Intel CEO Pat Gelsinger Retires

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts
  • mrg666
    Senior Member
    • Mar 2023
    • 1096

    #21
    Intel had a principle called "copy exactly" for design at the time Paul Otellini was appointed as the CEO. The principle meant "don't innovate unnecessarily" and always save cost. Combine this mentality with a marketing person like Otellini (and others after him) with limited technical knowledge who preferred making quick bucks without too much investment, Intel fell behind process technology, adopted copper interconnects last, avoided using EUV lithography until starting to miss ticks and tocks, missed all of the opportunities in the mobile technologies. They were trying to sell those stupid webcams at the time while trying to strongarm the market with Itanium. They lost a big share of even their key server market to AMD. The largest semiconductor manufacturer in the world lost the crown to TSMC.

    With all these missteps, any other company would have been bankrupt already. Their last two decades is a very sad story with a lot of lessons to learn. But, I believe, they have such a legendary history prior to this decay period, Intel can still be revived with a capable leadership. I had high hopes for Gelsinger for that reason ... disappointed.

    Comment

    • L_A_G
      Senior Member
      • Oct 2015
      • 1612

      #22
      Admittedly I haven't been paying especially close attention to what's been going on inside Intel, but from what I've heard its the usual "palliative capitalism" story.

      A relentless drive for profitability, causing them to cut down as much as possible on costs and especially investments. Driving short term profitability while consciously short changing the long term. Which obviously isn't great when you're in a market as capital-investment heavy as semiconductor manufacturing. Semiconductor design isn't exactly a low investment industry either, but on the manufacturing side the chickens come home to roost relatively quickly and that's an area where they'd be genuinely cooked if not for the CHIPS act.

      What most people refer to as "Moore's Law" ("The number of transistors on a leading edge semiconductor product doubles every 18 months") is actually called "Moore's First Law" and there is a second. It states "The cost of the factory required to produce those products doubles every 48 months" and unlike the first, it still applies. In a drive for short term profitability they didn't invest in new factories or trying to make foundry services a large business (you need BIG volume to make a 10 billion dollar factory profitable) until it was too late.

      In trying to get the foundry services going it obviously didn't help that Intel's problems getting the 10nm yields up causes some major issues for their long time undisclosed foundry services customer, Nokia. They badly lost out in early 5G deployments when their ASIC-based "Reefshark" products couldn't be produced in volume and had to rely on more expensive and lower performance FPGA-based products while competitors like Huawei and Ericsson were selling ASIC ones left right and center.
      "Why should I want to make anything up? Life's bad enough as it is without wanting to invent any more of it."

      Comment

      • r1348
        Senior Member
        • Jul 2007
        • 636

        #23
        November is stock vesting time...
        Even Tavares just resigned from Stellantis, the timing is not a coincidence.

        Comment

        • DumbFsck
          Senior Member
          • Dec 2023
          • 345

          #24
          I had high hopes for Gelsinger and honestly I'm a bit shocked, I think taking him out is if not wrong, maybe too early.


          5 years is barely a development cycle for CPU architecture, and from my (very external) POV he made a lot of good moves (some of which may not have paid out, although I still have them as good move because of positive EV). Opening their fabs to third party clients is a good move, focus on technological leadership is a good move, making the company leaner and possibly less bureaucratic is a good move. Products that came out under him are also acceptable, their foveros is good, their npus I actually think are the best designed (above apples, Qualcomm, AMD), and I don't know how long after or before him joining these projects were spearheaded.


          I understand that market pressure has to come to the top, and if the stock is not well the CEO has to suffer, but I wonder how much of the stock movement is because of Pat vs every other thing inside Intel.


          All I hope is that the next CEO isn't some liquidator brought in to make the intel plane crash less gruesome.

          Comment

          • DumbFsck
            Senior Member
            • Dec 2023
            • 345

            #25
            Originally posted by CommunityMember View Post

            Of course he was.

            There have been rumors for quite some time that the board and Gelsinger had some serious disagreements in strategy, and three months ago a senior board member resigned over such disagreements.

            At some point the board is forced to assert their authority. That point is now.
            Do you have some sources I can read more on? This sounds really interesting to me

            Comment

            • hwertz
              Phoronix Member
              • Apr 2008
              • 97

              #26
              Originally posted by botipua22 View Post
              i'm sure nvidia would love to absorb intel, but their attempt will certainly be blocked like it was during their pursuit of arm. intel is an american darling.
              Awkward part is, I don't know if they'd care. They uise embedded Risc-V in their GPUs and AI accelerators (for the tiny portion that isn't just EUs/CUs... in olden terms vector processors). And Arm instruction set on their embedded and "deskside superconputer" type prodcuts. So they give 0 cares about x86 instruction set, and they are fabless so (even if Intel's fabs were up to par with TSMC) they wouldn't necessarily be interested. I mean that said, given their relative market caps, buying them anyway wouldn't be a bad idea at all.

              Comment

              • hwertz
                Phoronix Member
                • Apr 2008
                • 97

                #27
                Originally posted by Weasel View Post
                No one cares because it's bullshit. That's the reason.

                But you will continue spewing that forever no matter what so keep burying your head in the s̵a̵n̵d̵ Apple's ass.
                I'm no Apple fan at all but the M series chips do indeed clean Intel and Arm's clocks; you can find an Intel or AMD CPU that is roughly as low power but it's a wheezy (in Intel's case..) Atom based 'all E-core' setup to get close; and find other models that can keep up performancewise but just chug throigh power. This isn't some Apple magic, it's just how Arms can be.

                I had Ububtu on a Nvidia Tegra K1 based Chromebook that got 22 hours battery life under typical use, and 11 hours pegging out all 4 cores video encoding (which it did faster than my quad core desktop I had at the time.). That was really a 5-core. it had alittle 1ghz core it'd kick over to on below 600mhz or so of system demand, it was like the kind you'd run off a watch battery but plenty to keep a cursor blinking and whatever 'the computer is almost but not quite idle' type stuff going (it was fast enough to play most videos on that core. )

                Frankly I'm curious on how long a Qualcomm ARM system could run when it's got a good OS on it rather than being saddled with Windows, given the very good performance and battery life I saw the last time I went down that route.

                Comment

                • bernstein
                  Junior Member
                  • Feb 2011
                  • 16

                  #28
                  Originally posted by avis View Post
                  People can laugh at me all they want, but when a fruit cult company has a much faster and more efficient uArch (M4 Pro destroys them while consuming much less power) than both Intel and AMD, it should be quite alarming, but for some reason no one cares.​
                  Originally posted by roviq View Post
                  I will not laugh at you, I find it very strange that Qualcomm can't do the same thing with ARM for years now, and neither does AMD nor Intel. But also the fruit cult company does handle everything end-to-end so they may micro-optimize at all levels, hardware, firmware, compiler, libraries, apps.
                  I'm always astonished at how people think, they can compete with the current apple. Over the last decade apple has stomped over the competition, no one has been able to compete. Apple outdid Intel at their core business (with the help of TSMC), Apple outdid ARM (by a huge margin), Apple outdid AMD, Apple outdid Qualcomm. Not to mention Apple completely stomped over Google, Microsoft and every other Mobile Phone Manufacturer on the Planet in the case of Smartphone revenue. So much so that Microsoft folded it's Mobile OS. Nokia even sold it's entire smartphone business.
                  The only part where Apple has struggled is in beating Qualcomm in cellular modems, notably after buying intel's employees. Oh and Apple so far has refrained from going head to head with Nvidia with a discrete gpu - but perf/watt of its igpu hint's that apple may be able to outdo them, if it wanted.

                  Whats troubling for intel (& amd) is that qualcomm was able to beat them on efficiency. On the other hand, that's the team that designed apple's current crop of SoC's...

                  Comment

                  • rmfx
                    Senior Member
                    • Jan 2019
                    • 764

                    #29
                    He retires... he was retired.

                    I understand he got the company is a bad shape, but things have not improved since... despite a humongous gigantic salary while he was firing the staff.

                    I think Intel should think post x86, because when Qualcomm and MediaTek both make M4-like chips, x86 is dead. That will become a transition architecture and the world will switch to ARM/RISC-V, and Intel will not have a lower market share, it will have none.

                    Comment

                    • bernstein
                      Junior Member
                      • Feb 2011
                      • 16

                      #30
                      Originally posted by geerge View Post
                      From what I can tell he made the right moves, it's just the hole that had already been dug was too deep for a quick turnaround.

                      Apple has two things going for it. 1) They buy out the best node years in advance. 2) They go for efficiency above all else; large die area, many accelerators, clocked at the sweet spot in the efficiency curve. Both of these things are possible because of how they've positioned themselves in the market. Everyone else is competing with the additional constraint of cost, in what boils down to a very different market with some overlap. On the same node with the same design goals everyone competent would be roughly at parity regardless of arch, there's nothing alarming about the situation.
                      Exactly

                      Comment

                      Working...
                      X