Intel straws and Camel's backs [WAS: life expectancy of 32-bit x86]

On 14 February 2018 at 14:32, D. Hugh Redelmeier via talk <talk@gtalug.org> wrote:
| > Almost NO 32-bit x86 chips are in current production. I think that | > Intel has some goofy SoCs for IoT applications that are limited to | > 32-bit but they really don't matter. | | No, they canned that line last year. They really were not very good.
This doesn't say that it is EoLed. But I do think I heard that they stopped work in this direction.
<https://ark.intel.com/products/91949>
The "segfault bug" (lock instruction (prefix?) bug) sounds bad: <https://en.wikipedia.org/wiki/Intel_Quark#Segfault_bug>
That's an awesome solution they implemented there: "don't use LOCK." Intel says "It's 2014, you don't need multi-threading!" There's this assumption among many computer geeks (until today I was among them) that Intel is the chip "gold standard." I _should_ have changed my mind on that after their spectacularly poor response to Spectre and Meltdown. And then there's the security nightmare of the IME, a horribly insecure computer-within-your-computer. But for some reason, I think this is the figurative straw that broke the camel's back: I'm going to be looking to other chip manufacturers after this. All my vehemence and reading https://en.wikipedia.org/wiki/Intel_Management_Engine shows that AMD has created their own IME-alike. Security is a thing of the past ... -- Giles https://www.gilesorr.com/ gilesorr@gmail.com

[warning: long, didactic and tangential.] | From: Giles Orr via talk <talk@gtalug.org> | There's this assumption among many computer geeks (until today I was among | them) that Intel is the chip "gold standard." The prime aesthetic in the computer world is simplicity. Every bit of complexity increases the chance of bug bites. (I admit that there is significant complexity inherent in many problems.) Complexity has an unrelated advantage for enterprises: it creates the proberbial moat around their business. This is in no way good for customers. In the computer world, all big businesses seem to depend on this kind of moat. Perhaps the only way to attack a business behind a moat is by disruptive innovation. (I know that term has been worn out; go read Christensen's "The innovator's Dilemma" from 21 years ago.) RISC architectures were disruptive but mostly lost anyway. Successful businesses are constantly trying to re-enforce their moat, but often the complexity gets beyond even them. Examples of complexity as moat: - x86 - MS Windows - MS Office - IBM mainframes Of course there are other kinds of moats: "intellectual property", domination of distribution channels, scale requirements, regulatory capture, bundling, cross subsidization,... They are often complementary. Intel has been very good at some things -- they had to be. I guess they haven't focussed on being good at things they don't have to be. ================ Speaking of Windows... I bought a couple of little computers last week and they arrived when I was out for TLUG last night. Today I decided to initialize Wondows before I wipe it. I tuned the first one on. Windows booted and we went through introductions. It decided unilaterally to apply updates. Luckilly, it didn't apply them all since I'm sure that would have taken a day. (There seemed to be no way to avoid setting up PIN security, which I did not want, but eventually I could cancell out of that process.) Then, before I get a desktop, it asks to apply updates (again). What could go wrong -- I'd hardly done anything to the box. Well, something did go wrong. But neither it nor I know what: Something wet wrong Windows can't be updated right no. We will attempt to update again later. You can contact Microsoft support for help with thie error. Here is the error code 0xc1900203e [OK] The "Microsoft support" is a button that fires up Edge browser on a page with a FAQ about upgrading to Win 10, which I was not doing. Sheesh. All my recent experience with Windows has been updating it and all has been bad. You would think that this would be something to get right. (Giles: the Windows experience I whined about last week is a completely different incident.)
participants (2)
-
D. Hugh Redelmeier
-
Giles Orr