ColdFire Project?

[wildstar1063]

Is the Coldfire project still active at all?
 the web page is still there, but it has not
 been updated for several years it seems.


Thanks

Wildstar1063

[Zac67]

And probably won't be updated at all. There's much consent here that a Coldfire creates major problems in software compatibility and seems to be no good choice for the Amiga, running legacy software. It's not impossible, but once you've mastered all problems, it's probably not much faster than an '060 - if at all.

[little]

but once you've mastered all problems, it's probably not much faster than an '060 - if at all.

If you relegate it to the role oe emulating a 68k then it is not a good solution, but if (and only if) amigaOS and a bunch of key applications were compiled for the coldfire procesor it could achive quite nice speed improvements, specialy if the 400mhz v5 coldfire was used.

[Zac67]

Of course, that would provide a tremendous speed increase. But you'd have to recompile, which isn't an option for the vast amount of legacy software. Another way would be a patch database, but I doubt the user base could reach the critical mass for that.
For the emulation option it's a lot smarter to start with a faster and more readily available CPU like a cheap & fast x86.

One day someone will do an '060 simulation in FPGA. In the VERY far future...

[A1260]

its dead........

[AJCopland]

wildstar1063 wrote:
I the Coldfire project still active at all?

I wish. I think that Oli got it so far but ran into problems that really would have required a lot more work to resolve and that's before he got into the software side of things.

Its been suggested that using the ColdFire as a co-processor instead of a replacement 68k might make more sense much like the PPC cards are but if you're going down that route you might as well use PPC.

Anyway the topic is dead in most peoples minds though I'd like to see one made as a replacement for the 68060/40 cards since they're so expensive, old and hard to find... I mean which would people rather have? A slightly slower than 060 card based on a ColdFire? Or a full speed 060 based card... that you can never afford off ebay?

(assuming that Oli fnished the ColdFire project of course)

Andy

[rkauer]

 I think the best solution for the Coldfire instruction set issue is adding a parallel "small" Spartan to catch the instructions who are not compatible with 68k.

 That could be a way to fix all problems in a snap.

 Then we get the best of two worlds: the speed and low cost of  a Coldfire (compared to a 060) and the complete set of 68k instructions, with no issues.

 My two cents.

[Piru]

@rkauer

I think the best solution for the Coldfire instruction set issue is adding a parallel "small" Spartan to catch the instructions who are not compatible with 68k.

How would that work exactly? How can some external chip know what the CPU is doing?

[Ragoon]

The coldfire since core V4e is fully compatible with the 68040. The first generation (5102) was already fully compatible with the 68EC040. It is an enhanced version of the originals 68k (410 MIPS @ 260MHz for actual coldfire and 110 Mips @ 75 MHz for 68060).
The coldfire embedded network, pci, usb, ... controllers. It is a great integrated processor but not enough powerful to face the PowerPC. It is a good solution to make an Amiga classic compatible computer with dedicated chipset in a FPGA. But the developpement costs would be too expensive for interested people.

[rkauer]

Piru wrote:

How would that work exactly? How can some external chip know what the CPU is doing?

 I mean use a Spartan as a translator table to catch the instructions the Coldfire can't process in the right way. Indeed it will generate some slowdowns, but in the end we can get an operational CPU who behaves like a 060 ~44MHz, who is a major upgrade from old accelerators around (I mean 030 and 040s).

[Piru]

@Ragoon

The coldfire since core V4e is fully compatible with the 68040. The first generation (5102) was already fully compatible with the 68EC040. It is an enhanced version of the originals 68k (410 MIPS @ 260MHz for actual coldfire and 110 Mips @ 75 MHz for 68060).

They are not compatible.

Supervisor mode is totally different. User mode is different, and even with emulation library loaded the user mode has a small difference which cannot be trapped.

[Piru]

@rkauer

I mean use a Spartan as a translator table to catch the instructions the Coldfire can't process in the right way.

But how can you do that? How can you make the spartan interface with the CPU in such way?

[rkauer]

Piru wrote:

But how can you do that? How can you make the spartan interface with the CPU in such way?

 What I think is only a well-intended guess.

 This is almost the same as a PPC board acts. Those Spartan CPUs have near the double IO pins as the 68k CPUs, so is not a big worry in the hardware connection, but must be a worry in SW (I think a nightmare is the correct word).

 The voltage level is the same, clocks can be decoupled, instructions can be interpreted BEFORE trowed to Coldfire...

 My 2 cents...
 

[Plaz]

The coldfire mail list still has some activity as discussions continue about possible solutions. Basically there about about 6 instructions that behave differently on coldfire that can't be traped. I think it can be solved (Elbox must have done it) but it would take alot more work using a card much different than the current prototype.

Plaz

[downix]

Well, could a recompilation system do it?  Using AROS you could make an OS which would, rather than JIT emulation, re-compile the apps.  The coldfire is similar enough that it could work.