Asus CEO: Intel 14nm CPU Shortage Has Affected Business

[BonScott]

Intel 10nm or Bigfoot or the Loch Ness monster. LARGE dies make 10nm too low yielding and Intel is getting is ass kicked at the high end, except for the rare $5K platinum CPU that still needs exotic cooling,

(and no reply to this comment or other FACTS spelling Intel's doom,, just nitpick AMD'd projected server share, which doesn't matter because its REAL market share is what matters) Lets see where we are one year from now CLOWNS

 

[BonScott]

8 core does with great yields at TSMC, where Lisa Su was born. Don’t worry about capacity, it will get there, worry about Intel.

 

[BonScott]

Oops “8 core dies”

 

[BonScott]

AMD (who was on 28nm just a couple years ago) will be on 7nm, while INTC has been at 14nm the whole time.

Well to be fair the way fabs name nodes, TSMC "7nm" (and Samsung's 7nm) will probably be roughly on par with Intel's 10nm... but Intel is struggling a lot more than TSMC, which is embarrassing! However, Intel's current tweaked 14nm is still pretty decent, and their Core designs are good... so it's not as rough for Intel as the Netburst era.

Anyway I am a little concerned that TSMC will also end up being extremely strapped for capacity. The good news is that AMD's I/O plus chiplets paradigm vastly boosts the number of CPU cores they can squeeze out of every leading-edge wafer, and enables them to optimize the I/O block to make best use of a more mature process.

Intel has UPI which is just as fast is not slightly faster (latency wise) than IF is.

Well, it's not as though AMD is standing still, either. They're claiming better and more predictable performance for IF as part of their new setup. I would bet these improvements trickle down to processors with less chiplets and a more diminutive I/O chip. All very interesting to see unfold.

Yea, TSMC is going to be real busy making Apple’s iPhone CPUs that no one is buying until Apple and everyone else has a 5G modem. Been paying attention what Apple is saying? TSMC ain’t busy with their business. Been paying attention to crypto? TSMC ain’t making GPUs like hot cakes.

Knowing the underlying tech as a design EE myself and the history going back decades and seeing Intel fail at everything including Optane, which was suppose to be the way they were going to force everyone to buy there overpriced (now inferior) CPUs.

This is $ in the bank. INTC short or AMD long. Wait until Apple makes peace with QCOM for 5G modem.

 

[kinggremlin]

I guess a fool thinks mid double digits in servers = 10%. From Lisa Su’s lips. If you had any clue about tech, how to read a benchmark, etc, you would see, but please please please buy INTC, it’s a good value, gives a dividend, has a shit overprices, low yielding CPU about to get trounced.

https://www.notebookcheck.net/AMD-s-server-marketshare-hits-1-for-the-first-time-in-4-years.321130.0.html

"The CEO said Rome, or the second generation of Epyc, will be a necessary step in reaching 10% share in the server market. Su went on to say AMD feels "good about our competitive position and the path to double digit market share." Regardless of what detractors say, AMD and its CEO remain very confident in the company's ability to reach 5% by the end of 2018 and then 10% sometime in 2019 with Rome"

https://www.extremetech.com/computing/277793-amd-may-regain-30-percent-desktop-market-share-by-q4-2018

"DigiTimes is predicting 5 percent server market share for AMD by the end of Q4, which fits with Lisa Su’s own mid-single-digit projections for the year and even Intel’s comments on the topic."

You're bad at reading lips. She said mid single digits in 2018 and 10% in 2019. Nothing about mid double digits.


"Its Ryzen Mobile chips have begun to take some market share but the last figures we saw for AMD in this space put it under 5 percent of the mobile market and no real presence in slate PCs."

Under 5% in the mobile market. 1% in the enterprise market. Yea, AMD is just around the corner from hitting 50% market share in the overall market....

 

[ravitheult]

I like how politely you are responding to his offensive words. Actual fool is none other than him. Can't read.. Can't accept the reality.. 😀
Just ignore him..

 

[BonScott]

I guess a fool thinks mid double digits in servers = 10%. From Lisa Su’s lips. If you had any clue about tech, how to read a benchmark, etc, you would see, but please please please buy INTC, it’s a good value, gives a dividend, has a shit overprices, low yielding CPU about to get trounced.

https://www.notebookcheck.net/AMD-s-server-marketshare-hits-1-for-the-first-time-in-4-years.321130.0.html

"The CEO said Rome, or the second generation of Epyc, will be a necessary step in reaching 10% share in the server market. Su went on to say AMD feels "good about our competitive position and the path to double digit market share." Regardless of what detractors say, AMD and its CEO remain very confident in the company's ability to reach 5% by the end of 2018 and then 10% sometime in 2019 with Rome"

https://www.extremetech.com/computing/277793-amd-may-regain-30-percent-desktop-market-share-by-q4-2018

"DigiTimes is predicting 5 percent server market share for AMD by the end of Q4, which fits with Lisa Su’s own mid-single-digit projections for the year and even Intel’s comments on the topic."

You're bad at reading lips. She said mid single digits in 2018 and 10% in 2019. Nothing about mid double digits.


"Its Ryzen Mobile chips have begun to take some market share but the last figures we saw for AMD in this space put it under 5 percent of the mobile market and no real presence in slate PCs."

Under 5% in the mobile market. 1% in the enterprise market. Yea, AMD is just around the corner from hitting 50% market share in the overall market....

 

[BonScott]

CLOWN, "double digit market" Lisa Su

Pay attention

 

[BonScott]

I guess a fool thinks mid double digits in servers = 10%. From Lisa Su’s lips. If you had any clue about tech, how to read a benchmark, etc, you would see, but please please please buy INTC, it’s a good value, gives a dividend, has a shit overprices, low yielding CPU about to get trounced.

https://www.notebookcheck.net/AMD-s-server-marketshare-hits-1-for-the-first-time-in-4-years.321130.0.html

"The CEO said Rome, or the second generation of Epyc, will be a necessary step in reaching 10% share in the server market. Su went on to say AMD feels "good about our competitive position and the path to double digit market share." Regardless of what detractors say, AMD and its CEO remain very confident in the company's ability to reach 5% by the end of 2018 and then 10% sometime in 2019 with Rome"

https://www.extremetech.com/computing/277793-amd-may-regain-30-percent-desktop-market-share-by-q4-2018

"DigiTimes is predicting 5 percent server market share for AMD by the end of Q4, which fits with Lisa Su’s own mid-single-digit projections for the year and even Intel’s comments on the topic."

You're bad at reading lips. She said mid single digits in 2018 and 10% in 2019. Nothing about mid double digits.


"Its Ryzen Mobile chips have begun to take some market share but the last figures we saw for AMD in this space put it under 5 percent of the mobile market and no real presence in slate PCs."

Under 5% in the mobile market. 1% in the enterprise market. Yea, AMD is just around the corner from hitting 50% market share in the overall market....

 

[BonScott]

You don't even know what is most important. Cloud is more than half of all servers now. AMD will rule Cloud. The highest margins, Intel's castle.

 

[BonScott]

nitpick all you want FOOLS
Remember what the scene in Cloud looks like a year from now. You know nothing but Intel hype. Things change.
Get some popcorn and watch the show.

 

[chickenballs]

as it stands now I may have to go AMD when it is time for me to upgrade the i7 4790K

 

[redgarl]

How about promoting AMD platforms...?

They do. Problem is branding is not AMDs forte and they have never had a heavy marketing presence outside of enthusiasts. There are people who will only buy Intel so pushing AMD to those people is the same as pushing Intel to AMD, after all I knew plenty of people who bought Phenom 9500s even though they were inferior to Core 2 Quad and plenty of people who bought a Pentium 4 over an Athlon 64.

Hell no, they don't. Asus has kept their AREZ brand. They don't push AMD and now that shortage is going to hit them hard for almost a year. AMD is having an incredible chip called the 2800h for laptops. It can make crazy thin form factor with 720p max setting in most of today games... but they don't do it.

I would just buy one if there was one available to the market, but OEMs are dumb and they are still putting 1030 GTX and 1050 GTX in their laptops while they are completely obsolete as of now.

There is no reason why AMD is not a major player in mobile except OEMs stupidity.

 

[BonScott]

I guess a fool thinks mid double digits in servers = 10%. From Lisa Su’s lips. If you had any clue about tech, how to read a benchmark, etc, you would see, but please please please buy INTC, it’s a good value, gives a dividend, has a shit overprices, low yielding CPU about to get trounced.

https://www.notebookcheck.net/AMD-s-server-marketshare-hits-1-for-the-first-time-in-4-years.321130.0.html

"The CEO said Rome, or the second generation of Epyc, will be a necessary step in reaching 10% share in the server market. Su went on to say AMD feels "good about our competitive position and the path to double digit market share." Regardless of what detractors say, AMD and its CEO remain very confident in the company's ability to reach 5% by the end of 2018 and then 10% sometime in 2019 with Rome"

https://www.extremetech.com/computing/277793-amd-may-regain-30-percent-desktop-market-share-by-q4-2018

"DigiTimes is predicting 5 percent server market share for AMD by the end of Q4, which fits with Lisa Su’s own mid-single-digit projections for the year and even Intel’s comments on the topic."

You're bad at reading lips. She said mid single digits in 2018 and 10% in 2019. Nothing about mid double digits.


"Its Ryzen Mobile chips have begun to take some market share but the last figures we saw for AMD in this space put it under 5 percent of the mobile market and no real presence in slate PCs."

Under 5% in the mobile market. 1% in the enterprise market. Yea, AMD is just around the corner from hitting 50% market share in the overall market....

 

[BonScott]

Aren't you the same CLOWNS that said AMD would go BK when it was $2/sh?
Or are you the CLOWNS that said Intel's Server castle was un-penetrable?
Now you think AMD isn't about to put a world of hurt on INTC, even though Epyc is now qualified for servers and SUPERIOR to anything Intel can produce for 5 years at least.

Remember CLOWNS:
AMD $20/sh on 11/15/2018
INTC $47/sh on 11/15/2018

 

[jimmysmitty]

AMD (who was on 28nm just a couple years ago) will be on 7nm, while INTC has been at 14nm the whole time.

Well to be fair the way fabs name nodes, TSMC "7nm" (and Samsung's 7nm) will probably be roughly on par with Intel's 10nm... but Intel is struggling a lot more than TSMC, which is embarrassing! However, Intel's current tweaked 14nm is still pretty decent, and their Core designs are good... so it's not as rough for Intel as the Netburst era.

Anyway I am a little concerned that TSMC will also end up being extremely strapped for capacity. The good news is that AMD's I/O plus chiplets paradigm vastly boosts the number of CPU cores they can squeeze out of every leading-edge wafer, and enables them to optimize the I/O block to make best use of a more mature process.

Intel has UPI which is just as fast is not slightly faster (latency wise) than IF is.

Well, it's not as though AMD is standing still, either. They're claiming better and more predictable performance for IF as part of their new setup. I would bet these improvements trickle down to processors with less chiplets and a more diminutive I/O chip. All very interesting to see unfold.

Not assuming they are standing still. I doubt we will see IF trickle down soon to the consumer mainstream market mainly because IF and UPI are really designed more for multi chip systems. Until we start using multiple CPUs or AMD/Intel decide to go to MCM designs for mainstream its not as useful to us.

Of course Intel will also make improvements so both will have to be on their toes.

 

[logainofhades]

Intel right now doesn't seem to be firing on all cylinders. They lacked competition for so long, that they seem to be struggling with it, now that they have it.

 

[stdragon]

... when a basically $4B/yr revenue, break even company goes from basically 0% market share of a $30B/yr market to having clearly the best CPU and roadmap to take at least half of that ... what happens to their stock price.

Its still has a LOOOOOOOOOOOOOOOOONG way to go.

There is more to it than just swappinf out. There are a lot of gears in a network and server farms. If you run VM you can't just swap out to AMD or Intel. Normally you would change out all VM Hosts at once. Not many companies will front that kind of money. However we will see. You are expecting AMD to grab 50% market share in a year or two. I think it would take much longer than that to get up to 50% and that is if it stays the current way it is with AMD having more cores per dollar.

I guess we shall see but I for one want to see Intel strike back to make AMD strke back harder.

For one, you don't need to replace all your hosts within the cluster so long as you're accepting of the minimum EVC compatibility in CPU features. Effectively the lowest common denominator. This allows you to use different generation CPUs within the cluster while still ensuring that all running VMs can vMotion between hosts without interruption.

The problem however is you can't mix AMD and Intel within the same cluster. Meaning, you can't vMotion between an Intel host to an AMD host and vice versa. You can however cold migrate a VM (in an off state) between different clusters however. So for a reduction in costs in a cloud hosting service, it might be worth it to "convert" your VM via shutting it down and having it hosted on AMD. Meaning, the down-time might be worth it. So unless it's mission critical, downing a website for a few minutes shouldn't be an issue 😉

BTW, not being able to live-migrate between Intel and AMD is not just a VMWare issue, but the same is true for Hyper-V, Xen, and basically any and all hypervisors (both type 1 and 2). Meaning, it's a function of abstraction between differing CPU architecture or microcode.

Anyways, that all said, I could totally see AMD hoovering up the data-center market.

 

[alextheblue]

Not assuming they are standing still. I doubt we will see IF trickle down soon to the consumer mainstream market mainly because IF and UPI are really designed more for multi chip systems. Until we start using multiple CPUs or AMD/Intel decide to go to MCM designs for mainstream its not as useful to us.

Of course Intel will also make improvements so both will have to be on their toes.

IF is *already* a big part of Ryzen from the beginning. Even if you discount MCM setups, IF is the special sauce for both dual CCX and single CCX + GPU chips. IF speed being locked to RAM speed is a large part of why faster RAM boosts dual CCX chip performance so much. So any improvements to IF can trickle down.

But that wasn't what I was suggesting. I was suggesting that future mainstream setups can easily end up with dual chiplets + a baby I/O chip on the same package. Could be one CPU and one GPU chiplet, or could be a pair of CPU chiplets. Even on desktop it has a lot of advantages in terms of production - bleeding edge process for the chiplets, power-tweaked older process for I/O. That's how I see this playing out in the long term. It will probably come to HEDT first, but I don't think it will stop there.