http://news.techeye.net/chips/globalfoundries-to-buy-ibm-semi


Apparently, Global Foundries is for now the preferred bidder, but talks are ongoing with Intel and (rumored)TSMC.

Should Intel win the bid, that would mean they would also be in possesion of PowerPC license and POWER chips.

So, in the end, Amiga might finally be running on Intel tech. Rejoice!

_________________

@WolfToTheMoon

The PPC development was a joint venture of IBM & Motorola, with some early input by Apple. While IBM may own patents on the design (architecture) and RISC instruction set, Motorola sold the 680x0 and PPC line to FreeScale and IBM sold the 32-bit embedded processor PowerPC products to Applied Micro Circuits Corporation. The 64-bit line of PPC CPU's used in application-specific integrated circuit (ASIC) has further been licensed to others. This confuses exactly which company makes what and owning licenses, with the specification now being maintained as an open Power ISA v.2.03 specification and a new reference platform for servers called PAPR (Power Architecture Platform Reference).

See: http://en.wikipedia.org/wiki/PowerPC from which I copied liberally and likely incorrectly.

@WolfToTheMoon

I always wondered why there wasn't an x86 chip that cloned the PowerPC. That is read in the RISC code like x86 reads in the CISC code and translate it to the internal RISC core. Could be faster going from RISC to RISC codes.

Well wondered if anyone would do it. Lots of clones around these days. Apple could go back to 64-bit PPC, AltiVec and stay with Intel.

@Hypex

Quote:

Hypex wrote: I always wondered why there wasn't an x86 chip that cloned the PowerPC.

At least there was a plan for a PowerPC cloning i386:

------------------------------------------------------------------------------------
PowerPC 615

The "PowerPC 615" is a little-known PowerPC processor announced by IBM in 1994. Its main feature was to incorporate an x86 core on die, thus making the processor able to natively process both PowerPC and x86 instructions.

It was pin compatible with Intel's Pentium processors and comparable in speed, so it could be used in standard PC boards

A version of Minix was written to run in PowerPC mode and launch x.86 and PowerPC applications. It proved the concept and a version of OS/2 for the platform also worked.
------------------------------------------------------------------------------------
Source: http://en.wikipedia.org/wiki/PowerPC_600#PowerPC_615

Last edited by Arko on 07-Apr-2014 at 06:23 PM.

_________________
AmigaONE. Haha. Just because you can put label on it does not make it Amiga.

I borrowed this comments from here (#27 & #28):
http://amigaworld.net/modules/newbb/viewtopic.php?topic_id=38873&forum=2&start=20&order=0

@WolfToTheMoon

Quote:

Should Intel win the bid, that would mean they would also be in possesion of PowerPC license and POWER chips.

It would?

@Hypex

Quote:

Hypex wrote: @WolfToTheMoon I always wondered why there wasn't an x86 chip that cloned the PowerPC. That is read in the RISC code like x86 reads in the CISC code and translate it to the internal RISC core. Could be faster going from RISC to RISC codes. Well wondered if anyone would do it. Lots of clones around these days. Apple could go back to 64-bit PPC, AltiVec and stay with Intel.

As you may recall, that is what AMD did to outperform Intel's chips in the 90's; they developed a fast RISC core and a translator that converted CISC X86 instruction to run on it. At the time though, the PPC was a "convertible" chip running big or little Endian and was not meant to copy Intel, but that a new direction for other (non-PC) operating system and environments (business, education, server... You know $$$, IBM's real initials).

@cymru

Quote:

cymru wrote: they developed a fast RISC core and a translator that converted CISC X86 instruction to run on it.

Thats the normal way like all Intel CPUs startiung with 8080 (maybe 4004) where built.

Inside was a VLIW/RISC core interpreting the CISC commands and this is the reason why the 8080 needed more than 4MHz external clock to compete with a 6502 that used Random Logic http://en.wikipedia.org/wiki/Random_logic.

The 68000 had a built in VLIW core with did the same using 70bit long nano instructions.

_________________
AmigaONE. Haha. Just because you can put label on it does not make it Amiga.

I borrowed this comments from here (#27 & #28):
http://amigaworld.net/modules/newbb/viewtopic.php?topic_id=38873&forum=2&start=20&order=0

@Arko

Uh, where did you read this?

@Arko

Quote:

At least there was a plan for a PowerPC cloning i386:

Amazing. Now, if this was around the time of the AmigaOne, and a PC motherboard could have been used with a special PPC chip it would have been cool to see.

Being that the AmigaOne was almost there, the only major difference between that and a PC was firmware and a CPU.

@cymru

Quote:

As you may recall, that is what AMD did to outperform Intel's chips in the 90's; they developed a fast RISC core and a translator that converted CISC X86 instruction to run on it.

I thought by now they would have dropped the translator and focused on the core so there were no bottle necks to hold it back. After all supporting an old machine code instruction set when assembler is past tense is not movng computing forward. Especially when you need x86 emulators to run x86 code on your modern x86, it becomes a moot point.

That said, I do wonder the specifications of the RISC core and how it compares with ARM and PPC. Better or worse? Is there a stnad alone version.?

@cymru

Quote:

cymru wrote: @Arko Uh, where did you read this?

"M68000 Familie Teil1 Grundlagen und Architektur" from 1984

But you could find a lot of resources about 68k internal Micro + Nanocode:

http://people.cs.clemson.edu/~mark/uprog.html
http://www.easy68k.com/paulrsm/doc/dpbm68k1.htm

---

But before praising VLIW (that was not present for coders) as ultimate solution, it was chosen because it made the design easier and used less on chip resources. There are free 68k soft cores available, that have a higher performance per clock ratio than the original 68k.

The 68k could have been faster it wasted on clock cycle just for polling the bus, something you won't find on a 6502.

Last edited by Arko on 11-Apr-2014 at 06:48 PM.
Last edited by Arko on 10-Apr-2014 at 05:38 PM.

_________________
AmigaONE. Haha. Just because you can put label on it does not make it Amiga.

I borrowed this comments from here (#27 & #28):
http://amigaworld.net/modules/newbb/viewtopic.php?topic_id=38873&forum=2&start=20&order=0