Apollo4060 @ 100MHz

[NovaCoder]

Apollo's have faster chipset access vs Blizzard's faster FAST ram access but you have to factor in the speed of 68060 libs and CPU speed. Also depends on what you are trying to run, fps are more important to me than mips

[Ratte]

I hereby send a sign of life.

[alexh]

Welcome back.

I don't know if anyone cares but there was an experiment earlier this year by Chucky (famous Amiga Hardware dude) with Apollo 4040 cards which have only MACH130 CPLDs and he was able to get his to have been able to become 4060 and run at 100MHz.

https://wordpress.hertell.nu/?p=1399

[derSammler]

I will never understand the silliness of overclocking hardware whose components can not be replaced in case of failure. Even crappy Apollo cards should be treated better.

[alexh]

Originally Posted by derSammler

I will never understand the silliness of overclocking hardware whose components can not be replaced in case of failure.

BFG9060 could be considered a replacement? (Ok so it doesn't have SCSI but the SCSI on Apollo A4060 wasn't great.) It's only a matter of time before PiStorm32 comes to bigbox Amiga's.

It wasn't silly at the time. Overclocking was common place. Catastrophic failure was low. It enabled you to run things on your Amiga that would have otherwise been too slow. There was no alternative (other than not use an Amiga)

[Ratte]

After a long break (the children are older now) I have decided to replace my Apollo and Mediator with a BFG9060 and a FirebirdPCI.
The Apollo is nice and fast, but the BFG also has the faster RAM.
The SCSI interface of the Apollo is not really worth mentioning, as it is a passive port with CPU driver and on top of that has a problem with the 4GB barrier and splits the disk into several disks, each of which has a maximum of 4GB.

[Bruce Abbott]

Originally Posted by derSammler

I will never understand the silliness of overclocking hardware whose components can not be replaced in case of failure. Even crappy Apollo cards should be treated better.

Any evidence that higher clock speed burns them out?

[Calabazam]

Originally Posted by Bruce Abbott

Any evidence that higher clock speed burns them out?

In the late 90's, i remember to have read articles and seen pictures made with electronic microscopes showing internal damages inside CPU, like deformed pathways and displacement of substance, but i can't find again those sources online now.

[Calabazam]

Originally Posted by alexh

Welcome back.

I don't know if anyone cares but there was an experiment earlier this year by Chucky (famous Amiga Hardware dude) with Apollo 4040 cards which have only MACH130 CPLDs and he was able to get his to have been able to become 4060 and run at 100MHz.

https://wordpress.hertell.nu/?p=1399

Exactly! The need for the MACH131 for 1260/4060 Apollo was a false legend that lasted decades.
What is the true difference between MACH130 and MACH131, now only Jens could tell by he will never reveal the CPLD.

Nowadays, it is still nice to have an Apollo 68060 board but with CPU shortage one might prefer a modern board instead.

[tom256]

Originally Posted by Calabazam

In the late 90's, i remember to have read articles and seen pictures made with electronic microscopes showing internal damages inside CPU, like deformed pathways and displacement of substance, but i can't find again those sources online now.

How higher frequency can deform pathways and make displacement of "substance"?
The only enemy is heat and very high voltage, so keep You CPU cool, and all will be fine.

[Calabazam]

Originally Posted by tom256

How higher frequency can deform pathways and make displacement of "substance"?
The only enemy is heat and very high voltage, so keep You CPU cool, and all will be fine.

It’s not high frequency but of course heat and higher voltage to achieve overclocking and the phenomenon was called electromigration.
Here is an old reference with electronic microscope pictures I could find again (you can Google translate) https://www.hardware.fr/articles/165...migration.html


Envoyé de mon iPhone en utilisant Tapatalk

[Calabazam]

More about the electromigration :

https://www.howtogeek.com/829636/how...kill-your-cpu/

As for the Amiga, the question is : Is the 68060@rev6 working at high speed like 100 MHz within the normal capacity of this CPU or should it be considered an overclocking?

[tom256]

Originally Posted by Calabazam

More about the electromigration :

https://www.howtogeek.com/829636/how...kill-your-cpu/

As for the Amiga, the question is : Is the 68060@rev6 working at high speed like 100 MHz within the normal capacity of this CPU or should it be considered an overclocking?

Interesting, anyway I think You can sleep safe with Your MC68060 overclocked:
https://www.anandtech.com/show/15839...-kill-your-cpu

BTW MC68060 max voltage is 4,5V. Just keep it cool and everything will be fine.

[eXeler0]

Originally Posted by Calabazam

More about the electromigration :

https://www.howtogeek.com/829636/how...kill-your-cpu/

As for the Amiga, the question is : Is the 68060@rev6 working at high speed like 100 MHz within the normal capacity of this CPU or should it be considered an overclocking?

All rev6 060s are not exactly equal, but consider this:
In 1994 the 060 launched at both 50 and 66MHz. It was manufactured on a crappy 600nm (well back then it was actually measured in micrometers --> 0.6um).
The rev 6 came along in 1999 ish, at 320nm (or 0.32 um).
That's a pretty significant improvement in production process. So if 600nm parts could reach 66MHz, the rev 6 should handle 75MHz like a walk in the park.

And to have some practical reference. I have been running a rev 6 in my 1200, overclocked to 66MHz for many, many years with just a slim heatsink with no issues. (the limit is the Apollo card, not CPU but its fast and very stable so I havent touched it)
In the A4000, I run a rev 6 at 100MHz however, here I actually use a big heatsink and a fan just because there is plenty of space for it, so why not right. Its a permanent overclock on the accelerator so it never runs slower than that, and so far no issues with that one either.

I think the only reason they continued to ship 50MHz versions even at rev 6 was because their market wasn't "computers" anymore, it was a bunch of other stuff like telecom equipment etc and I think lots of those products didn't require faster clock, they were designed for a 50MHz part and low power and cool operation was a priority over higher clock speeds.
I have in the last 20 years never heard about a rev 6 that can't overclock above 75MHz.

[Ratte]

https://www.digipart.com/part/MC68060FE133

True or fake ?!

[eXeler0]

Originally Posted by Ratte

https://www.digipart.com/part/MC68060FE133

True or fake ?!

I think this was something the Natami ppl (Now Apollo team) ”discovered” and afaik its fake

http://warpclassic68k.blogspot.com/2...3-eng.html?m=1

[AJCopland]

Originally Posted by eXeler0

I think this was something the Natami ppl (Now Apollo team) ”discovered” and afaik its fake

http://warpclassic68k.blogspot.com/2...3-eng.html?m=1

I was part of that Natami effort long ago and Thomas did indeed get them and run them. I think they were missing the FPU but he had them at upto 120MHz. He did however have go through the same problem with fakes.

The problem was that no-one new where they came from, Motorolla/Freescale deny the existence of them but they were 68060 parts.

I got fed-up of the whole Natami affair not long after that so that's all I know

EDIT: my assumption is that they were remarked 75Mhz EC/LC chips or something