Intel Expands 10nm Manufacturing Capacity

No, no, no Intel....

Yeah, we've heard that before. Years. Time and time again.

 

All reports are that 10nm is actually "fixed" and able to be produced in volume.

 

good news for Intel.
but as Intel's 10nm equals everyone else's 7nm, they're still a node behind

 

That doesn't mean anything because Intel can use everyone else as well. Their options is actually more like everyone else +1.

 

Intel density is higher on 10nm than TSMC on 7nm, so they are actually ahead if they manage to fix it


didn't research if this is true tho, just heard it + googled it

 

We make fun of Intel's 14nm and all the +++ they add, but, it has very comparable performance to TSCM's 7nm. If Intel can pull off the same sort of refinement on 10nm, it will be better than we expect. The problem is, they're 3 years late.

 

While trashing on Intel is trending now, we always need competition in the market. And Intel can actually provide enough supply, so I do hope they can come up with good products soon.

 

That video....omfg....I can't even....

 

Don´t tell me you don´t believe the guys from Intel just because they lied a few times about this...

For me that´s not even the best part. Their 14nm process is so refined by now that using it is basically like printing money. The 14nm process must be the most profitable in the history of semiconductors by far!

The only "downside" is that a big chunk of the profits are being burned on their failed 10nm process...

 

It's even more profitable when you consider how little engineering Intel has done to their 14nm CPUs. AVX512 the most work they've done, and although it isn't minimal, it didn't come with an architecture refresh.
Even during Intel's worst years, they still rake in more cash than they know what to do with. Their net profits are stupidly huge, so although they have burned a lot on 10nm, it's not really much of a loss.

 

Exactly. It would take years of blunders and mismanagement to topple Chipzilla. Like any other corporate behemoth they take a while to change course, but it won't take them long to catch up and overtake again. That said, it's fun to take pot shots at them, so let's enjoy it while we can.

 

If we're comparing nodes, then we're not comparing actual performance as a LOT of it is tied to the architecture, as we've seen recently. Intel has only made gains of a few hundred Mhz in recent years, yet they've made much larger gains in overall performance. The nodes aren't comparable, but the different approaches (monolithic vs chiplets, low latency vs more cache, etc) of both companies have resulted in similarly performing products, at least in the metric of absolute performance.

The Radeon 7 is a great example of how a great node can underperform massively and why nodes don't mean anything by themselves. They can point to what's possible, but I don't think saying Intel's 14nm performs similarly to 7nm is objectively true.

 

It's not bad news. *and this is public knowledge* nothing which isn't already known... these terms "nm" etc not in all cases but have often been obfuscated by marketing departments, the finer points - in essence, AMD's (Qualcomm, GoFlo etc, whom ever for that matter) hmm you know what one moment I shall link a known good source instead such as a news article...

• 10nm, 7nm, 5nm…. Should the Chip Nanometer Metric Be Replaced? (hpcwire.com)

• Do nanometers matter? (fudzilla.com)

• What Do “7nm” and “10nm” Mean for CPUs, and Why Do They Matter? (howtogeek.com)

• Intel: The Nanometer Games (NASDAQ:INTC) | Seeking Alpha

• Intel 14 nm Node Compared to TSMC's 7 nm Node Using Scanning Electron Microscope | TechPowerUp

"TSMC's node is still much denser compared to Intel's 14nm - TSMC's 7 nm produces chips with a transistor density around 90 MT/mm² (million transistors per square millimeter), which is comparable in density to Intel's 10 nm node used on recent mobile processors."

 

there is a big upside to intel's 14nm+

Spoiler: gaming

and a big downside to still using 14nm on high frequency +6 core cpus

Spoiler: efficiency

all in all,these 14nm cpus are still a good buy for gaming/office/light workstation use provided the price is right.10900F costs a lot less than 5800X and in most cases you'll be getting a faster cpu provided you can deal with such a power hog for your home build.
imo they're doing better against r5000 than they did against r3000 due to zen 3 low availability and high prices.

of course we would rather have seen new arch and node from intel like 3 years ago,but it is what it is
10850K sells for 2000pln here,10900F for 1800,10700F for 1300pln and 5600x/5800x for 1600/2200 respectively.it would be an easy decision in terms of value for me.

 

Performance per core intel does fine. Power consumption is where they loose by a huge margin obviously 10nm will help em with consumption big time. Now the 10nm for intel is really late to the party , what remains to be seen is how fast intel can nail their 7 nm , did they get valuable knowledge from all the problems of 10nm ? Or scaling down to 7 is a beast of it's own ?

 

What I've learned from this video is that there's too many people that does not know how to use a face mask.
Is that why 10nm is having issues?

 

I do wonder how true it is considering intel's line up on the consumer side remains mostly 14nm for the time being.

The lack of a current 6-8 core tiger lake just adds to the question. If its bugs have been ironed out, I would have hoped that an 8 core cpu would be a feasable product.

 

doubling what could have been 10-20% of manufacturing capacity they need is still not gonna cut it.

 

Intel has massive capacity, period.
it does not need to do anything more than keep marketing out of engineering (source of 1/3 their issues), and by marketing i mean the executive suite which is overflowing with marketing types.
the Hillsboro, OR 10nm fab is twice the size of the original TSMC 7nm fab, and they have two other 10nm fabs (another in US, one in Israel).

they will catch up relatively fast on clock speeds and ipc, power draw is and will continue to be a weakness until they go to a sub 10nm node or new uArch or both.
what gets lost is the fact that the 10nm mobile cpu has been a huge success, defined by numbers shipped and laptops on the market. i'm writing this on one now.