Stories
Slash Boxes
Comments

SoylentNews is people

posted by martyb on Wednesday January 06 2021, @03:27AM   Printer-friendly
from the bit-flip-out dept.

Linus Torvalds On The Importance Of ECC RAM, Calls Out Intel's "Bad Policies" Over ECC

There's nothing quite like some fun holiday-weekend reading as a fiery mailing list post by Linus Torvalds. The Linux creator is out with one of his classical messages, which this time is arguing over the importance of ECC memory and his opinion on how Intel's "bad policies" and market segmentation have made ECC memory less widespread.

Linus argues that error-correcting code (ECC) memory "absolutely matters" but that "Intel has been instrumental in killing the whole ECC industry with it's horribly bad market segmentation... Intel has been detrimental to the whole industry and to users because of their bad and misguided policies wrt ECC. Seriously...The arguments against ECC were always complete and utter garbage... Now even the memory manufacturers are starting [to] do ECC internally because they finally owned up to the fact that they absolutely have to. And the memory manufacturers claim it's because of economics and lower power. And they are lying bastards - let me once again point to row-hammer about how those problems have existed for several generations already, but these f*ckers happily sold broken hardware to consumers and claimed it was an "attack", when it always was "we're cutting corners"."

Ian Cutress from AnandTech points out in a reply that AMD's Ryzen ECC support is not as solid as believed.

Related: Linus Torvalds: 'I'm Not a Programmer Anymore'
Linus Torvalds Rejects "Beyond Stupid" Intel Security Patch From Amazon Web Services
Linus Torvalds: Don't Hide Rust in Linux Kernel; Death to AVX-512
Linus Torvalds Doubts Linux will Get Ported to Apple M1 Hardware


Original Submission

 
This discussion has been archived. No new comments can be posted.
Display Options Threshold/Breakthrough Mark All as Read Mark All as Unread
The Fine Print: The following comments are owned by whoever posted them. We are not responsible for them in any way.
  • (Score: 0) by Anonymous Coward on Wednesday January 06 2021, @04:13AM (6 children)

    by Anonymous Coward on Wednesday January 06 2021, @04:13AM (#1095458)

    In software, "shit happens" has been the accepted notion. We sent out update when we were forced to.

    Hardware used to be different. But that was like half a century ago. Hardware is as flaky as software, and that's been the case for a few decades now.

  • (Score: 3, Insightful) by canopic jug on Wednesday January 06 2021, @05:48AM (5 children)

    by canopic jug (3949) Subscriber Badge on Wednesday January 06 2021, @05:48AM (#1095502) Journal

    In softwareMicrosoft products, "shit happens" has been the accepted notion. We sent out update when we were forced to.

    Hardware used to be different. But that was like half a century ago. Hardware is as flaky as software, and that's been the case for a few decades now.

    There. Fixed that for you. The problem has not been software but rather Microsoft products. Over time that has translated to a general expectation of bad engineering in software and to an acceptance of bad design everywhere else. With both computer software and hardware that has become an expectation. Everywhere else, it has merely become accepted, much to the threat of our continued survival as a society. Bill Gates' most lasting legacy, if there is a civilization left after a few years, will be that he made bad engineering acceptable.

    --
    Money is not free speech. Elections should not be auctions.
    • (Score: 2) by Immerman on Wednesday January 06 2021, @06:17AM (4 children)

      by Immerman (3985) on Wednesday January 06 2021, @06:17AM (#1095510)

      Really? Microsoft is particularly bad, but I can't say that I've ever used flawless software. And I don't think it's just a problem with expectations - software engineering is HARD. A typical car only has around 1800 parts, and cars always have their faults, because engineers aren't perfect.

      If you take a single line of code as very roughly comparable in design complexity to a single car part (ranging from a single bolt out to a complex cast manifold), that means a single large scale piece of software can have around 1,000-5,000x as many "parts" as a typical car, and in any sane world you would expect at *least* a similar increase in flaws.

      • (Score: 0) by Anonymous Coward on Wednesday January 06 2021, @06:22AM (2 children)

        by Anonymous Coward on Wednesday January 06 2021, @06:22AM (#1095513)

        > Microsoft is particularly bad, but I can't say that I've ever used flawless software.

        On Linux they call this User Error or WONTFIX.

        • (Score: 0) by Anonymous Coward on Wednesday January 06 2021, @01:50PM (1 child)

          by Anonymous Coward on Wednesday January 06 2021, @01:50PM (#1095606)

          On Linux they call this User Error or WONTFIX.

          On Linux, they call it "unreproducible". More seriously, with free software, the user can actually fix the damn problem too.

          • (Score: 2) by barbara hudson on Wednesday January 06 2021, @03:54PM

            by barbara hudson (6443) <barbara.Jane.hudson@icloud.com> on Wednesday January 06 2021, @03:54PM (#1095655) Journal

            The user can fix a flipped bit on Linux? Don't think so.

            When you get weird errors , and you swap the ram, and it goes away, problem solved.

            There have even been cases where one stick will be more susceptible to EM interference from the power supply, and swapping slots so the vulnerable stick is now furthest from the power supplyfixes the problem.

            --
            SoylentNews is social media. Says so right in the slogan. Soylentnews is people, not tech.
      • (Score: 0) by Anonymous Coward on Wednesday January 06 2021, @03:46PM

        by Anonymous Coward on Wednesday January 06 2021, @03:46PM (#1095650)

        Yep, ACPI was a particular mess because MS refused to enforce Intel's rules about how the DSDT was written. You'd have systems shipping with DSDT so badly written that they would not compile at all on the reference compiler that companies were supposed to be using, but MS would include special code to tolerate it rather than just allowing the system to have working ACPI support. I remember having to dump and recompile it myself because the manufacturer was too lazy to properly code it. The thing is that in that case, it wouldn't have even cost them any development time as the core logic was correct, there were just a few minor things they chose not to do correctly. No programming knowledge necessary.

        Going back to the bad old Wintel days, MS has used manufacturer laziness to make it hard for any other OS to compete with them. So, this isn't exactly new news or particularly shocking for those paying attention. Not only were those Wintel modems less functional, but you had to pay a premium for having some of the working offloaded onto the CPU.