Linked by Thom Holwerda on Thu 23rd Feb 2017 00:10 UTC
Google

Turns out the processor/SoC in the latest two ChromeBooks - the Samsung models - are part of a wider program by Google.

The OP1 is built by Rockchip, which has made ARM processors for a while and isn't especially well-regarded among US consumers. And, strangely enough, even discovering that Rockchip makes the OP1 took a bit of sleuthing. The company doesn't have its brand anywhere near the Chromebook Plus. Also, the chip is called the OP1, which implies that there's going to be an OP2 and OP3 and so on. What exactly is going on here? Just what is OP?

Well! Turns out there's a website for answering that exact question, helpfully named whatisop.com. OP is a designation for SoCs that are optimized for Chrome OS. Naturally, I assumed it was a Rockchip brand - but that's not the case at all. And the website ostensibly designed to explain OP to us doesn't tell us who owns it (and it's even registered anonymously), so OP strangely mysterious.

Mystery solved: OP is a trademark owned by Google, and bestowed on SoCs that meet a Google spec for a good Chrome OS device. Basically, if a Chromebook has an OP processor, it means that Google certifies that it’s been optimized for Chrome OS.

Everybody is racing towards ARM laptops. Intel's decision to sell Xscale is probably going to be looked back upon as one of the worst decisions in technology history.

Order by: Score:
Comment by stormcrow
by stormcrow on Thu 23rd Feb 2017 01:09 UTC
stormcrow
Member since:
2015-03-10

"Intel's decision to sell Xscale is probably going to be looked back upon as one of the worst decisions in technology history."

I really doubt it. The Xscale line was created to replace their aging RISC core line (i860, i960) out of their purchase of the StrongARM line from DEC in the late 90s. For all intents and purposes, what they retain from the sale of Xscale to Marvell still takes care of those needs. Xscale as a full GP CPU for standalone mobile devices was tangential to their needs and business focus.

Intel still holds an ARM license for architecture designs, so they can still incorporate whatever improvements ARM Ltd releases in the future into their future products in line with their product road maps.

Reply Score: 3

RE: Comment by stormcrow
by ahferroin7 on Thu 23rd Feb 2017 13:36 UTC in reply to "Comment by stormcrow"
ahferroin7 Member since:
2015-10-30

Entirely accurate, although there are many people who don't know the history behind this and it therefore looks to them like a bad business choice (although it's still nowhere near as bad as the whole iAPX 432 thing (object-oriented memory at the machine level is never a good idea)).

That said, if Intel is planning on doing any kind of GP ARM CPU, they had really better start working on it soon. AMD, APM, and a bunch of other companies already have good options on market and there's already support in the two OS's that matter for Intel's business (Windows and Linux, Apple's moving towards ARM laptops too, but they're pretty obviously using internally designed hardware), and it looks at least from outside like Intel thinks that the whole ARM 'craze' isn't worth their time, which may be the real worst business decision ever for them.

Reply Score: 6

RE[2]: Comment by stormcrow
by Megol on Thu 23rd Feb 2017 18:36 UTC in reply to "RE: Comment by stormcrow"
Megol Member since:
2011-04-11

Entirely accurate, although there are many people who don't know the history behind this and it therefore looks to them like a bad business choice (although it's still nowhere near as bad as the whole iAPX 432 thing (object-oriented memory at the machine level is never a good idea)).


That wasn't in the top 10 of i432 problems. Being slow (due to design choices), expensive, hot and badly supported by compilers and operating systems were among the most important failures of the project. The segmented memory model while causing some problems (partially because of short-sighted design and partially because of software portability) wasn't that painful in comparison.

My "favorite" hate of the i432 is the bit-aligned instruction set. The instructions are complicated in ways that makes pipelining etc. very hard and are designed to be microcoded.


That said, if Intel is planning on doing any kind of GP ARM CPU, they had really better start working on it soon. AMD, APM, and a bunch of other companies already have good options on market and there's already support in the two OS's that matter for Intel's business (Windows and Linux, Apple's moving towards ARM laptops too, but they're pretty obviously using internally designed hardware), and it looks at least from outside like Intel thinks that the whole ARM 'craze' isn't worth their time, which may be the real worst business decision ever for them.


ARM revenue per company v.s. Intel revenue for their x86 chips tells another tale. Intel can reduce per chip costs if they wanted to making x86 more popular (they don't because they want to make money - not inexpensive chips).

But even if Intel want to skip x86 why would they go to ARM? There are many alternatives including rolling their own RISC chip, making a x86 "light", make RISC-V chips etc. There isn't anything magic about ARM and most ARM software can be ported to any other reasonable architecture.

Reply Score: 3

RE[3]: Comment by stormcrow
by Vanders on Thu 23rd Feb 2017 22:26 UTC in reply to "RE[2]: Comment by stormcrow"
Vanders Member since:
2005-07-06

There are many alternatives including rolling their own RISC chip, making a x86 "light", make RISC-V chips etc.

"Going their own way" doesn't tend to be a winning strategy for Intel.

Reply Score: 3

RE[4]: Comment by stormcrow
by timl on Fri 24th Feb 2017 19:14 UTC in reply to "RE[3]: Comment by stormcrow"
timl Member since:
2005-12-06

"Going their own way" doesn't tend to be a winning strategy for Intel.


Even the Apple Mac has shifted towards more commoditized hardware over the years.

Nevertheless, "Go your own way" remains a big hit for Fleetwood Mac ;)

Reply Score: 2

RE[3]: Comment by stormcrow
by bugjacobs on Thu 23rd Feb 2017 22:48 UTC in reply to "RE[2]: Comment by stormcrow"
bugjacobs Member since:
2009-01-03

For example rolling in a ARM core or four into a x86 design :-D

Reply Score: 1

RE[3]: Comment by stormcrow
by viton on Thu 23rd Feb 2017 23:26 UTC in reply to "RE[2]: Comment by stormcrow"
viton Member since:
2005-08-09

Intel can reduce per chip costs if they wanted to making x86 more popular (they don't because they want to make money - not inexpensive chips).

Where have you been? Intel not only reduced the price of their smartphone product to $0 and offered free engineering services but they also paid some money to manufacturers.

That's how they burned 10 billions. Without any positive result.

There isn't anything magic about ARM and most ARM software can be ported to any other reasonable architecture.
ARM and x86 are the only two widely supported architectures.
One step aside and you are left alone.

Edited 2017-02-23 23:27 UTC

Reply Score: 3

RE: Comment by stormcrow
by Bill Shooter of Bul on Thu 23rd Feb 2017 14:27 UTC in reply to "Comment by stormcrow"
Bill Shooter of Bul Member since:
2006-07-14

Could you elaborate more on what they do with what they retained from the sale of Xscale?

I thought I understood the general outlines of what goes on in chip makers. But apparently not. The i860, i960 lines were colossal flops, but I never heard of any new Intel RISC arriving after they bought strongARM.

Reply Score: 2

ronaldst
Member since:
2005-06-29

That was the worst decision. I think it cost their ceo his job. Intel is still in a pretty comfortable spot.

Reply Score: 3

Reminds me the bad old time
by Kochise on Thu 23rd Feb 2017 06:11 UTC
Kochise
Member since:
2006-03-03
Connecting the strategy dots?
by BlueofRainbow on Thu 23rd Feb 2017 14:57 UTC
BlueofRainbow
Member since:
2009-01-06

Could there be a link between the development of an OS from scratch (Fuchsia) and this custom SoC?

One interesting observation about the Samsung Chromebook Plus/Pro is that the screen is 12.3", is touch enabled, has a 3:2 aspect ratio over a 2600 x 1600 resolution. These specifications are closer to those of the Google Pixel 2013/2015 than previous Chromebooks. This could aid in "unification" of the visual experience.

Wouldn't be nice to start afresh without being hindered by huge X86 legacies of Linux, OS X, and Windows?

Reply Score: 2

Bill Shooter of Bul Member since:
2006-07-14

No, it would not. Linux anyways, is perfectly capable of running on whatever arch you can think of, and probably has been ported. Its pretty flexible. It also has a lot of thought put into it and a wide range of applications.

For a mobile niche phone like device, I can live without gnu userland. But not on anything bigger than that.

Since I'm certain I'm probably the most perfect target consumer of such a device, Google should take heed.

Reply Score: 3

Intel will do well...
by dionicio on Thu 23rd Feb 2017 15:08 UTC
dionicio
Member since:
2006-07-12

Something many times dismissed, is that even Corps self-impose ethical limits.

Reply Score: 2

What Else Does The OP Do?
by Pro-Competition on Thu 23rd Feb 2017 22:35 UTC
Pro-Competition
Member since:
2007-08-20

Setting minimum standards for compatibility is a good idea.

My question is: what else does the OP do? (I don't trust Google very far with privacy.)

Reply Score: 2