After 10 years, PA product lands on Mac finally, solute for DEC

[tempestglen]

From 2004-2006, PA semi (ex califonia PA division of DEC) designed a good chipset for Mac with PowerPC ISA, the PT6A, which was the only PPC microarchitect outside of AIM from scratch.

But Apple "betrayed" PA semi by chooosing Intel x86 product, the reason is unknown.

However, from today's situiation we can guess that the reason is not intel is better than PA semi, on the contrary, PA is better than intel, at least in SOC field. Because in 2006 apple's secret weapon was pad/phone and bought PA semi in 2008 eventually, and let PA team to design A6-10, now it's time for apple to finish its promise:

PA product on Mac!

https://developer.apple.com/library...al/APIDiffsMacOS10_12/Objective-C/Kernel.html


Added #def CPUFAMILY_ARM_HURRICANE

If apple use aARM Chip on Mac, a new era is coming, long live DEC, long live PA!

 

[Phynaz]

FYI, P.A. Semi was never a part of DEC.

 

[Eug]

Also, Steve Jobs came right out and said that Apple didn't buy PA Semi for PowerPC. He said they bought PA Semi for its talent.

Furthermore, Apple's A chips are not PA Semi's products. They are Apple's products. PA Semi never had an ARM product AFAIK. OTOH, its star talent already had prior experience designing ARM chips before PA Semi, which presumably was one of the reasons Apple bought PA Semi's talent.

BTW, the DEC connection is that the PA Semi founder was the DEC lead designer also dealt with ARM there, and then he left DEC before starting PA Semi. So if there is any "betrayal" it was done by PA Semi's founder with regards to DEC, but of course that's a stretch.

There is also a bit trivia here in that Apple has had a relationship with ARM all they way back in the 1980s, almost 30 years ago. Apple Newton, which some consider the father of the iPhone, was an Apple ARM product.

 

[tempestglen]

There were nearly 200 engineers in PA branch of DEC before 1997 when DEC was bought by HP, after 1997, the PA team leader leaded the whole team to become PA semi, none of them wanted to join Intel.

So PA semi was a leftover from DEC's Califonia branch. And from 1991 to 1996, their product was famous STRONGARM, the best arm microarchitecture at that time, after 1997 Intel bought strongarm and named it as Xscale.

This team is the best in designing arm chipset,

iPad - Newton PDA 2.0
A6 TO A10 - Strongarm 2.0

 

[Nothingness]

There is also a bit trivia here in that Apple has had a relationship with ARM all they way back in the 1980s, almost 30 years ago. Apple Newton, which some consider the father of the iPhone, was an Apple ARM product.

There's more than that: Apple also was one of the largest investors in ARM back then.

 

[tempestglen]

They found 'cyclone/twister" in MacOS 10.11. So the machine.h may not be a strong evidence.

 

[tempestglen]

There's more than that: Apple also was one of the largest investors in ARM back then.

Apple was the father of arm holding, who was called 'British apple' in 1990s, and apple provided most of the capital for the joint adventure because another partner acorn only provided "technology", a third company only provide 5% equity.

The reason why Apple founded arm is to design Newton PDA chipset, and in 1996? Apple was on the edge of bankrupt so had to sell most arm shares. But before that, Apple invited DEC's PA branch to successfully design strongarm and promote the arm chip to a new level.

For arm, Apple was the farther, DEC PA was the teacher, and now "farther'&'teacher" becomes into one, and Apple will teach every other arm chip company a lesson of how to design the best arm SOC.

Thanks for the talents from DEC and other new cruited engineers, Apple&PA was, is and will be the best arm CPU developer, will make Intel very uncomfortable in the next 10 years.

 

[tempestglen]

BTW, the DEC connection is that the PA Semi founder was the DEC lead designer also dealt with ARM there, and then he left DEC before starting PA Semi. So if there is any "betrayal" it was done by PA Semi's founder with regards to DEC, but of course that's a stretch.

https://en.m.wikipedia.org/wiki/Daniel_W._Dobberpuhl

He founded PA semi in 2003 and Sibyte in 1998, both after the DEC's disappeared in 1997. I am sure that he was not fighting alone, there was a team with him. It is said that seldom DEC's PA branch employees would join Intel in 1997 when DEC was finished.

https://web.archive.org/web/2008042...le-buys-pasemi-tech-ebiz-cz_eb_0422apple.html
In February 2007, P.A. Semi debuted a 64-bit dual core microprocessor which the company asserted was 300% more efficient than any comparable chips. It consumes only 5 to 13 watts running at 2 gigahertz

 

[Phynaz]

Because somebody that worked at DEC left to found another company doesn't mean the new company had anything to do with DEC.

 

[Nothingness]

[OT] For those interested in ARM history: https://www.semiwiki.com/forum/content/6258-free-copy-mobile-unleashed-history-arm.html
I don't know how accurate it is as I didn't read it yet.

 

[Roland00Address]

There is also a bit trivia here in that Apple has had a relationship with ARM all they way back in the 1980s, almost 30 years ago. Apple Newton, which some consider the father of the iPhone, was an Apple ARM product.

Apple was one of the 3 companies that FOUNDED arm, it was a joint venture between Apple (Apple was there for they wanted a chip for the newton and nothing existed so they help organized this joint venture), Acorn Computers (developer of RISC, the tech who engineered and designed the ARM chips), and VLSI Technology (a silicon hardware maker, aka who was going to make the ARM chips).

ARM originally standed for "Acorn RISC Machine." Note this joint venture for ARM for the newton had nothing to do with Steve Jobs for he was gone from Apple for several years then. It was Sculley apple who helped formed ARM (Sculley was the pepsi guy who forced out Jobs, note Jobs was trying to organize a coup for Jobs wanted more independence to spend money on creating new departments and projects). Now Sculley apple was not really a bad apple, it was good but not great, but the guys who followed Sculley (aka post 93) Spindler and Amelio are the people who almost bankrupt Apple.

Now Apple during the dark ages where they nearly went bankrupt, well when Jobs returned as intern CEO they only had enough money to keep the company alive for 90 days , well Jobs sold all of Apple's holding of ARM to keep the company alive. At the time (1997 to 2000) ACORN computers was facing a hard times and went defunct, yet their subsidiary ARM Holdings did an IPO in 1998 and eventually evolved into the ARM we all know and use today in over a billion mobile devices (plus many other computers like devices and hardware devices)

 

[dark zero]

Sadly ARM is now sold to Softbank and they easily could sell it to Intel... Killing all ARM chances to outperform x86.

Well... There is still MIPS alive.

 

[Ajay]

Sadly ARM is now sold to Softbank and they easily could sell it to Intel... Killing all ARM chances to outperform x86.

Well... There is still MIPS alive.

Somehow I think this wouldn't be easy.

 

[HiroThreading]

Sadly ARM is now sold to Softbank and they easily could sell it to Intel... Killing all ARM chances to outperform x86.

Well... There is still MIPS alive.

Yeah... I don't think this is likely.

 

[dark zero]

Somehow I think this wouldn't be easy.

It shouldn't be easy (China and Europe will automatically ban it, to not to say South Korea). The deal won't happen unless USA orders them to do it.

 

[Ajay]

It shouldn't be easy (China and Europe will automatically ban it, to not to say South Korea). The deal won't happen unless USA orders them to do it.

OK. I'm a bit confused as to why you said "they easily could sell it to Intel" in the first place then.

 

[dark zero]

OK. I'm a bit confused as to why you said "they easily could sell it to Intel" in the first place then.

Money. Softbank don't want to get all ARM for them. So getting the important patents and then sell the rest to Intel is more logic. Just like Google did with Motorola.

 

[Ajay]

OK, so from a their business perspective, that would possibly be an easy option for Softbank, but the likely protests from regulators, many in the ARM chip business and competitors to Intel would make it unlikely to be realized. Sound about right?