Intel launching H270 rebrand called B365 due to 14nm production problems

[pas008]

42 minutes ago, samcool55 said:

Yes and no, Intel switched to 10nm assuming they could use it to mass-produce cpu's by now. But they can't.

Changing node takes months so it has to be done in advance and it's not worth to switch those 10/7nm fabs back to 14nm and then back to 10nm again because that would cost a lot of time and money. So for now the 10nm fabs mainly are barely producing anything because they basically can't produce decent cpu's with it.

 

They produce cpu's, don't get me wrong, but they aren't faster than their 14nm bothers (i think slighty slower like a few% ?) and also use more power (i think) so they are basically worse than what's already available. Also yields are probably quite trash as well because it's a new node and it takes years to master.

yes and no?

what? did you just state pretty much what I said but in detail?

 

[pas008]

31 minutes ago, AnonymousGuy said:

 

 

Exact same boat.  I don't have a single USB 3.0 device in my house, let alone USB 3.1.  Storage?  I have a server and 10Gbit, don't need external drives.  Everything else that might use that bandwidth I have running off PCIe slots.   If anything I want more USB 2.0 ports and headers and shit on my Z390 motherboard instead of 3.1, because that's what I actually need.

 

 

If by not looking good you mean "selling every wafer that is produced at 2x the profit margin that AMD has" then yeah things are just awful for Intel.  This isn't a capacity decrease problem, it's a demand increase problem.

dont think people really understand, they produced enough for their forecasts not an elevated demand

 

[Drak3]

10 hours ago, firelighter487 said:

so now they're selling a year old stuff as new? 

That's precisely what everything post Skylake is.

[Guest]

12 hours ago, NelizMastr said:

This is a case of desperation. What's next? Rebranded X58 for HEDT?  

Rebranded G41 for 'Business and light use'. Rebrand Core 2 Duo for Intel Atom. 

[Canada EH]

13 hours ago, NelizMastr said:

So now we have a Kaby lake PCH on a Coffee Lake chipset

no worries, they will have it all worked out in 10 years time!

[orbitalbuzzsaw]

12 hours ago, NelizMastr said:

Which makes me even more curious about AMD's unveiling of Ryzen 3/ Zen2 at CES. If the rumors are even remotely true, it's really going to be

Athlon 64 all over again.

Rumored $100 6c12t 4.5-5GhZ Ryzen 3 3200...

[BuckGup]

Oh my god how can anyone follow this. Just buy Ryzen because you actually understand what you are buying lol

[Fnige]

AMD sitting there just like: You're still more popular than me but damn, yo got problems up there

[Fnige]

16 hours ago, NelizMastr said:

Athlon 64 all over again

Also, Whatcha mean by that

Edited December 15, 2018 by SlimyPython
Rewordation!!

[Crunchy Dragon]

16 hours ago, NelizMastr said:

Rebranded X58 for HEDT?  

I wouldn't exactly say no to that, X58 is one of my favorite platforms.

[ZcanKal]

Wow, I see no similarities at all! (FX, Pentium 4, Pentium D)

[ZcanKal]

3 hours ago, RorzNZ said:

Rebranded G41 for 'Business and light use'. Rebrand Core 2 Duo for Intel Atom. 

Honestly I kinda want to see what a Core 2 Duo would look like on 14nm/22nm.

[Crunchy Dragon]

14 minutes ago, ZcanKal said:

Wow, I see no similarities at all! (FX, Pentium 4, Pentium D)

What was FX a rebrand of?

 

The only thing I can think of that comes close is the Phenoms, but those were quite a bit worse than the FX chips.

Wasn't the Pentium D just two Pentium 4 dies duct taped together?

[captain_to_fire]

23 minutes ago, ZcanKal said:

Wow, I see no similarities at all! (FX, Pentium 4, Pentium D)

Uhm Pentium D is two Pentium 4’s that are glued together 

[Dabombinable]

23 minutes ago, ZcanKal said:

Honestly I kinda want to see what a Core 2 Duo would look like on 14nm/22nm.

Or perhaps something from one of its predecessors. I'd so love to have dual PIII that actually ran stable at 1.33GHz.

[TheDankKoosh]

9 minutes ago, Crunchy Dragon said:

What was FX a rebrand of?

 

The only thing I can think of that comes close is the Phenoms, but those were quite a bit worse than the FX chips.

Wasn't the Pentium D just two Pentium 4 dies duct taped together?

Wut, phenom 2 was leagues better than fx, I wish they had just made phenom 3 on that beefy af 32nm with a 10% ipc bump.

[ZcanKal]

1 hour ago, captain_to_fire said:

Uhm Pentium D is two Pentium 4’s that are glued together 

 

1 hour ago, Crunchy Dragon said:

What was FX a rebrand of?

 

The only thing I can think of that comes close is the Phenoms, but those were quite a bit worse than the FX chips.

Wasn't the Pentium D just two Pentium 4 dies duct taped together?

FX was AMD's (and currently Intel's) bad strategy of just strapping more cores on.
The Pentium D I'm pretty sure was actually Pentium 3 dies if I'm not mistaken?

[ZcanKal]

1 hour ago, Dabombinable said:

Or perhaps something from one of its predecessors. I'd so love to have dual PIII that actually ran stable at 1.33GHz.

I just want a new line of ancient CPU's done on the newest node.

[samcool55]

10 hours ago, pas008 said:

yes and no?

what? did you just state pretty much what I said but in detail?

 

I will try to explain it again.

 

"i know they have other fabs up for 10/7 which is why they werent able to meet demand"

Yes they have indeed fabs up for 10/7nm, most of these used to be 14nm or something older, but the lack of cpu's isn't because of the lack of fabs, they have plenty of fabs they can use right now.

 

"if you look at the shortages is becuz they didnt forecast as big as a demand while they were transitioning."

The demand didn't really change, their forecast for cpu's is fine and the transition at least for the fabs itself also went fine.

So, what you say about the fabs is correct, they do own 10/7nm fabs. But the reason why there is a shortage is incorrect.

 

The problem is that Intel right now has a bunch of 10/7nm fabs, mostly 10nm but are unable to actually use them for mass production. The fabs work but they keep failing in actually making them produce something useful.

Intel guessed that they would be able to use those 10nm fabs for mass production for a while actually, but they still can't. Problem is because transitioning takes months, a bunch of fabs were transitioned to 10nm before they were sure they could produce useful products for them. The transition is partially done, but they cannot use them.

 

There are cpu's out there produced by those 10nm fabs right now, but they are worse than their 14nm cpu's and a company as Intel cannot afford to launch a new generation of cpu's and say "they are worse than what we had!" because that would be a disaster.

 

So basically, they cannot use the 10nm fabs which are ready to mass-produce cpu's because they cannot make cpu's with it that are better than what they already have.

 

What they are doing now is using older fabs (that were probably planned to be transitioned to 10/7nm or to be phased out) and put them back into production as is (so in this case 22nm) so more 14nm fabs can produce cpu's because you can't just change the node of a cpu, but for a chipset it's fine.

 

 

 

 

[Ashleyyyy]

10 hours ago, Drak3 said:

That's precisely what everything post Skylake is.

i suppose... the only real issue now with kaby lake and older is 4 core's... 

[mr moose]

1 hour ago, samcool55 said:

I will try to explain it again.

 

"i know they have other fabs up for 10/7 which is why they werent able to meet demand"

Yes they have indeed fabs up for 10/7nm, most of these used to be 14nm or something older, but the lack of cpu's isn't because of the lack of fabs, they have plenty of fabs they can use right now.

 

"if you look at the shortages is becuz they didnt forecast as big as a demand while they were transitioning."

The demand didn't really change, their forecast for cpu's is fine and the transition at least for the fabs itself also went fine.

So, what you say about the fabs is correct, they do own 10/7nm fabs. But the reason why there is a shortage is incorrect.

 

The problem is that Intel right now has a bunch of 10/7nm fabs, mostly 10nm but are unable to actually use them for mass production. The fabs work but they keep failing in actually making them produce something useful.

Intel guessed that they would be able to use those 10nm fabs for mass production for a while actually, but they still can't. Problem is because transitioning takes months, a bunch of fabs were transitioned to 10nm before they were sure they could produce useful products for them. The transition is partially done, but they cannot use them.

 

There are cpu's out there produced by those 10nm fabs right now, but they are worse than their 14nm cpu's and a company as Intel cannot afford to launch a new generation of cpu's and say "they are worse than what we had!" because that would be a disaster.

 

So basically, they cannot use the 10nm fabs which are ready to mass-produce cpu's because they cannot make cpu's with it that are better than what they already have.

 

What they are doing now is using older fabs (that were probably planned to be transitioned to 10/7nm or to be phased out) and put them back into production as is (so in this case 22nm) so more 14nm fabs can produce cpu's because you can't just change the node of a cpu, but for a chipset it's fine.

 

 

 

 

It sounds to me like you are saying the same thing.  You're just putting a lot more detail into it.

 

 

The more important question here is why do some people think this is AMD doing?  Do people still think Intel sat on their hands?  No company does that.

[samcool55]

32 minutes ago, mr moose said:

It sounds to me like you are saying the same thing.  You're just putting a lot more detail into it.

 

 

The more important question here is why do some people think this is AMD doing?  Do people still think Intel sat on their hands?  No company does that.

It sounds the same because it is the same, just to be sure i'm being clear and other people understand what i mean.

 

I personally wouldn't say this is all because of AMD, but i do think it's thanks to AMD the problems are bigger than they would without them.

 

The cpu's with more cores come at a cost, not only actual cost but also less amount of cpu's per wafer because it's physically a larger thing so it's normal that they produce less cpu's atm. Without AMD I would say intel 6-core cpu's would be realistic, but 8-core cpu's so soon? That's because of AMD.

 

The big problem is their 10nm node that they just can't get working right and that has nothing to do with AMD, but it does add pressure to get it sorted quickly because it's getting a bigger and bigger problem.

[RejZoR]

Chipsets being on older nodes are not all that uncommon or weird. After all, you need stable operation, not peak possible performance. And power consumption differences are minimal as well. Chipset using 8W or 10W is irrelevant. CPU using 95W or 130W or even more is quite significant.

[pas008]

9 hours ago, samcool55 said:

I will try to explain it again.

 

"i know they have other fabs up for 10/7 which is why they werent able to meet demand"

Yes they have indeed fabs up for 10/7nm, most of these used to be 14nm or something older, but the lack of cpu's isn't because of the lack of fabs, they have plenty of fabs they can use right now.

 

"if you look at the shortages is becuz they didnt forecast as big as a demand while they were transitioning."

The demand didn't really change, their forecast for cpu's is fine and the transition at least for the fabs itself also went fine.

So, what you say about the fabs is correct, they do own 10/7nm fabs. But the reason why there is a shortage is incorrect.

 

The problem is that Intel right now has a bunch of 10/7nm fabs, mostly 10nm but are unable to actually use them for mass production. The fabs work but they keep failing in actually making them produce something useful.

Intel guessed that they would be able to use those 10nm fabs for mass production for a while actually, but they still can't. Problem is because transitioning takes months, a bunch of fabs were transitioned to 10nm before they were sure they could produce useful products for them. The transition is partially done, but they cannot use them.

 

There are cpu's out there produced by those 10nm fabs right now, but they are worse than their 14nm cpu's and a company as Intel cannot afford to launch a new generation of cpu's and say "they are worse than what we had!" because that would be a disaster.

 

So basically, they cannot use the 10nm fabs which are ready to mass-produce cpu's because they cannot make cpu's with it that are better than what they already have.

 

What they are doing now is using older fabs (that were probably planned to be transitioned to 10/7nm or to be phased out) and put them back into production as is (so in this case 22nm) so more 14nm fabs can produce cpu's because you can't just change the node of a cpu, but for a chipset it's fine.

 

 

 

 

Again you are just starting same shit by getting into detail but are assuming shit and trying to get into details

You know Intel makes shit load of other products that take up fabs

 

[samcool55]

1 hour ago, pas008 said:

Again you are just starting same shit by getting into detail but are assuming shit and trying to get into details

You know Intel makes shit load of other products that take up fabs

 

Like what do they produce?

Atom cpu's are REALLY small so you can get a LOT of them out of a single wafer, don't need much for that.

Apart from that, what do they produce on the 14nm node? Can't think of much apart from cpu's and chipsets atm (and they are moving away chipsets from 14nm)

 

And 10nm cpu's exist, like you can buy them so there are reviews of them out there

https://ark.intel.com/products/136863/Intel-Core-i3-8121U-Processor-4M-Cache-up-to-3-20-GHz-

https://www.zhihu.com/question/277156338

If you want to know why it's in Chinese, it's because they were originally launched in products for China only. These days they are also available in NUC's that are sold in the US but no reviews of it so far. (at least i couldn't find them)