On Fri, Jul 1, 2016 at 1:42 PM, Solomon Peachy <pizza@xxxxxxxxxxxx> wrote: > On Fri, Jul 01, 2016 at 11:35:27AM -0500, Bruno Wolff III wrote: >> If I could afford it, I'd pay a significant premium to get a Talos board. >> But as it is, I'm going to be stuck using second hand, pretty old computers >> for the foreseeable future. > > And even though "significant premium" is a polite understatement in this > case, the Talos was the cheapest [proposed] POWER8 solution by a > substantial margin. > > I've also been quite disappointed in the AArch64 offerings so far. > > With very few exceptions (eg Gigabyte's mostly-unobtainable MP30-AR0 or > some locked-down appliance-type servers) they've all been ultra-dense > high-end boxes (dozens of cores per U) priced into the stratosphere or > glorified cell phone devkits (eg Tegra-based stuff) that don't sport > enough memory (or other expandability) to run a modern desktop OS. So the thing is, I'm genuinely curious why you would want to run a desktop OS on either architecture. >From a development perspective, they likely are not going to be wide enough to attract a ton of manufacturer attention in the laptop/desktop form factors. They're a niche market at best, likely to be filled by Chromebooks in the aarch64 case. As a developer, you'd be targeting ChromeOS or Android in that case, not plain Linux. Certainly not Windows. All of that seems rather speculative for making it an attractive development target for a desktop OS and applications that run on one. >From a manufacturer perspective, you aren't going to make laptop/desktop boards. The margins on that would be extremely small, or your product would be so expensive nobody would buy it. What you're going to do is focus on things where you can get bang-for-the-buck. Either a consolidation play in the datacenter or in a massive scale-out workload that's energy efficient. In either case, you make the big expensive machines or you make smaller boards that can expand out in a rack. That leads to a datacenter/cloud market. Cloud instances are relatively cheap, and can be paid for by the hour. I would think we'd want to focus our efforts around good Cloud and Server Editions for these architectures. Doing so attracts both manufacturers and developers at the same time. Sure, it would be neat to have a desktop solution and I'm not saying we shouldn't try there. I simply think that unlike x86_64, it is likely the least important thing to focus on because the hardware won't really be suitable and the developer market is going to be mostly focused elsewhere for these architectures. josh -- devel mailing list devel@xxxxxxxxxxxxxxxxxxxxxxx https://lists.fedoraproject.org/admin/lists/devel@xxxxxxxxxxxxxxxxxxxxxxx
