A POWER'ful Announcement Is Expected Tomorrow Changing The Open-Source Landscape

[CrystalGamma]

As much as I am hoping they will bring back the IBM Thinkpad; this announcement is about POWER chips and I don't think those things should really go anywhere near a laptop! Yes, desktops would be very interesting though!

In the Talos IRC channel there have been a few vague hints dropped that indicate that POWER may indeed be making a move toward the laptop market … though not with the current generation of silicon (POWER9), which as you noted is only optimized for constant load (like in a server), and rather with POWER10.

I for one am hyped about any announcement they will make (though 'industry-shaking' is a high bar I'm not sure they will meet), even though I'm not in the market for a POWER-based 'new classic ThinkPad' myself for the next few years.

[mos87]

Nippy X86 binary translation please =P

[edwaleni]

Here is my 2 cents. It won't be as big as we would like, but big as far as OpenPOWER goes.

Probably POWER10 in form factors outside of Sforza.

[Guest]

As much as I am hoping they will bring back the IBM Thinkpad;

That's simply impossible, as IBM doesn't own the brand, and they would have to do a major shift in their company again. They don't deal with customers these days.

If you expect something that can match proper ThinkPads, look at other companies. The patent for TrackPoint IV died recently, so anyone can put that pointing device in their designs.

[Guest]

In the Talos IRC channel there have been a few vague hints dropped that indicate that POWER may indeed be making a move toward the laptop market … though not with the current generation of silicon (POWER9), which as you noted is only optimized for constant load (like in a server), and rather with POWER10.

I for one am hyped about any announcement they will make (though 'industry-shaking' is a high bar I'm not sure they will meet), even though I'm not in the market for a POWER-based 'new classic ThinkPad' myself for the next few years.

Is POWER even viable for mobile computers? From what I understand it goes for perf at the cost of power consumption, and I assume that a POWER laptop is as viable as a Powerbook G5. I've said that somewhere else before, but RISC-V has a better potential of becoming the new mobile architecture, as it seems to be quite power efficient, which implies sensible heat output as well.

[-MacNuke-]

A POWER based SoC / SBC?

[CrystalGamma]

Is POWER even viable for mobile computers? From what I understand it goes for perf at the cost of power consumption, and I assume that a POWER laptop is as viable as a Powerbook G5. I've said that somewhere else before, but RISC-V has a better potential of becoming the new mobile architecture, as it seems to be quite power efficient, which implies sensible heat output as well.

Power ISA is an architecture that can have implementations that occupy very different points in the power/performance/area design space. The POWER brand of processors by IBM are just one series of chips, which has traditionally been designed for high-power high-load scenarios (which mostly means they didn't spend a lot of engineering time on intricate power management, particularly in the uncore). There is nothing in the way of ISA stopping them from designing a chip that will work well on a battery-powered system (laptop or even tablet/phone, if they cared about that). There are definitely microcontroller-class implementations of Power ISA out in the wild (and POWER9 embeds dozens of them for various odds-and-ends functions).

Power and RISC-V are too similar to make a big difference in terms of viability for mobile computers from the ISA angle. Whether any of them can displace ARM and x86 in those markets depends entirely on whether companies involved with them are interested in the mobile computing market and have the business sense to make inroads.

[Sonadow]

With virtually every single open-source application written with the underlying assumption that they will be compiled under x64, I consider POWER to be nothing more than just IBM's personal toy (and RISC-V to be just a novelty).

I'm not going to give POWER and RISC-V (and practically every other processor architecture) a second look until we reach the day where every application that makes up the Linux desktop can be built as-is and unpatched directly from upstream sources by an end-user with just:

• ./configure, make, make install,
• cmake make make install or
• meson ninja ninja install

as is the current situation with x64.

[sdack]

I really hope they'll announce a POWER platform that is price/performance competitive with amd64. Other than that I can't image anything else that might be big news.

We did already have this in the past. It was called PREP - PowerPC Reference Platform - and later CHRP - Common Hardware Reference Platform. Those were designed to compete with Intel x86 platforms, but it fell short in the end. It was all a bit too expensive and the performance couldn't quite reach that of x86. Nevertheless was it an awesome idea and I'm sure there are still many who wished it had come true.

However, I cannot imagine they're going to try this again. All signs point to x86 and ARM being the major platforms for the foreseeable future (x86 for servers/desktop and ARM for IoT). If anything then perhaps a mini POWER SoC solution for IBM to get into the IoT market and to compete with the likes of RaspberryPi & Co.. Now that I could get behind if the price is also right. But it would have to be outstanding, meaning, more powerful while less power-hungry and also not as hot as current ARM-based SoCs...

Maybe IBM is going to bury their POWER technology and starts favouring AMD CPUs?! This, too, would be big news.

[kpedersen]

I'm not going to give POWER and RISC-V (and practically every other processor architecture) a second look until...

This is silly; if everyone thought like this.. we would still be using i386.

It used to be that it was more likely a build would succeed with i686 and that x64 needed patches.

Just use a capable OS that has a decent package build system and can apply patches easily. The Raspberry Pi has shown that it can be done quite easily. More things run on that out of the box than a lot of smaller x64 Linux distros.