PPC motherboard A1222 "Tabor" revealed

[Retrofan]

Few info yet: http://www.amiga-news.de/en/news/AN-...-00027-EN.html

[eXeler0]

(If true...) Why oh why choose a CPU that will cause incompatibility problems and decreased performance in existing software?

[jbenam]

Originally Posted by eXeler0

(If true...) Why oh why choose a CPU that will cause incompatibility problems and decreased performance in existing software?

I'm starting to wonder if Trevor has got some interest/shares in PowerPC companies, otherwise I can't understand why he's sticking to that

[Yakumo9275]

lol an incompatible fpu of all things. strange

[Locutus]

It makes you wonder what kind of design meeting they must have had to discuss this.

[eXeler0]

Originally Posted by Locutus

It makes you wonder what kind of design meeting they must have had to discuss this.

Lol, and to leave that meeting with the idea in your head that... "Yea! Let's go through all that trouble of making an entire friggin mobo with a "crippled" CPU and try to sell it to the most fragmented niche market in the entire world! What could possibly go wrong?? Let's commit to this great idea!"

[Schoenfeld]

The P1022 is a dual-core CPU. Even if you were to emulate all FPU commands in software, you'd have a spare core to do that, as OS4 is still a single-core OS.

Besides, it's not all-incompatible, but throws a few illegal instruction traps on commands not implemented (according to one of the developers I talked to last saturday). That's about the same as what the 060 CPU does for unimplemented commands of the earlier members of the 68k group. Would you make fun of those who designed with the 060 back then? Surely not.

Jens

[strim]

Also, NetBSD features FPU emulation for BookE CPUs with SPE since 2012. NetBSD also originally required FPU. The emulation works 100%.

Surely, if ACube/A-ONE chose this particular CPU, they knew it is possible.

[wawa]

possible perhaps, but it doesnt mean practicable. whether this is aeon or acube, they products usually contain design decisions andlimitations that couldnt be explained for years. that seems to continue.

[FOL]

If its better than whats currently available and works, whats the problem?
I love the way the word "crippled" is used. Its a nice little board, for entry level.

I agree with Jens, no one is moaning at the 060. Why did they design it the way they did.

[wawa]

I agree with Jens, no one is moaning at the 060. Why did they design it the way they did.

none was moaning at the 060? maybe because it was at least three times faster than what was available up that point and there was no alternative. neither one or the other is apparently the case here.

instead look at coldfire accelerators that have been as much hyped at their time as the board in question. check for availability of ultimate ppc containing the very same processor. sure, problems can be solved, but it remains to be seen if the result is satisfactory enough for it to become viable product. according to repeated attempts the projects with the cpus in question did not reach practicable state.

[FOL]

Originally Posted by wawa

none was moaning at the 060? maybe because it was at least three times faster than what was available up that point and there was no alternative. neither one or the other is apparently the case here.

instead look at coldfire accelerators that have been as much hyped at their time as the board in question. check for availability of ultimate ppc containing the very same processor. sure, problems can be solved, but it remains to be seen if the result is satisfactory enough for it to become viable product. according to repeated attempts the projects with the cpus in question did not reach practicable state.

How about we all wait and see. Pointless debating this until there are facts.

[wawa]

How about we all wait and see. Pointless debating this until there are facts.

maybe as pointless as it was to discuss ultimateppc, when it has been pointed out how this was an unneccessary effort, and apparently it occured to be truth. similarly it has been argued that coldfire is not suitable for 68k replacement, people tried anyway, retried, and realized that it was really the case.

the board here has been publicly presented while an amiga gathering, provoking expectations and discussion, you cannot now blame the public for it. you simply shouldnt have shown it off. there were much more interesting developments out there like apparently working natami board or sonnet warpos accelerator. we could now spend time talking of them instead.

[Schoenfeld]

Originally Posted by wawa

maybe as pointless as it was to discuss ultimateppc, when it has been pointed out how this was an unneccessary effort, and apparently it occured to be truth. similarly it has been argued that coldfire is not suitable for 68k replacement, people tried anyway, retried, and realized that it was really the case.

Errmm.. wait. First of all, everybody should know by now that I believe the PPC was about the worst thing that could have happened to the Amiga (mostly because of that red vs. blue war). However, I have the highest respect for the OS4 team and Gideon - I do believe that both have the required skills to make it happen. However, Gideon obviously chose to spend his spare time in a different way. Now the OS4 team is on the same CPU, and I am pretty sure that they wouldn't have started this work if there's no perspective for success.

Coldfire? That's a totally different story. There was only one person on it, and he chose to do what he could do a lot better: Stay within his core business (to deal with boats if I remember right). In other words: There was never any serious effort to use a Coldfire processor in the Amiga. That's "for good reason" in my opinion: While the Coldfire attempts to be binary-compatible, but behaves differently on *implemented* commands, the OS4 core team member I talked to claimed that implemented FPU commands work as they should, and unimplemented commands throw an exception. Although I haven't read any details in a datasheet, I trust Cosel Mincea on this: The P1022 may have an FPU with fewer commands, but it also provides the bits&pieces to work around that, just like the 68060 did.

Originally Posted by wawa

the board here has been publicly presented while an amiga gathering, provoking expectations and discussion, you cannot now blame the public for it. you simply shouldnt have shown it off. there were much more interesting developments out there like apparently working natami board or sonnet warpos accelerator. we could now spend time talking of them instead.

Sure - but in a different thread, please. This one is about Tabor - something that has all chances to become a working alternative for OS4.

Jens

[wawa]

the PPC was about the worst thing that could have happened to the Amiga

the red vs blue war wasnt caused by the choice of cpu, otherwise it would be aros vs mos/os4. the problem with ppc was imho (im just a user) that it resulted with overcomplicated and overpriced solutions to behin with, and made a budget hardware from the start almost impossible.

I have the highest respect for the OS4 team and Gideon

you are probably talking of people you know and deal with personally. i lack this perspective. i have to judge the achivements vs problems, promises and expectations. looking from here individuals within os4 team may be well competent and gifted, no doubt, but the track of record as a whole doesnt convince me that essential problems are to be overcome within years, maybe even decades.

what concerns ultimateppc, i doubt one throws the towel on a project which came so far and has been invested in that much, without a reason. personally i think the problem with ultimateppc was probably not even technical but conceptional one. the ppc part/bus practically being isolated from 68k/amiga side doesnt make much sense to me as an expansion. rather the ppc side should be designed as standalone, which is what we are witnessing now. thats much more reasonable so far, admittedly.

Coldfire? That's a totally different story. There was only one person on it

one person? i could count at least three attempts, an a3/4k coldfire cpu card design. the elbox approach (being demoed working, but not giving a desirable boost?) and the natami team. natami team has been given development boards by gensi, even though gunnar has been warned on a1k, him being still coldfire concept supporter. im not sure what they have tried or not, but they have given up. for good reason i guess. i would hate to see another block of incompatible binaries appearing on aminet, even if the system had been patched to boot on cf.

[matthey]

Originally Posted by Schoenfeld

The P1022 is a dual-core CPU. Even if you were to emulate all FPU commands in software, you'd have a spare core to do that, as OS4 is still a single-core OS.

Even if the 2nd core could be set to trap for the 1st core, wouldn't the registers still be from the 2nd core? I would expect overhead even if the cache is shared between cores. The 1st core would have to wait until the 2nd core is done before continuing.

Originally Posted by Schoenfeld

Besides, it's not all-incompatible, but throws a few illegal instruction traps on commands not implemented (according to one of the developers I talked to last saturday). That's about the same as what the 060 CPU does for unimplemented commands of the earlier members of the 68k group. Would you make fun of those who designed with the 060 back then? Surely not.

My limited understanding of the P1022 is that there are no FPU registers (GPR registers are used for floating point) so *all* PPC FPU instructions must be trapped. This in contrast to the 68060 where the most common 6888x instructions are executed in hardware like FMOVE(M), FADD, FSUB, FMUL, FDIV, FSQRT, FNEG, FTST, FCMP, FABS, FINT and FINTRZ. These are the core FPU instructions in hardware vs *no* FPU instructions in hardware.

Originally Posted by wawa

one person? i could count at least three attempts, an a3/4k coldfire cpu card design. the elbox approach (being demoed working, but not giving a desirable boost?) and the natami team. natami team has been given development boards by gensi, even though gunnar has been warned on a1k, him being still coldfire concept supporter. im not sure what they have tried or not, but they have given up. for good reason i guess. i would hate to see another block of incompatible binaries appearing on aminet, even if the system had been patched to boot on cf.

There were multiple attempts to execute 68k code on the ColdFire as you say. Executing 68k code on the ColdFire is even worse, IMO, than trapping all PPC FPU instructions on the P1022 because:

1) integer instructions are more common than FPU instructions
2) ColdFire has a few conflicting encodings, instructions which set the CCR differently and the stack uses different alignment which is incompatible

ColdFire requires pre-execution or on the fly patching of 68k executables. It looks like P1022 trapping of the standard PPC FPU instructions is straight forward but has much more overhead than trapping unimplemented 6888x instructions on the 68060. The P1022 FPU emulation would likely be adequate for light floating point use and a joke for heavy floating point use like using Blender.

[iggybeans]

Originally Posted by Schoenfeld

The P1022 is a dual-core CPU. Even if you were to emulate all FPU commands in software, you'd have a spare core to do that, as OS4 is still a single-core OS.

Besides, it's not all-incompatible, but throws a few illegal instruction traps on commands not implemented...

Jens

Jens, since the trapping will occur on the primary cpu, halting it, using the second core to execute the translation stage will not produce an increase in speed.

Originally Posted by matthey

...The P1022 FPU emulation would likely be adequate for light floating point use and a joke for heavy floating point use like using Blender.

Unfortunately, no one seems to follow this point and we are still getting "wait and see" comments.

Or comments that we somehow are acting against the community's best interests by pointing this out.

The cure to the problem is easy, btw.

Once these first 1000 boards are sold, design a board that supports a PPC with a standard fpu.
The price increase would amount to about $10 in parts.

I'm willing to wait AND pay a slight premium (which in this case could be really slight).

[utri007]

Will see how it performs. But it surely looks nice and Trevor did said that 500-700€ sounds expensive. So hopefully it will be cheapest Amiga OS4 hardware ever.

If there is a need for two separate versions, it is not a problem IF two versions is not required. There is already altivec versions of programs.

[rave]

That price tag!
In this day and age, for something so niche?
We amiga users must be pretty rich monetarily to even consider something at that cost.
Fair play to them though, I hope they do well, you never know what will happen.

[trixster]

That price indeed. Alsolutely bonkers.