--- (---.delete@this.redheron.com) on January 27, 2023 9:31 am wrote:
> Anne O. Nymous (not.delete@this.real.address) on January 27, 2023 12:09 am wrote:
> > Chris G (chrisg.delete@this.chrisg.com) on January 26, 2023 5:02 pm wrote:
> > > I thought this quote from the NY Times article was interesting: “She also determined that Intel had
> > > scheduled more products than its engineers and customers could easily handle. So she streamlined that
> > > product road map, including pushing back a successor to Sapphire Rapids to 2024 from 2023.”
> > >
> > > In other words, Xeon Emerald Rapids is also delayed again. Emerald Rapids is just Sapphire Rapids with
> > > some more errata fixed, a 17% increase in DRAM bandwidth and a 25% increase in intersocket bandwidth
> > > on the same process technology. It could be that the complexity of x86 and the hairball of their numerous
> > > instruction set extensions is just too much for anyone to deal with. I would also guess a big part of
> > > Intel’s problem is poor management. Andy Grove was probably the best technology manager of all time
> > > but that was a long time ago. Now the type of innovation
> > > we get from Intel is renaming the process technology
> > > from Intel 4 to Intel 3 because a name with a smaller number must surely be better.
> >
> > You know that Pat Gelsinger became CEO of intel just ~ 2 years ago? It takes time to change things,
> > so maybe give the gentleman a bit more time to pull things around (well possible he might fail
> > given the "institutional inertia" of a behemoth company like intel, but he might prevail, but we
> > only know after giving the whole exercise enough time to play out one way or the other).
> >
> > Now, it looks pretty clear that intel was resting on its laurels for too long and
> > the whole 10nm process cluster-fsck certainly did not help, but not sure whether
> > all is doom and gloom, it is still a pretty rich and resourceful company...
>
> Outsiders obviously have to work with limited data; we can't prove anything,
> we can just point to patterns and say what they suggest to us.
>
> Having said that, I compared Pat with Steve Job's return to Apple, or even
> Satya's first two years. Both made it clear in their first year that it was
> no longer business as usual, that things had to be shaken up and changed.
> Conversely I don't see Pat as doing anything similar. Steve, for example, dealt with
> complexity (of a different sort, sure) by dramatically pruning the product line. But
> Pat has done nothing similar.

Intel seems to have gotten out of the memory business under Pat. NAND flash went to SK Hynix and 3DXPoint was just killed.

Pat seems to be trying to get Intel's AI/Accelerator product mix under control. At one point Intel was pushing: x86, FPGAs, Movidius, Habana, Nervana and maybe GPUs. Nervana products are dead an I think Movidius doesn't have a future.

Intel unloaded OmniPath in late 2020 and just announced they were getting out of the network switch chip business.


So I think I see Pat trying to get the product line down to something sane. Or at least manageable.

Fixing the x86 roadmap will take time, but at least we are seeing new micro-architectures with noticeable performance increases.

Next will be to see if the process cadence executes according to plan and whether Intel can successfully run a foundry business, but 24 months is a bit early to tell for either.