July 17, 2009 11:12 AM
Posted on July 17, 2009 11:12
Paul, I think most people would agree with you - both for engineering reasons and business reasons.
Posted by David Manners
|
July 17, 2009 11:18 AM
Posted on July 17, 2009 11:18
Wot, no 8051?!
Posted by Bill Davy | October 22, 2009 11:21 AM
Posted on October 22, 2009 11:21
For sheer longevity and flexibility my own choice is the 8086/8. I am a little biased as we had to run systems in excess of 150'C and the 8088 would tackle temps up to 190'C quite happily, RAM was a bit of problem though. I really grew to love the little guy.
Posted by rob gillatt | October 22, 2009 11:33 AM
Posted on October 22, 2009 11:33
It was the 8086/88 that did it for me. This might have been to do with the fact that the 86/88 was able to bridge the gap between embedded and desktop and had the most amazing (for its time)set of tools to work with. After nit picking with assembler and horribly tedious tools for the 68000 I had Intel's PLM at my disposal. The tools would still be burning rubber if they were available today.
The much criticised page memory mode was a real boon for context switching. You could write a near real time and really well protected multitasking embedded system back in the 1980's with no more than the PLM compiler. Of course it was designed for the Intel RM10 operating system, hugely deployed by telecoms and other industrial sectors in the 80's, not to mention the military.
As an added bonus the 8088 (if you bought the Harris MIL product)would run at 180'C almost indefinitely, quite a baby to have in your toolkit if you wanted to run in oil wells and aero engines.
Posted by rob gillatt | October 23, 2009 9:54 AM
Posted on October 23, 2009 09:54