Hacker News new | ask | show | jobs
by zaarn 2891 days ago
This is speculation to some extend, but I don't think it'll be that easy.

Intel does have margins on their products but at the same time, the products are still expensive. I fault their production method for that. I bet that even if they sold their CPUs at a loss, AMD could undercut them in the important market segments at minimum.

AMD has struck gold with being able to, in Intel's words, glue together a bunch of desktop CPUs.

Intel's biggest CPU, a 28 core, is a monolithic die. All in one go. Which means the yields are bad. AMD produces basically 1 die and glues together as needed. The yields here are much better due to the smaller die and being able to use basically any die that has functioning 2 cores.

Intel doesn't have a response to that since if their 28 core die has 26 defect cores, they don't have a housing for it to sell in. Their segmentation also means that they don't have compatible sockets in size to even conceivable put the 28 core to use at all.

IMO Intel will have to come up with a similarly modular response or reduce their production cost.

Of course, Intel has a while, they can idle a few years on marketshare and mindshare before they run a risk of loosing the CPU business (and they sell other silicon too where AMD doesn't compete). So likely, they will be able to do it.

Question then is how much marketshare AMD will have eaten by then.

(Personally, I hope AMD will release server cpus for the lower market segment, building an Intel-based Server that doesn't have the suck is rather expensive and low-cost baremetal would be a delight for hobbyists and SMB)

4 comments

The Zen architecture was built to be able to be glued together with their "Infinity Fabric." That is their trump card. Intels architectures just aren't built with this in mind, so it will be back to the drawing board for them.
What difference does monolithic vs. modular really make? If Intel has a core on their 28 core die that is defective, they just disable it and sell a 27 core chip. Perhaps it's easier for AMD to find 4 8 core chips with no defects than for Intel to find a 28 core chip with no defects, but Intel can just make a 34 core die sold as a 32 core chip with better yields than AMD's 8 core dies.
There is a huge difference.

The number of defects is an average over an area. So if you double your die area then you get double as many defective chips.

Making a bigger chip means you throw away more dies. So Intel's 34 Core would be even more expensive to produce.

Imagine it like taking a low res photo of a paper with dots drawn on it. The bigger the pixel the bigger the black spots on the paper will be. The higher resolution your camera is, ie the smaller the pixel, the more pixels that aren't black, ie defective silicon.

AMD has the advantage here, their 4 core dies are very small so they can produce them with a high yield rate. The defect one with 1 or 2 defect cores but otherwise okay are recycled. Then everything is binned according to their performance (chips can have wildly different performance).

Another advantage Intel does not have. If one core on the 28 core chip does not manage to run stable on 2.8 GHz (example if they sold the chip at that) that means either bin it on the next lower frequency to run at or trash the die.

For AMD a low running chip means getting it into a lower bin. If they build a 32 core CPU with high core they just need to select from 4 core dies that run fast.

The difference here is very surprising, IIRC AMD stated their yield rate is 98%. For comparison, a chip like the 28 core Intel Xeon would be expected to have around 60% if not worse. And Intel can't rebin a bad Xeon for a low power desktop CPU in the lowest market segment since the die is too big for those sockets.

The original Core 2 Quad processors were dual-die, and they got a lot of negativity for being a fake quad core... Funny how things change.
My memory is cloudy on this one, but did they not use the Front Side Bus to communicate with each other and thus take a performance penalty?
That seems correct AFAIR.
If one 28 core cpu is likely to have 26 defective cores then yields are going to be bad indeed.