Intel Core i9-9900k 8c/16t, i7-9700K 8c/8t, i7-9600k 6c/6t 2nd Gen Coffee Lake CPU's + Z390

Just meant that i did not have the chip long before delidding so did not really do much testing. I was running stock at 1.21 Vcore before and after the delid (I should have noted). I was also working on a 5.0 overclock but did not have it completely stable yet so the tests i ran would not be completely accurate. Seemed to be around 8C difference from no delid to delid with copper IHS. I have no tests with stock IHS vs Copper but Linus Tech Tips test he saw about a 2C drop from stock IHS to copper IHS. I would assume i saw about a 6C drop from the delid with CLP and 2C with stock IHS to copper.

 

Ahh.. 8C difference. Gotcha.

 

I know safe voltage recommendation for 8 cores is lower (1.35v by der8auer). But for 4 Core i7 as yours...

Put it this way... Kaby lake, almost same chips as Skylake is recommended from der8auer with up to maximum core voltage at (safe 1.4 volts). CaseKing sell their binned chips with own 24-month King-Mod-Guarantee. I wonder if the buyers have a case if their chips degrade in same way as yours the first 1 or within 2 years. Aka need more voltage than they are advertice with for specified clock speed. Or you have to downclock. Why advertice it with their 24 month guarantee if they know the chips will run into same degredation problems as you have seen. I don't think your case is the normal.


But of course we have to expect that chips can have some variances. Some more fragile than others.

 

I got lucky with my 9900K it seems. Reading other posts it seems mine does pretty low voltage 5.1Ghz all cores with 0 offset. 5.1Ghz at 1.252v load with LLC7 (max 8 on my maximus XI). I think the BIOS is 1.3v but vdroop is real on this board. Temps are also in check.

My daily driver is 5.0Ghz all cores with 1.27v BIOS and 1.217v load. I'm happy with that as it produces excellent results, low temps, and will likely never degrade my chip.

 

Cheers, I hadn't realised that the safe recommendation for 8 core was 'only' 1.35V, I assumed it would be the same as the 4 core Skylake CPUs due to it being manufactured on the same process.

Perhaps my CPU with it's small amount of degredation is an outlier, but it does tally with some degredation tests that one user did with a number of Kabylake CPUs, here's my original post where I describe the situation & give a link to my 'research': http://forum.notebookreview.com/thr...ers-welcome-too.810490/page-296#post-10766430. My CPU only degraded by 8mv over a 1yr period of running at 1.4V, and I was surprised to see that because CPU was never above 65 degC max, normally the 50's when gaming. I thought about leaving it at 4.7Ghz and increasing the voltage, but realised that would just accelerate the degradation. Instead I decided to lower the voltage & lower the overclock - and then a couple of months ago I fine tuned it by using BCLK overclocking to gain 4.65Ghz (@1.37V), which had an added bonus of bumping up the RAM overclock by a fraction, and my RAM is not stable at the next multiplier step, so this fine tuned the RAM overclock at the same time. As a result of the fine tuning to CPU & RAM I've hardly lost any performance at all by running the CPU at a lower frequency & voltage.

 

According to silicon lottery you would be in the top 11 percent of what they have tested.

• CPU Multiplier: 51
• BCLK: 100.0
• CPU Vcore: 1.312V
• AVX Offset: 2

As of 12/07/18, the top 11% of tested 9900Ks were able to hit 5.1GHz or greater.

 

Z390 AORUS MASTER - Aka the VRM MASTER (en)
der8auer
Published on Dec 18, 2018

 

LoL - Athlon 200GE rekt's 9900k IGP in gaming...

i9 9900K (HD 630) vs Athlon 200GE (Vega 3)
Benchmark
Published on Dec 17, 2018
♦ CPU1: AMD Athlon 200GE with Radeon Vega 3 Graphics
♦ CPU2: Intel Core i9-9900K with Intel UHD Graphics 630
♦ RAM: G.Skill RipJaws 16 Gb (2x8Go) DDR4 3200 MHz


arthur 1 day ago
"haha. a 50$ cpu wins against a 600$ cpu haha"

 

... That comparison makes about as much sense as saying a bicycle beats a ferrari because fuel economy

 

Actually, it's more like when "fossil fuel" cars are banned, the top end speed of a bicycle will be an order of magnitude faster than a "fossil fuel" powered Ferrari.

 

Oof poor poor Ryzen.

 

Oww Lordy, it's a blood bath, that Ryzen just can't keep up...2 FPS slower @ 1440p... my'o'my... Good thing too, since the 2080ti can only run RTX on Battlefield V.
And, don't forget, you can get 2x 2700x for the same money as 1 9900k, so you have to multiply 2 x 2700k FPS against 1 x 9900k FPS = 9900k loses totally.

 

Don't see a Crossfire..,err ummm dual socket 2700X motherboard around, so....that kind of cancels out a total loss... but yet...the drama is real.

 

The price savings are real. Whereas the FPS difference won't be felt at all, the thinner wallet will be immediately apparent

And, look at that input laggy 9900k in Windows 10, what a shame...

Input DELAY - The 1680 v2 Vs. 2700X Vs 9900K - Part 2/3

Oww, that 9900k knows how to suck... power as well as cash!

 

And everyone that buys BGA says the exact same thing about the 2700X/9900K and the thinner wallet will be immediately apparent 10 times over.

Side note:
2013 still in the game!

Side note:
And he totally hits on what your saying about price.
Starting at 54 seconds on part 2.... beating a dead horse and all!

 

1440p Ultra in BFV singleplayer is pretty GPU-bound, making it pointless for comparing CPUs, yet the 9900K is still 18% higher in 1% lows.

 

Ryzen gets absolutely thrashed in gaming va the 9900K or basically any Intel CPU with a 2080 Ti. It’s really bad. Nice cherry pick of one example, in multiplayer it gets destroyed in BFV.

It’s sad when an 8/16 core Ivy Bridge-E is still beating Ryzen.

At this point Ryzen just might catch up to Intel performance with Ryzen 2 just to get passed up again with Intel 10nm Sunny Cove in the same year.

 

And yet, the FPS with the Ryzen 2700x is more than fast enough to enjoy gaming without spending $300 more for your CPU, or $800 more for your GPU - if you get a nice used 1080ti for $500... also overkill actually.

"More than fast enough."..."Learn not to Burn, Cash"... - Smart Buyers

"10nm?...? What 10nm? We're going directly to 7nm!, in 4 more years...!!" - Intel, Xmas 2019

Intel, the author of: " Intel, Our Book of Lies and Deceptions - The Suckers Eat it Up Edition"

Even odds Intel delays again past 2019, I really hope Intel doesn't, but they've kinda got a groove going with 10nm failures to deliver as promised.

What kind of cooling will that "Sunny Cove" 14nm 10c/20t CPU need, @ "95w TDP" require? Molten Thorium Salts??

 

Listen for 2 more seconds and he says he is letting the 9900K throttle down to an idle CPU clockstate. If he was using 5Ghz all day like most overclockers, the input test would have reflected what it showed in CSGO with the 9900K in the lead. At the end of the day these tests are somewhat silly imo even with the Ryzen CPU doing worse in gaming tests where input lag might matter. Those differences in gaming would likely never be noticeable between all of those CPUs.

 

Exactly, why spend 2x the $ on a 9900k vs a 2700x / 2700 when the differences in gaming would likely never be noticeable between all of those CPU's?

See it didn't take that long to figure out what I've been saying for a very long time, stop giving Intel / Nvidia all of your money - a large premium over what AMD charges for adequate performance.

That's being a good consumer, to reward those that reward you with good products at sensible prices. And, that's not Intel, Nvidia, Apple, etc.

 

1680 v2 Vs. 9900K Vs. 2700X - Productivity - Part 3/3
Tech YES City
Published on Dec 21, 2018
Here is the conclusion to part 3 of this showdown with 8 cores from 2013 (1680 v2) vs 8 cores from 2018 (i9-9900K and Ryzen 7 2700X). The Little xeon, it ain't no joke.

 

If that's the case then Ryzen 2 has no place in the market for gamers..

Look I get what you're saying about Ryzen, and RX GPUs (Vega sucks) having a place in the market for the budget consumer. I've said it before as well. Unfortunately for the high end consumer, enthusiast, AMD sucks right now aside from tile based rendering needs with Threadripper. For everything else there's Nvidia and Intel.

Claiming it offers "adequate" performance means nothing. That's like saying why do you want a sports car when the Honda Civic also gets you to the grocery store. Not everyone lives on the same budget or has the same desires. We could use guitars, coffee makers, or literally any other inanimate object and the result would be the same. There are high end and budget markets for anything.

 

Not everyone drives a Honda because they can't afford a Ferrari. Most of the people that can afford a Ferrari don't buy one because they aren't interested in wasting the money they have on frivolous superfluous spending - they think it's silly and wasteful.

Rolls Royce car engine output specifications for the longest time said: "Adequate", because that's what matters.

People with money didn't get rich by spending money needlessly in any area where money is spent. Budget's are limited by good common sense or by an absolute desire to empty your pockets whenever they are full enough to apply to some frivolous acquisition. Whether you can afford it or not isn't relevant to whether you need it or not.

The Ryzen 2 generational upgrade is a process bump with better production yields for continuing manufacture. The previous Ryzen 1 production is done - there is product in the pipeline until sold out, but no more are being produced.

It's like if Intel made 10nm versions of their 14nm parts, it's a process bump to graduate to better yields and more efficient production - if that actually worked for Intel, but it has been working for AMD - which they will continue with 7nm + an architectural change.

It's a matter of spending our money where it matters, where it actually improves your life vs your "magic bean benchmark counter" which doesn't deliver anything but "puppy dog tails" - you may not notice it's effect in real life but it sure is cute and gives you a warm feeling where your cash used to be.

It will be interesting to see the progression for both Intel and AMD entering new process production next year, as always YMMV - actually TMMV (their) - as every process bump brings new challenges along with new potential.

Benchmark Chart Placement = Cash out of pocket. The more cash out of pocket you spend, the higher you can place on the charts. Is it worth it? Not really.

 

Intel Core i9-9900KF, i7-9700KF, i5-9600KF, i5-9400F Prices Listed Tomshardware.com | Dec 22, 2018

Now major Norwegian and Finnish computer hardware retailers have listed four unannounced Intel 9th-generation Coffee Lake Refresh processors: The Intel Core i9-9900KF, Core i7-9700KF, Core i5-9600KF, and Core i5-9400F. The list prices are likely placeholders, so you should take them with a grain of salt.

Both retailers expect to have the new processors in stock by January 3, 2019 which could mean that Intel will announce the chips in a couple of weeks.

 

Finally, I've said since the iGPU came out that was a poor idea - waste of good power and space budget - to include an on-die / carrier GPU that vampires power and thermal headroom from the CPU package.

Maybe with that little bit of power and heat removed from the CPU it can now be stable at higher clocks, draw less power, and cool easier. For the lower power budget mobile CPU's the iGPU is the kiss of death for OC headroom... and don't get me started on "Optimus".

This will also be a good excuse for Intel to greatly drop the price so 9th gen can be a reasonable cost vs performance buy, with removing the iGPU as enough "reason" to not loose face for dropping the price significantly.

 

Intel Xeon E5 2670 vs i9 9900K
Benchmark
Published on Dec 21, 2018

Spoiler: Specs...

♦ CPU1: Intel Xeon E5-2670 @ 3.5GHz OC
♦ CPU2: Intel core i9-9900K @ 5.0GHz OC
♦ MOB1: ASUS P9X79 PRO
♦ RAM1: 32 Go (4x8Go) DDR4 2400 MHz
♦ RAM2: 32 Go (4x8Go) DDR4 3200 MHz
♦ GPU: EVGA GeForce RTX 2080 XC ULTRA GAMING

 

Sounds like a great idea. Always wondered why they even bothered with the integrated graphics on the higher end chips; however, this is just coming from a gamer and I am sure there are many who are not buying dedicated graphics.

 

In order to compete with AMD's CPU-only products Intel needs to ditch the iGPU to compete on even ground - as Intel reaches the limits of their resources in top end CPU's at 14nm, every little bit counts.

At the top end of the CPU power range an iGPU bottlenecks the CPU. Soon after the iGPU was born the succeeding generations of CPU's gained performance such that now somewhere around lowest mid-range, the iGPU is a mismatch with the CPU, not giving enough GPU power to not bottleneck the CPU.

Resulting in the ridiculous pairing of an onboard iGPU with the 9900k:
http://forum.notebookreview.com/thr...-lake-cpus-z390.811225/page-133#post-10835652

For Intel there was a fleeting moment where it made sense to add an iGPU on board with the CPU, but soon it was a waste of silicon, thermal and power budget, and Intel was never able to increase the performance of the iGPU enough to make sense in any but the lowest end CPU's.

AMD has done a much better job building onboard GPU's in the APU ranges, matching performant GPU's onboard with lower performance CPU's. Even so AMD doesn't include onboard GPU's with all of their CPU's, because it is a bad idea.

 

lol where exactly do you call home, "make a wish and get a pony land"?

these KF skus will just be chips that Intel would otherwise have declared as defective due to flaws in the iGPU area. so its gonna be rebranded, defective silicon, thats it. no more OC headroom, MAYBE 5 bucks less in pricing and more money to intel

 

You forget one thing... More chips Intel is able to throw out can stagger the retailers greed. Less chips for sale = Higher prices. Would you produce more (chips) and earn less?

 

in theory yes. but so far im not seeing any huge price decreases... so lets wait and see how this goes

 

You can't see price decrease as long the new chips ain't out

 

well duh i was actually talking about the chips staying in stock and not being sold out inside 5 min every time theyre offered

 

The longer chips staying in stock, something has to be done. See... Norwegian retailer. 4 in stock. And many havent a single chips in stock. This won't make big change in prices You can only guess what would happen if all the retailers here home had +20 each and one and not a few with 1-4

 

Today isn't the first day of Intel's life generating CPU's, and failed units with bad iGPU's must have happened before now, right?

So where are the Intel CPU products of failed iGPU's in CPU SKU's before today?

Are you saying that Intel never had enough failed units in production before the 9th generation CPU's, so that's why Intel never did a similar release of CPU's without iGPU's?

Failed 9th generation die's with bad iGPU's aren't why these new 9th generation KF CPU's are being made and released.

These are good 14nm dies, not failed 10nm production, without iGPU's.

 

right, so you think Intel just came up with a completely new 8 core design missing iGPU? i highly doubt that...

also, i never said that Intel never had defective CPUs before. just that they now start to sell those, which they apparently did not do before!

 

The key focus should be on the fact that Intel is releasing their top 9th generation CPU's without onboard GPU's- whether the iGPU missing or deactivated isn't pertinent - without the iGPU onboard Intel is now shipping it's top CPU's with the same functionality as AMD's top CPU's - without an onboard GPU.

Is this another result from the influence of AMD's Ryzen / ThreadRipper CPU's competition on Intel's product shipments?

Possible reasons why is Intel doing this was the subject of my previous posts, no need to repeat it here again.

I would add that the Intel KF CPU price information is still not known as is Intel's reasoning for these new sku's - publicly given or privately held, there has to be a good business reason for producing and shipping these new KF CPU's.

 

Igpu area is still around ~30% of the die area of CFL-R so theoretically nearly that amount of GPUs with a critical defect could be salvageable as KF

Its just another aspect to the standard binning process (a defect in a physical or logical core could be why a die is 9600K or 9700K not 9900K) and makes more sense at the top end where 99% of users don't use igpu ever not even for quicksync

The igpu is completely unpowered in the LGA Clevos for example, if a few bucks savings are to be had (or a KF instead of a no stock cpu) it makes sense

Probably hasn't happened in the past because Intel's supply has not been constrained and its not been worth their while to have the additional product lines in the channel.

 

Possibly, but it seems like an unusual step to take, the production failure rate would need to be really high to make this feasible - profitable - and even so it's not going to be many units - an order of magnitude more limited in availability than the normal successful die's with iGPU.

Given the longevity of the 14nm process, I would think that the failures specific to iGPU die area would be vanishingly small. Failed die's would hardly generate quantities worth creating whole new sku's, you'd be assured to not have enough to go around, creating another failed sku more unavailable than the existing sku's with iGPU's.

That is unless Intel is having some "surprises" with their 14nm+++(...)+ current generation, and Intel is seeing an unexpectedly higher number of failed dies per wafer. That might be the underlying reason for the production shortages.

I've thought it's very odd Intel was so far off the mark estimating 14nm production yields so late into their process - and with so much advance notice that their 10nm process was failing - Intel had plenty of time to adjust production facility allocation to 14nm production to produce enough product.

Given AMD has taken away so many sales, Intel should be flooded with inventory, not having shortages. Every AMD CPU sold that replaces an Intel CPU sale is one less CPU Intel needs to produce, that should be millions fewer CPU's required for Intel to deliver in the consumer market.

AMD CPU sales reducing the need for Intel CPU's in the consumer market was the reason I attributed to Intel switching 2 million units of production from consumer to commercial production, not production shortages.

I wanted to wait for someone to "scratch off the die" to find out what the actual CPU layout look's like, and compare it to the current iGPU sku's before "guessing"...

Given Intel's recent 10nm experience with shipping CPU's without iGPU's it shouldn't have been difficult for Intel to re-layout the 9th generation KF CPU's without iGPU's to save die space.

30% of area saved per CPU, if designed and created without that iGPU area on the die for the new KF CPU's would mean more CPU's per wafer, and given the smaller area likely fewer failed dies per wafer.

That to me is the best most logical path for Intel to be taking, reducing the area for each CPU die to increase yield output per wafer.

Intel could generate as many as 2x good KF dies per wafer without the iGPU area, taking into account more dies per wafer + fewer failed dies per wafer.

In order to "sell the idea" of no iGPU CPU's Intel is going to have to motivate people to buy the KF CPU's over the iGPU CPU's. Besides availability only price comes to mind as a motivator, unless Intel can squeeze more performance out of the KF CPU's.

Intel is going to have to make the KF CPU's an attractive purchase by lowering the price significantly.

 

It is highly unlikely that Intel is having production issues with 14nm, but rather delay on 10nm causing shortage on their 14nm since their entire production stack relies on 14nm.

AMD did take some sales away from Intel but mostly on CPU retail market which only represent very small percentage of overall market. AMD has yet to make a huge impact on OEM shares and they have almost no presence on notebook market which now makes up majority of client PC sales. Along with very healthy growth on datacenter/server market, overall industry is growing faster than Intel's ability to supply and AMD's ability to replace. Remember Intel had record revenue last quarter despite AMD's growth on CPU division.

 

How many units do you think Intel can recover by disabling the iGPU in bad dies and shipping those 9th generation CPU's as KF models?

If the 14nm process is refined, with barely 3%-5% failed dies overall, what percentage of die failures will only be in the iGPU part of the die?

How could that miniscule number of recovered dies allow Intel to ship enough of each KF sku to make the new sku's worth creating?

I don't think there are enough failed iGPU dies in any 9th gen sku to generate enough inventory to justify creating the KF CPU sku's.

There must be another reason for Intel to begin shipping 9th gen KF CPU's without iGPU's. It would be nice to know if Intel is disabling the iGPU's or producing dies without iGPU's to begin with. Or, if there really are so many failed dies with only-iGPU failures.

If there are enough failed iGPU dies available to fill inventory in all of the KF sku's without disabling good iGPU's in fully functional dies, then Intel is having a much higher than expected die failure rate for their mature 14nm process node.

Also, if Intel is disabling perfectly good iGPU's in 9th generation CPU's to create the KF sku's, then Intel is lying about the 14nm production shortage. Intel wouldn't ruin perfectly good iGPU CPU's to sell them as KF non-iGPU sku's if they really had a 9th generation CPU 14nm production shortage.

If Intel is having production problems with 14nm resulting in shortages, so much so that they need to recover the tiny % of dies with iGPU-only failures then Intel is in great trouble.

I think Intel is in great trouble for so many reasons, and this is just another contributing factor.

 

Fire up new factorie(s) lines (to spit out a lot more chips fastest as possible)may increase the production failure rate. And if Intel removed 30% of the whole die it will run hotter. Not a good idea on this gen chips.

 

They are simply just trying to produce more dies per wafer instead of increasing capacity for client Desktop CPU to ease up the shortage, since they are bottom of the list in terms of Intel's production priority anyways.
It is kinda funny to see "Intel is in great trouble" narrative when they are enjoying record quarters for a while. They would be in trouble if they are oversupplying, but so far, demand is much higher than Intel's ability to supply and will be for a foreseeable future, thanks to demands from datacenter/server market due to rapid increase in cloud computing market.

 

Actually, the current DC sales growth is halved from the beginning of the year, and is continuing to slow.

Why Intel's Data Center Growth Is Set to Slow
Tough comps make it harder for the chipmaker to post fast growth.
Ashraf Eassa (TMFChipFool), Nov 20, 2018 at 9:37PM
https://www.fool.com/investing/2018/11/20/why-intels-data-center-growth-is-set-to-slow.aspx

"...Although that figure is solid in itself, it would represent year-over-year growth of just 13% -- a far cry from the growth rates that the segment enjoyed in the first three quarters of 2018.

Let's go over why the growth rate of Intel's DCG segment is set to come down in the fourth quarter of 2018.
...
After a huge 2018 for DCG, don't be shocked if that growth rate comes down substantially in 2019. "

Amazon unveils its own server chip, challenging Intel on price
Ian King, Bloomberg News, Nov 27, 2018
https://www.bnnbloomberg.ca/amazon-unveils-its-own-server-chip-challenging-intel-on-price-1.1174339

"...The company said the Graviton-backed cloud service is available a "significantly lower cost" than existing offerings run on Intel processors.

Wells Fargo analyst Aaron Rakers estimated the new service will be up to 45 per cent cheaper than the equivalent offering based on Intel chips. "This will add another competitive question on Intel’s positioning within AWS Cloud," he wrote in a research note Tuesday.

It’s the second time this month that Amazon has taken a swing at Intel’s server chip business. The cloud provider said on Nov. 6 that it’s offering services based on computers that use Advanced Micro Devices Inc. (AMD.O) processors. That marked a breakthrough for AMD’s efforts to compete against Intel."

------------------------

Given Intel's slowdown in sales, for desktop, laptop, mobile, and data center, coupled with Intel's failure to move to 10nm - and now 7nm - Intel's competitors across all product lines will have a much greater effect on Intel's bottom-line, Intel can hide their problems for only so long before it's clear what is happening.

Even if Intel starts shipping 10nm products, I doubt the products will outperform the current 14nm performance by much if any significant percentage. 14nm being a mature process, and 10nm starting out again at the beginning of the optimization cycle, it would be years before the best of 10nm will ship.

While the competition is releasing 7nm, then 5nm, and then 3nm all in practically no time.

Starting now, within the same span of years that Intel has frittered away failing at 10nm, all of the rest of the industry will be at 3nm.

Where will Intel be then? Likely still stuck on 14nm / 10nm.

 

I will see it before I belive it. What said now doesn't be as easy as you think In what time frame?

 

Intel's 10nm is slightly more denser than other competitor's 7nm and Intel's future 7nm is going to be comparable to TSMC's 3~5nm. Lets not go crazy about nm values since they are no longer represent standard actual node sizes.

Intel has been getting trouble with 10nm for a long time but when they released the info on Sunny Cove, it is first time that they actually detailed the architecture information so looks like they will actually hit the market in 2019-2020, which wouldn't be so behind AMD's schedule. Intel's 7nm's relatively unknown at this point but it doesn't mean 10nm troubles will carry over to 7nm so we will wait and see on that one.

 

In "2 Shakes of a Lamb's Tail"?

TSMC gets approval to build 3-nanometer factory in southern Taiwan
Construction will begin in 2022
By Shawn Knight on December 20, 2018, 1:40 PM
https://www.techspot.com/news/77963-tsmc-gets-approval-build-3-nanometer-factory-southern.html

TSMC to Start 5nm Production With Full EUV in April 2019
Friday October 05, 2018
https://www.hardocp.com/news/2018/10/05/tsmc_to_start_5nm_production_full_euv_in_april_2019

TSMC to Start 5nm Production in April
5 October 2018, Rick Merritt
https://www.eetasia.com/news/article/18100502-tsmc-to-start-5nm-production-in-april

"SAN JOSE, Calif. — TSMC taped out its first chip in a process making limited use of extreme ultraviolet lithography and will start risk production in April on a 5-nm node with full EUV. Separately, the foundry forged partnerships with four partners to support online services for back-end chip design."


Intel is so far behind they really don't have a chance of catching up. Intel is already looking to TSMC to take on production.

Global Foundries already came to that conclusion - GF were too far behind and the market to sustain multiple players @ 7nm / 5nm ( / 3nm ) wasn't lucrative enough compared to the investment. GF wanted to be first in to capture the market for 12nm / 14nm / 22nm / etc foundry business before Intel figured it out.

Although, Intel has just recently given up their Custom Foundry business and given the market away to GF and other competitors:

Intel Discontinues the Custom Foundry Business!
by Daniel Nenni, Published on 12-17-2018 06:00 AM
https://www.semiwiki.com/forum/content/7912-intel-discontinues-custom-foundry-business-e.html

Yet another Intel failure, IDK about you but to me it looks like Intel is losing ground rapidly.

 

Maybe for coming phone processors... Yes. For computers is another task. And don't forget the cost. 7nm process shall be paid first.

 

TSMC 7nm is building AMD's Ryzen, ThreadRipper, Epyc CPU's and Navi / Polaris GPU's - with 5nm / 3nm for future generations of AMD products, not just phones or mobile devices.

AMD CEO Su Claims Navi Radeon GPUs Will Challenge NVIDIA High-End Turing In 2019
by Paul Lilly — Tuesday, December 04, 2018
https://hothardware.com/news/amd-dr-lisa-su-claims-next-gen-radeon-gpus-challenge-nvidia-turing

" AMD is investing heavily into 7-nanometer manufacturing technology, both for its CPUs and GPUs, only one of which currently competes at the high end of the market. That would be AMD's processor family, and specifically its Ryzen, Threadripper, and Epyc chips.

What about graphics? AMD has mostly focused on the mid-range and low-end sectors with several refreshes to its Polaris GPU architecture, but is not willing to concede the high-end market to NVIDIA.

" We believe, we will be very competitive overall and that includes the high-end of the GPU market. Obviously there are new products out there from our competition. We will have our set of new products as well and we will be right there in the mix," AMD CEO Dr. Lisa Su said at the 22nd Annual Credit Suisse Technology, Media, and Telecom Conference."

Intel is going to have a very tough road ahead for the next 3-5 years, which is about how long I think it will take them to catch up to TSMC / Samsung.

That is only if Intel can afford the capital investments along the way, which are double digit $Billion's and progressively more at each node along the way.

Does It Matter If TSMC Dethrones Intel and Takes the Semiconductor Lead?
Joel Hruska on November 29, 2018 at 11:16 am
https://www.extremetech.com/g00/com...hrones-intel-and-takes-the-semiconductor-lead

 

Phones chips will be the Guinea pigs. From what I know they won't jump right on high performance computer chips first. 3nm for future generations of AMD products. Yees

¨Will they afford lose sales? Nope.