add Coffee Lake to the mixall Comet Lake -K SKUs were out of stock constantly for over a month
Now, let’s talk about availability. The top-of-the-range Core i7-8700K model is not available from major retailers at all. Even on auction sites, there were only two sellers to offer this chip for $800 - $900 at press time. The Core i7-8700 is more readily available from large suppliers, but the majority of online stores cannot ship it right now.
The 10900K was binned to extreme degree where it's barely available even today, but that didn't happen with all the other dies.
But the difference is that Intel is slowly moving their critical products(mobile and server) off their most profitable node. On server things aren't as clear-cut but mobile chips are definitely moving to 10nm. I see mostly Ice-Lake and Tiger-Lake in the U-laptops worth buying these days, with the Comet Lake CPUs being relegated to entry-level models except in a few models like the pro ones which require vPro which isn't available on Ice-Lake or Tiger-Lake.Really? I'm pretty sure Intel use substrates, right?
Turns out N7 isn't the only bottleneck in the industry.
And substrates were a problem before a fire at one of the factories producing them a month ago.
I did not state a word about availability. I'm simply comparing value at the suggested MSRP, stating that if by any chance the i7 Rocket Lake with 8-cores and HT is priced closer to $350, it will be a much better deal than the 5800X which costs $450. I also don't care about comparisons with the i9, as the difference between the i9 and i7 in this generation is rumored to be simply a clock speed difference of 300 MHz.Did you conveniently forget how all Comet Lake -K SKUs were out of stock constantly for over a month after launch?
The 10900K was binned to extreme degree where it's barely available even today, but that didn't happen with all the other dies.
10850K availability:The other 8 and 10 core models haven't been that available either. The 9700K and 9900K have been easier to get. Didn't help that Comet Lake-H went with the 8 core die. With no mobile version of the die coming, I would agree that DIY availability won't be great.
This.
Worth keeping in mind Intel are pulling back availability two months from March -> January. The chances that Rocket Lake will be available in with decent volume is aren't all that high.
Honestly after what happened with the 10900K I'm expecting the Rocket Lake i9 to also be practically unobtainable for a while.
Huh, well things may have gotten better. The 10700K doesn't really make sense at those prices since you can get the 9700K and 9900K for less.
This. The pressure on 14nm is going to be less this time around. Plus, Alderlake-S is just around the corner, on 10nm, so RocketLake-S is going to be mainly a limited release just to keep AMD honest for 2021.But the difference is that Intel is slowly moving their critical products(mobile and server) off their most profitable node. On server things aren't as clear-cut but mobile chips are definitely moving to 10nm. I see mostly Ice-Lake and Tiger-Lake in the U-laptops worth buying these days, with the Comet Lake CPUs being relegated to entry-level models except in a few models like the pro ones which require vPro which isn't available on Ice-Lake or Tiger-Lake.
I see Rocket Lake as Intel's main product for the gaming community, which is pretty substantial these days, but still far fewer units than corporate desktops. I expect it will be competitive with Zen3, and as such, will have value till Zen4 shows up. Alder Lake, as Intel's first large CPU on 10nm could be capacity strained for it's entire existence - we'll have to wait and see.This. The pressure on 14nm is going to be less this time around. Plus, Alderlake-S is just around the corner, on 10nm, so RocketLake-S is going to be mainly a limited release just to keep AMD honest for 2021.
I see Rocket Lake as Intel's main product for the gaming community, which is pretty substantial these days, but still far fewer units than corporate desktops. I expect it will be competitive with Zen3, and as such, will have value till Zen4 shows up. Alder Lake, as Intel's first large CPU on 10nm could be capacity strained for it's entire existence - we'll have to wait and see.
It is looking more and more likely that Intel's 7nm node will be another stopgap till 5nm (as Intel will be using TSMC for the 'Atom' line and some server CPUs). It would be good for consumers if Intel can get it together by 5nm so there is better competition - but who knows, things still look pessimistic.
Worth noting, Intel's specs for their 7nm is 15-30% more dense than TSMC 5nm. Intel is reportedly 202-220 MT/mm vs TSMC 5nm being 173 MT/mm.
Worth noting, Intel's specs for their 7nm is 15-30% more dense than TSMC 5nm. Intel is reportedly 202-220 MT/mm vs TSMC 5nm being 173 MT/mm.
Worth noting, Intel's specs for their 7nm is 15-30% more dense than TSMC 5nm. Intel is reportedly 202-220 MT/mm vs TSMC 5nm being 173 MT/mm.
Besides density, especially something as broad as "process density" is such a useless metric to describe processors. RWT has a good article on that subject (and transistor count):For context, the Intel density number came from an off-hand remark made by the CEO years ago, who has since been fired, and the process is nowhere in sight. It has since been parroted repeated by witeken (aka Arne Verheyde) on seekingalpha.com, who has zero credibility and veracity.
TSMC 5nm on the other hand is already on the market in volume.
Just what I was thinking. Intel's 10 nm, despite its relatively high density seems really bad for power consumption. Even 10 nm SF seems decent for performance, but not outstanding for power consumption. Or maybe it is just the core architecture itself that uses too much power relative to a clean sheet design like Zen.Besides density, especially something as broad as "process density" is such a useless metric to describe processors. RWT has a good article on that subject (and transistor count):
Transistor Count: A Flawed Metric
Transistor count and transistor density are often portrayed as technical achievements and milestones. Many vendors brag about the complexity of their design, as measured by transistor count. In reality, transistor count and density varies considerably based on the type of chip and especially the...www.realworldtech.com
For context, the Intel density number came from an off-hand remark made by the CEO years ago, who has since been fired, and the process is nowhere in sight. It has since been parroted repeated by witeken (aka Arne Verheyde) on seekingalpha.com, who has zero credibility and veracity.
TSMC 5nm on the other hand is already on the market in volume.
I don't have the faintest clue who you're talking about at SeekingAlpha, and don't care. Maybe you should stop reading that crap?
What you're saying is at least 50% false.
It is not just the previous CEO, it's the current one as well.
Source below :
7 nm lithography process - WikiChip
The 7 nanometer (7 nm) lithography process is a technology node semiconductor manufacturing process following the 10 nm process node. Mass production of integrated circuit fabricated using a 7 nm process began in 2018. The process technology will be phased out by leading-edge foundries by...en.wikichip.org
"the company's current CEO claims it will feature a density that is 2x that of Intel's 10-nanometer node. "
That would mean 202 MT/mm
"Intel's prior CEO, Brian Krzanich, mentioned that 7-nanometer will have "2.4x the compaction ratio" of 10 nm. "
That would be around 242 MT/mm.
Hence it would be 202- 242 MT/mm, with the upper bound actually being higher than what I posted.
Both of those numbers are far denser than anything TSMC or Samsung has said about their 5nm nodes.
This fits the pattern given that Intel's 10nm node is roughly the same density as TSMC's so-called "7nm" node.
Krzanich was a bullshit artist simply playing to the Wall Street crowd. There is a reason he was fired and it wasn't because he was playing house with another employee. 10nm couldn't hit its density targets and there is no way 7nm will be that dense when it's released. Intel can't push 7nm that aggressively after dropping the ball on 10nm (were they tried to one up TSMC on both density and performance - and failed spectacularly)."Intel's prior CEO, Brian Krzanich, mentioned that 7-nanometer will have "2.4x the compaction ratio" of 10 nm. "
Atom and Xeon SoCs, not Core and Xeon, will be fabbed at TSMC.As for 7 nm... well, it's looking increasingly likely that Intel is going to announce that they are going to fab Core & Xeon at TSMC in January. Bob may say that 7 nm is not dead, but who knows if you will ever see it.
Atom and Xeon SoCs, not Core and Xeon, will be fabbed at TSMC.
I don't think that it's poorly worded.Atom and Xeon as two categories doesn't make much sense to begin with. Inclined to believe the job posting was poorly worded.
Atom and Xeon SoCs, not Core and Xeon, will be fabbed at TSMC.
Rocket Lake would have been a nice product if it had come out 6 months ago. Now that Ryzen 5xxx is out, it just seems kind of meh.
I sure hope we get to see RKL-S with good availability and pricing, but given the current market conditions I won't be holding my breath.
The only reduction in pressure on their 14nm fabs comes from reduction in overall sales volume.