This Project is Going to Create Linux Laptops … Based on PowerPC

Recently, a new group has appeared in cyberspace with the goal of creating a line of Linux computers based on the PowerPC architecture. They are currently raising money and are in the early planning. The question is: Can they pull it off? And, if they do, will it be worth it?

What is PowerPC again?

First off, how many of you have heard of PowerPC? Go ahead and raise your hand if you have. Undoubtedly, most people remember PowerPC from old Macs. For you whippersnappers, Apple used PowerPC processors in all their computers until the early 2000s.

PowerPC is a computing architecture created an alliance between Apple, IBM, and Motorola. PowerPC technology is based on reduced instruction set computer (or RISC). This means that RISC chips work by processing a series of short, general instructions. This is different than the x86 processors that run most computers today. Those chips use the Complex instruction set computer (or CISC) architecture, which uses complex or multi-step instructions to work.

Currently, PowerPC can be found in embedded systems and niche computers, like the AmigaOne series.

Suggested read
This Open Source Mini Computer Will Delete Data If Tampered With

Meet Linux PowerPC Notebook Project

Linux PowerPC

The Linux PowerPC Notebook Project is made up a group of Linux fans who want to revitalize PowerPC use. To get the ball rolling, they are currently working to raise money in order to hire a company to design a modern PowerPC motherboard. In their plan, this is just the phase one of creating a Linux PowerPC laptop. The company that they want to hire (ACube) has experience creating PowerPC hardware, including the AmigaOne 500. So far the group has raised € 4,310 out of a total goal of € 12,600.

Even though creating the motherboard is step one of a long road, the team already has an idea of what they want the end product to look like. Below is a list of specs that they hope to have in the future PowerPC laptop. These specs could change in the future

  • CPU: NXP T208x, e6500 64-bit Power Architecture with Altivec technology – 4 x e6500 dual-threaded cores, low-latency backside 2MB L2 cache, 16GFLOPS x core
  • RAM: 2 x RAM slots for DDR3L SO-DIMM
  • VIDEO: MXM Radeon HD Video Card ( removable)
  • AUDIO: sound chip, audio in and audio out jacks
  • USB: 3.0 and 2.0 ports
  • STORAGE:
    • NVM Express (NVMe), M.2 2280 connector
    • 2 x SATA
    • 1 x SDHC card reader
  • NETWORK:
    • 1 x ethernet RJ-45 connector
    • WiFi connectivity
    • Bluetooth connectivity
  • POWER: on-board battery charger and power-management
  • CHASSIS: standard notebook case 15,6”
Suggested read
This Startup Promises An Energy Efficient Linux Desktop

Advantages of PowerPC over x86

Why PowerPC? To begin with, the PowerPC architecture is much newer than the x86 architecture that we use every day. x86 was released in 1978 and PowerPC was released in 1991. Undoubtedly, this means that the creators of PowerPC learned from the mistakes and short comings of x86.

The rest of the reasons on the Linux PowerPC Notebook Project site are too technical for me to understand. For those of you who know about this stuff read on:

  • 64-bit architecture with a proper 32-bit subset
  • Wide vector instructions with large register file allow efficient data moving without use of off-chip memory
  • RISC architecture introduces Superscalar concept of multiple execution units: Branch, Fixed Integer, Floating Point
  • AltiVec SIMD vector processing
  • ISA 2.04/2.05/2.06 support multicore/multithreading, virtualization, hypervisor and Power Management

Based on my own research, the CISC architecture used more power and thus runs hotter than RISC. This is because a CISC chip completes fewer instructions per process, but each step is more complicated.

Also, it seems as though, PowerPC has an advantage when it comes to processing large amounts of data. According to an article in ComputerWeekly, PowerPC shines when it comes to “enterprise workloads”.

This includes “databases, data warehouses, data transaction processing, data encryption/compression, and certainly in high-performance computing, which most in business think of as analytics.” PowerPC comes out ahead in this regard because virtualization and hypervisors are included in the architecture. In the x86 architecture, most hypervisors are third-party products.

Suggested read
Gratipay: An Open Source Startup Helping Open Source Projects

Final Thoughts

Would I buy a PowerPC based Linux laptop? It would depend on two things:

  • If I could afford it
  • whether it had a big performance boost over x86.

My question is: has PowerPC development kept up with x86 development? According to one article I read, the reason that Apple switched to Intel x86 was because they were disappointed with IBM’s speed in improving the PowerPC chips. Of course, that was over a decade ago. Who knows what advanced IBM would have made in that time.

I also wonder why not focus on creating an ARM powered laptop? After all, ARM used the same RISC instruction set as PowerPC and is used in a wide range of computering devices. ARM has exploded in popularity with the release of the RaspberryPi. I’m sure there are more Linux distros with support for ARM than have support for PowerPC.

On the flip side, PowerPC is a lot closer to actually being ready for desktop use. IBM has continued to pursue development of the PowerPC architecture through the OpenPOWER Foundation.

Regardless, the PowerPC Linux laptop is not going to appear overnight. Righ now they are at the planning stages. There are multiple hurdles they will have to get over and more money they will have to raise. I wish them good luck in the endeavors.

Would you buy a PowerPC Linux laptop? Where do you stand in the PowerPC vs x86 discussion? Let us know in the comments below. If you want to donate to the project or learn more, visit their website.

Add comment

E-mail is already registered on the site. Please use the Login form or enter another.

You entered an incorrect username or password

Sorry that something went wrong, repeat again!

27 comments

by Newest
by Best by Newest by Oldest

ARM DOES NOT use the same RISC instruction set as Power architecture!

Get your facts right before you state something like that.

ARM and Power5/6/7 are RISC, yes, but they don't have the same instruction set!

If that's the case, then you're also saying that SPARC from Sun/Oracle has the SAME instruction set as Power and ARM?

"After all, ARM used the same RISC instruction set as PowerPC"
This is not correct: both are RISC, but they use different instruction sets. RISC is a design philosophy, not an architecture.

Author

Thanks for the clarification. Which do you prefer ARM, x86, or PowerPC?

You may want to take a look at RISC-V.

Writing programs in GNU-C for GNU Linux, there are almost no difference between the 3. Going down to the internal HW peripherals and registers, one will experience a change in endianness between x86 and PowerPC originally made as a co-operation between Motorola (later Freescale), IBM and Apple - big endian was used by Motorola and IBM . The Linux kernel for x86 used to be the most tested of the 3, and PowerPC the second most tested. After Android for has captured the major part of the Smartphone market running on Cortex-Axx - 16/32 bit Thumb-2 instruction set and not the original 32 bit ARM RISC instruction set - this architecture may run on the largest number of systems.
BTW. The PowerPC T208x is from Freescale who merged with NXP which is now taken over by Qualcomm.

If I should select one of architectures for an embedded project, I would never select x86 because they have a very short lifetime. Freescale used to guarantee more than 10 years of lifetime through their longevity programme and they have both Cortex-A ARMs and PowerPCs in their QorIQ portfolio. I don't know if Qualcomm will continue that path.

Well, I learned the ARM instruction set for assembly language programming back in 1987 when the Acorn Archimedes was released. At that time, ARM was without doubt the best architecture, it was about 4 times faster than x86 (I used to emulate x86/MS-DOS on my Archimedes at the time) while still running cooler (the Archimedes did not need a fan).

Since then, a lot has changed, x86 has overtaken ARM in terms of speed and ARM found its niche in low-power embedded systems. So now I would say: it depends on what you need - but I do hope ARM will jump over x86 again one day and their new 64-bit architecture may be a first step in that direction. For assembly programming, RISC is nice and easy to learn, I never felt inclined to learn the rather arcane x86 instruction set. I have no hands-on experience with POWER/PowerPC, so I cannot comment on that.

And if I am allowed to dream freely, I hope one day an open architecture (such as OpenRISC or RISC V) will overtake them all.

That part of the Wikipedia article on PowerPC is wrong. It was the heat produced by PowerPC chips which made it unsuitable, especially for laptops. IBM was using the chip in high-end servers. What sense does it make that Motorola and IBM would make underpowered chips? Also the Wiki article makes references to clock speeds, which is totally inappropriate when comparing different architectures. It's not even that reliable for measuring speeds for comparable x86 chips, let alone risc vs. cisc. Someone should really correct the wikipedia article. That section was apparently written by an Apple fanboy who has his own fantasies about how the company makes decisions.

Ooops. It wasn't the Wikipedia article on PowerPC misreferenced and linked that I read, it was the article on Macintosh Computers with clock speeds. So I need to directly blame John Paul here. They were not "too slow". I think there's wishful thinking here that Jobs had such wonderful judgement and must have made the best decision possible. Not so. Many modern game consoles, including the Xbox 360, use powerpc and do so with plenty of speed. Yellow Dog Linux was the premier distro for PowerPC for quite some time. Don't know if they're still out there.

The last YDL release was 6.1, in November, 2008:

http://www.fixstars.com/en/technologies/linux/

So they're still out there, but the project is pretty comprehensively no longer under development.

I seem to remember that about the time Apple switched to x86, IBM were close to bankrupt and were getting rid of lots of their non-core businesses [eg. Xyratex hard-drives and their printers]. I suspect there would be little confidence in IBMs ability to produce the goods and they were always too expensive with everything they did. As an aside; someone bought the Alpha chip [china ?], which may still be in production.

Compaq bought the Alpha chip when they acquired properties of Digital (DEC). The IBM issues and PowerPC aren't really related. Remember it was a consortium including Motorola. The Xbox 360 processor, Xenon,
Nintendo GameCube, Wii, and Wii U processors are all PowerPC. Sony and Toshiba, use the PowerPC Cell processor inside the PlayStation 3. Most of the major CPU chips in automobiles are PowerPC. This article quotes exactly the wrong facts from Wikipedia. The author should have fact checked with another source.

ARM processors are plenty, why go for PowerPC? IBM has always been reticent with their PowerPC development. No doubt PowerPC is a very good processor, but the combination of PowerPC and IBM is a bit dicey.

In my opinion the article completely misses the point on why would one want to use a PowerPC laptop instead of an x86 one: no hardware backdoors (like Intel ME).

The Chromebooks offer x86 or ARM variants. The Samsung and Rockchip ARM options tend to be a little slower and quite inexpensive. The new Intel options seem to compete well with the ARM Chromebooks for battery life. PPC is an interesting architecture, but not sure if it is worth the trouble. We might not notice the difference except for the difficulty in keeping pace with Intel's improvements.

Apple stopped using PPC in the mid-2000s, namely 2005. Not in the early 2000s.

If it can run AIX in a VM, it would be a great development and learning platform. Hardware availability was one of many reasons Linux overtook the commercial UNIXes.

I hope the Linux laptop wouldn't be as expensive as the Ubuntu phone turned out. If priced competitively, it could do well. With a low power processor employed, it might have outstanding battery life ; something laptop users value highly. Would be good if it came in various form factors.

I can't see any practial reason for this whatsoever, for PPC to provide competition and push x86 and ARM would require billions of dollars. and a large market share. I understand how people get attached to certain technologies, many people feel that way about old consoles. But that's all this seems to be, people had good times on PPC back in the day and want to re-live it.

Security really is the key issue. The Intel / AMD universe is utterly compromised -- https://libreboot.org/faq.html#intelme Not that maybe PowerPC doesn't have its own issues, but probably not as bad as Intel / AMD. I'd encourage Linux users to work more with the AMD laptops available as Chromebooks

oops, I meant to say ... I'd encourage Linux users to work more with the *ARM* laptops available as Chromebooks. Here's a link on ARM-compatible Linux distros: https://www.linux.com/learn/4-fine-linux-arm-distros

I would. For what I do, which is mostly internet and wordprocessing locally, why not if it's intelligently priced.

Because using PPC brings back good memories your ok with losing performance, value and flexibility (in terms of your choice OS, distro, software)? I'm not judging I'm just curious as to why someone would consider this a good idea.

Abhishek Prakash

"if it's intelligently priced"
That's the part I am afraid of ;)

Yes. Sometimes a novel approach to upgrading/ modifying existing technologies yields huge dividends (in the form of lower costs, improved performance, new applications, etc.)

RISC architecture goes a bit further back than your article states. Motorola developed it with the 6803 and 6809 8-Bit CPUs used in the Color Computer Series from Radio Shack (excluding the micro Color Computer.) Using OS-9 one could run a Real Time Pre-Emptive Multi-Tasking system on the Color Computer 2, 3, and on a Color Computer 1 if it had the 6809 CPU. I loved my Color Computer 3 back when I had one and running OS-9 Level 2 on it made it one great computer. I had a BBS running and was formatting a HDD when a user dialed in and I was able to chat with him. You couldn't do that with MS-DOS. If the Linux Kernel for the PPC system has the Real Time Pre-Emptive Multi-Tasking abilities included it will make for an extremely fast, efficient and powerful system and I would definitely want one.

If they are attempting to use a "modern" or "different" platform, why choose PPC? Why note HP PARISC (relatively modern and very different) or try to reincarnate the Alpha AXP (modern even though it has been buried for more than a decade and still an arms length in front of the Wintel clash).

Personally I think the world would gain a lot by having AXP brought back to life. Much more so than jumping on the self-stalling PPC bandwagon.