Discussion Intel - the cost of BACKPORTING

Page 4 - Seeking answers? Join the AnandTech community: where nearly half-a-million members share solutions and discuss the latest tech.

Kocicak

Senior member
Jan 17, 2019
982
973
136
Seeing the preliminary results in the Anandtech 11700K review, the performance of this CPU compared to the previous CPU overall seems to be just marginally better, in some cases even worse than the previous generation product.

What is the economic sense and impact on the company of the decision to backport this new CPU to the 14 nm technology? How much money does this backport actually cost? Was this cost really worth it, when the only result is that you will have some "fill in product" for a few months before the Alder lake comes?

If it does not make sense financialy, then why did they do it? They must have known well that the end result of this will be.
 
  • Like
Reactions: krumme

DrMrLordX

Lifer
Apr 27, 2000
21,582
10,785
136
Kocicak asked what this intel gen will do and what it will do is going to sell.
Nobody will not buy an intel CPU because a Ryzen CPU is better at something that the user will never use.

I'm sorry, but the entire post you made was based on blatant falsehoods that it's hardly worth picking it apart. But if you think people won't use the MT performance of chips like the 5900X and 5950X, you are completely bonkers.
 
  • Like
Reactions: Tlh97

TheELF

Diamond Member
Dec 22, 2012
3,967
720
126
I'm sorry, but the entire post you made was based on blatant falsehoods that it's hardly worth picking it apart. But if you think people won't use the MT performance of chips like the 5900X and 5950X, you are completely bonkers.
Your reply makes no sense at all.
"Huh?! How many of the people that use computers do you think actually use any sort of multithreading?!"
The people that get chips like the 5900X and 5950X are obviously covered in my statement, I asked how many that would be and the answer is in the sales of those CPUs, everybody else is not going to be sad about getting a CPU that does everything they need but doesn't do the multithreading they don't need.
 

TheELF

Diamond Member
Dec 22, 2012
3,967
720
126
Why did Intel not just stick with what they had?

Smaller die, could produce more and sell at a reduced price and still make more money... and not be out the cost of the backport.


Oh right, because it appears performance does matter. :rolleyes:
What?! You afraid that intel is not going to make their money back or something?!
They wanted to release a CPU with the much stronger Xe iGPU before releasing the discreet ones, to iron out any possible bugs.
 

ondma

Platinum Member
Mar 18, 2018
2,718
1,278
136
What?! You afraid that intel is not going to make their money back or something?!
They wanted to release a CPU with the much stronger Xe iGPU before releasing the discreet ones, to iron out any possible bugs.
They had that already with TL, and that is a stronger Xe than RL. The thing I dont understand is how they could make a chip with a 15% IPC gain and bork the cache structure so that it is basically a wash with CL. (edit: in gaming, which was really what the chip was designed to do)
 
Last edited:
  • Like
Reactions: Tlh97

Mopetar

Diamond Member
Jan 31, 2011
7,797
5,899
136
It may be a case of a sound design on a 10nm process becoming unfeasible when ported back to 14nm. Maybe we'll eventually get a deep dive into Rocket Lake that pin points what creates the biggest bottlenecks for the CPU, but it's hard to imagine that something unwanted and unexpected didn't creep in along the way. At some point you find out there's an unanticipated problem and when you look at the product schedule you have it doesn't make sense to spend additional months on a respin that may not be fully able to solve the problem.
 
  • Like
Reactions: Tlh97

DrMrLordX

Lifer
Apr 27, 2000
21,582
10,785
136
Your reply makes no sense at all.
"Huh?! How many of the people that use computers do you think actually use any sort of multithreading?!"
The people that get chips like the 5900X and 5950X are obviously covered in my statement, I asked how many that would be and the answer is in the sales of those CPUs, everybody else is not going to be sad about getting a CPU that does everything they need but doesn't do the multithreading they don't need.

EVERYBODY USES MULTITHREADING THESE DAYS. Even games want 6 cores minimum, with some showing improvements in performance on 8 cores (especially wrt 1% and .1% frame times). This is not 2006!
 

Racan

Golden Member
Sep 22, 2012
1,101
1,969
136
Dr. Cutress did this video related to this topic:

It is a success, and it means it's probably the last time Intel will take so long to bring a new architecture to market. It will be harder from now on for AMD to stay ahead imo.
 
Last edited:

Topweasel

Diamond Member
Oct 19, 2000
5,436
1,654
136
It is a success, and it means it's probably the last time Intel will take so long to bring a new architecture to market. It will be harder from now on for AMD to stay ahead imo.
Let's remember AMD has gone up 80-90% increase in IPC in the last 4 years. Also not worth forgetting that next major release is Zen 4, the third straight gen that they felt the need based on projected performance gains to be considered a new architecture.

Think Intel has more to prove on improving their incremental performance it isn't just their 5 years of skylake. Till rocket IPC hadn't seen any drastic increase since Sandy Bridge. 10 years.
 

Racan

Golden Member
Sep 22, 2012
1,101
1,969
136
Think Intel has more to prove on improving their incremental performance it isn't just their 5 years of skylake. Till rocket IPC hadn't seen any drastic increase since Sandy Bridge. 10 years.

Well, Alder lake and Meteor lake are also supposed to be drastic improvements, we'll see how it goes.
 

Topweasel

Diamond Member
Oct 19, 2000
5,436
1,654
136
Yeah I am not debating whether or not Intel has the development capability to keep up and succeed. But to many times people put it on AMD to somehow match Intel. Nah it's the other way around. Intel has a lot to prove going forward. As far Zen is concerned AMD has been hitting on all cylinders and it's easier to see now AMD keeping that up then Intel changing their ways aft over of a decade of barely moving the arrow.
 

Atari2600

Golden Member
Nov 22, 2016
1,409
1,655
136
They wanted to release a CPU with the much stronger Xe iGPU before releasing the discreet ones, to iron out any possible bugs.

Yes... because that is how billion dollar turnover companies work.

They release products to the general public to iron out any possible bugs.

They don't do it in R&D, they do it in your PC.

Listen to yourself. Talk about flailing around desperately trying to grab hold of any excuse, no matter how flimsy. :rolleyes:
 
  • Like
Reactions: Tlh97 and MangoX

dmens

Platinum Member
Mar 18, 2005
2,271
917
136
Dr. Cutress did this video related to this topic:

Nope, you do not learn about designing to multiple processes by moving a design back to an in-house exclusive process with a tool stack that has been used repeatedly for years. That is not co-design, that is just brute-forcing an existing HDL database to your own single customer process.

It is like saying you learned something new by redoing something you did last week.
 

IntelUser2000

Elite Member
Oct 14, 2003
8,686
3,785
136
It is a success, and it means it's probably the last time Intel will take so long to bring a new architecture to market. It will be harder from now on for AMD to stay ahead imo.

Actually, Intel has tremendous difficulty in making CPUs that are consistently better than previous generations. History tells us.

Pentium III? Pentium II with SSE instructions added. Pentium 4? 50% higher clocks but barely outperformed the 1GHz Pentium III.

Once in a blue moon they make Banias, and Core chips, but in between it's all minor changes and rehashes.
 

DrMrLordX

Lifer
Apr 27, 2000
21,582
10,785
136
Saying Rocket Lake-S is a success is like saying, "hey, Intel didn't blow their brains out today". Judge for yourself based on the performance of the chip whether you'd want it over an existing (and likely discounted) Comet Lake product. But don't say it's a success just because Intel managed to release some kind of product as opposed to releasing, well, nothing? At all?
 

Ian Cutress

Junior Member
May 10, 2014
2
34
91
Nope, you do not learn about designing to multiple processes by moving a design back to an in-house exclusive process with a tool stack that has been used repeatedly for years. That is not co-design, that is just brute-forcing an existing HDL database to your own single customer process.

It is like saying you learned something new by redoing something you did last week.

Pretty sure you misunderstood the video. The point of the video is to say that because this processor *wasn't* co-designed, the future processors will be co-designed. I specifically went through the shoehorning in process in the video to showcase the difference of designing a chip focusing one PDK and porting it to a second, vs starting with two known PDKs.
 

dmens

Platinum Member
Mar 18, 2005
2,271
917
136
Pretty sure you misunderstood the video. The point of the video is to say that because this processor *wasn't* co-designed, the future processors will be co-designed. I specifically went through the shoehorning in process in the video to showcase the difference of designing a chip focusing one PDK and porting it to a second, vs starting with two known PDKs.

If you are referring to Intel designing big cores to external processes, that will never happen until the fabs are spun off and Intel is no longer an IDM.

If you are referring to Intel adopting design methodologies that ease process movement, it would take a lot more than one disaster to change Intel worship of in-house, proprietary tooling and methods. Many political empires within Intel depend on not changing and not innovating.
 

IntelUser2000

Elite Member
Oct 14, 2003
8,686
3,785
136
If you are referring to Intel adopting design methodologies that ease process movement, it would take a lot more than one disaster to change Intel worship of in-house, proprietary tooling and methods. Many political empires within Intel depend on not changing and not innovating.

I agree with this. But also because new processes give so many benefits. I don't believe they'll fully optimize for older processes either.

Goldmont/Goldmont Plus is an example of what properly designed for 14nm can do for you. But it was only because they were purposely relegating those chips to an older processes. I don't see that happening for Core.

Think Intel has more to prove on improving their incremental performance it isn't just their 5 years of skylake. Till rocket IPC hadn't seen any drastic increase since Sandy Bridge. 10 years.

Sort of agree on Haswell, but that was a great CPU for mobile.

Also Skylake does quite a bit better on high memory speeds. The newer memory standards need to be a speedgrade higher to be equal to the older generation. So DDR2-533 = DDR-400, DDR3-1066 = DDR2-800, etc.

Skylake's HD 530 was actually a nice advancement over the one in 4790K. Yea sure it wasn't GT3e in 5775C, but a solid 25% faster
 

Exist50

Platinum Member
Aug 18, 2016
2,445
3,043
136
If you are referring to Intel designing big cores to external processes, that will never happen until the fabs are spun off and Intel is no longer an IDM.

If you are referring to Intel adopting design methodologies that ease process movement, it would take a lot more than one disaster to change Intel worship of in-house, proprietary tooling and methods. Many political empires within Intel depend on not changing and not innovating.

Simply wrong.
 

dmens

Platinum Member
Mar 18, 2005
2,271
917
136
Goldmont/Goldmont Plus is an example of what properly designed for 14nm can do for you. But it was only because they were purposely relegating those chips to an older processes. I don't see that happening for Core.

This is because Intel Austin operates at the maximum amount of wrongthink that is allowable within the company.
 
  • Like
Reactions: lightmanek

IntelUser2000

Elite Member
Oct 14, 2003
8,686
3,785
136
This is because Intel Austin operates at the maximum amount of wrongthink that is allowable within the company.

I'm saying, they need to prioritize on moving to a new process, because that's still better than being on the older one.

I noticed on a GPU review that the graphs were separated into two groups - those on a newer process, and those on a older process. The real wrongthink happened was what made them screw up on 14nm and 10nm.
 

TheELF

Diamond Member
Dec 22, 2012
3,967
720
126
EVERYBODY USES MULTITHREADING THESE DAYS. Even games want 6 cores minimum, with some showing improvements in performance on 8 cores (especially wrt 1% and .1% frame times). This is not 2006!
Everybody leaves apps open, that is a far cry from multithreading/tasking, all you need is lots of ram and a fast ssd.
The amount of multitasking normal people are doing can be done on a dual core with smt.
Games that see any sort of improvement on 8 cores also need a $1000 (normally not now) GPU, that's luxury and not a necessity. Going from 80FPS to 100FPS 1% isn't going to be noticeable by anybody.
 

coercitiv

Diamond Member
Jan 24, 2014
6,151
11,681
136
The amount of multitasking normal people are doing can be done on a dual core with smt.
A modern browser alone can make good use of a quad core with multi-threading. You have a very poor understanding on what type of resources people can end up using just doing office work at home, let alone in professional environments.
 

Thibsie

Senior member
Apr 25, 2017
727
752
136
Everybody leaves apps open, that is a far cry from multithreading/tasking, all you need is lots of ram and a fast ssd.
The amount of multitasking normal people are doing can be done on a dual core with smt.
Games that see any sort of improvement on 8 cores also need a $1000 (normally not now) GPU, that's luxury and not a necessity. Going from 80FPS to 100FPS 1% isn't going to be noticeable by anybody.

Did you read what you were replying to ?
Not that DrMrLordX was necessarily right.
But you don't seem to even read before replying.