News AMD Raising EPYC CPU Pricing 10 to 30%, Intel's Sapphire Rapids Delayed: Report

waltc3

Reputable
Aug 4, 2019
420
223
5,060
If true, it would appear that AMD is raising the price of new EPYC CPUs based on Zen3, or for MilanX not yet shipping. So apparently AMD is not raising the price on existing EPYC CPUs. But--this is one of those "reports" which has a 50% chance of being accurate....;) Supply has to increase substantially and soon, else all the chip manufacturers will be in trouble, imo. Let's hope current estimates of supply increases to near-normal are accurate in the 2H.
 
  • Like
Reactions: VforV

spongiemaster

Admirable
Dec 12, 2019
2,273
1,277
7,560
Article repeatedly states that Sapphire Rapids is based on Intel's 7nm process. It is based on Intel 7, which is their 10nm process. There is no "nm" rating attached to Intel's new process labeling. Intel's actual fake 7nm process will be called Intel 4. A tech site this large needs to get basic facts like this correct.

Edit, article has been corrected. and the above no longer applies. Thank you.
 
Last edited:
  • Like
Reactions: sgh55fg

MasterMadBones

Distinguished
Article repeatedly states that Sapphire Rapids is based on Intel's 7nm process. It is based on Intel 7, which is their 10nm process. There is no "nm" rating attached to Intel's new process labeling. Intel's actual fake 7nm process will be called Intel 4. A tech site this large needs to get basic facts like this correct.
Although it is weird that Intel changed its naming like this, it makes more sense when you look at its competition. The "old" 10nm proces that was renamed to 7, has transistor dimensions and density similar to TSMC 7nm, not 10nm. I don't really have major issues with calling it 7nm then, but it does lead to confusion over whether the old or new naming scheme is used.
 
  • Like
Reactions: shady28

spongiemaster

Admirable
Dec 12, 2019
2,273
1,277
7,560
Although it is weird that Intel changed its naming like this, it makes more sense when you look at its competition. The "old" 10nm proces that was renamed to 7, has transistor dimensions and density similar to TSMC 7nm, not 10nm. I don't really have major issues with calling it 7nm then, but it does lead to confusion over whether the old or new naming scheme is used.
The article has been corrected, so my post won't make sense to anyone reading the article now.

I am in agreement with Intel's name change. It needed to be done based on what their competitors were doing.
 

Co BIY

Splendid
10% price increase would just be matching inflation in US Dollar denominated transactions. So not that large an increase given the environment.

Not at all surprised that customers are in a "Shut up and take my money" mode.
 

VforV

Respectable
BANNED
Oct 9, 2019
578
287
2,270
Don't kid yourself, AMD has what... 16% of server market. A few weeks/months ago it was listed 3-6months wait period for new Epyc cpus from AMD.
Don't kid yourself thinking you actually talk sense in this pathetic attempt of whatever you trying to prove, starting with your 1st **** remark "worthless to be in the leadership position on the server side " and ending with this one.

Here, some brain matter:
Further analysis reveals that revenue from the Computing and Graphics segment has grown at a rate of 46%, while that from the Enterprise, Embedded and Semi-Custom segment has grown at a rate of 286% during this period.
LINK > https://businessquant.com/amd-revenue-by-segment

286%, that's for 2021 for Servers. NOW please tell me how that is worthless without you sounding like an [***].... You're just digging yourself deeper.

Actually you know what, I don't need any more noise from you, you have proven what you are and what you lack: reason. Now stay silent on my IGNORE list. I'm done wasting my time with the types of you. Bye.
 

msroadkill612

Distinguished
Jan 31, 2009
202
29
18,710
The finalhalf is a mass of contradicrtions leading to an inescapablly pro amd bottom line.

"AMD's Milan holds the peak performance, and perhaps more importantly, performance-per-watt crown in the data center. Milan's higher core counts also offer superior performance density, thus allowing data centers to cram in more performance per server and further lower their operating costs, all of which is obviously attractive enough to justify the price increases. "

"Intel has ramped Ice Lake production and projects a possible 50% year-over-year increase in supply in 2022. In addition, in an attempt to retain market share, Intel isn't raising prices for its Ice Lake processors. Both of those factors could help blunt AMD's market share gains.

However, the report claims that Intel's full-scale Sapphire Rapids volume launch will be delayed until 'maybe' Q3 2022 (Intel had previously projected the formal launch in Q2 2022). Furthermore, the firm doesn't anticipate Sapphire Rapids will help Intel much, predicting that Intel will see further market share degradation due to 'materially higher' BOM pricing for Sapphire Rapids. "

"Intel CEO Pat Gelsinger recently said that Sapphire Rapids should put the company close to parity with AMD's data center chips (Sapphire Rapids will win some, lose some), but that he doesn't expect to gain a clear lead over AMD until Intel fields the next generation of data center chips. Wu sees that a bit differently, projecting that Milan and Genoa will retain the performance advantage over Xeon, resulting in AMD's continuing its explosive growth in the data center.

AMD is more supply-constrained than Intel, so its supply of silicon is obviously holding the company back from gaining market share more quickly. Still, Intel faces a tough challenge as TSMC continues to ramp 7nm capacity and AMD's 5nm Genoa comes to market later this year"

~ "they have more supply but no supply, not putting upprices but higher cost & uncompetitive perf, TSM is ramping 7nm & epyc will grow explosivedly"

the key is intel have nothing like Infinity Fabric - w/o which chiplest are just a buzz word.

all those cores are pointless if there are too many hops for inter chiplet core to core links - that has been amdS focus for a decade~

what it says is even when they place themselves at a huge bom disadvantage & throw everything at the problem, they are screwed on power & perf

amds margins will grow & grow, & intel will slowly starve while pretending to compete.

They are doomed but small investors wont know til too late.
 

kwohlt

Commendable
Oct 7, 2021
35
37
1,560
Article repeatedly states that Sapphire Rapids is based on Intel's 7nm process. It is based on Intel 7, which is their 10nm process. There is no "nm" rating attached to Intel's new process labeling. Intel's actual fake 7nm process will be called Intel 4. A tech site this large needs to get basic facts like this correct.

Edit, article has been corrected. and the above no longer applies. Thank you.

But...you're the one with your basic fact that is wrong. You say that Intel doesn't use nm in their process labeling, and then also label Intel 7 as 10nm. What exactly makes it 10nm? There're no measurements that are 10nm in that process. Density, perf/watt, and node size are certainly similar to TSMC 7nm. So what is 10nm? Do you correct articles on Zen Mobile 6000 and say "Acktually, TSMC 6nm is TSMC's 7nm process that just has some improvements"?
 

guru7of9

Reputable
Jun 1, 2018
58
7
4,545
observation - over the last 3-4 years all i see is Intel copy AMD in quite a few things. First they bag them for it, then they copy it! Intel are no longer the technology leader ! Latest they are trying to copy the Ryzen 3d v cache!
They are trying hard now to get back to parity but they still have a way to go! Server side of things yes AMD cant build them fast enough but these days they are miles better than anything Intel has currently or for a a good while to come! Unheard of, AMD server cpus are vastly superior to Intel ! Intel used to brag about how servers where it really matters cos that where all the money is ! Mind you they still have dominance of market share as it takes a lot longer for change in the server space but AMD have made massive inroads!
Alderlake is good and innovative but only out of neccessity as it is a reaction to Ryzen superiority and core count at the time of first release! Intel cant work out how to make cpu cores faster using less power. If they could they would"nt need to use e cores for desktop! Now for laptops i understand e cores are very uselful , but for desktop just cant see the point of it for the vast majoritory of users! Will be interesting to see how it all pans out, especially with Raptorlake supposedly coming out with 16 e cores and 8 p cores. and supposedly significantly larger cache!
If only more cheaper Intel motherboards can overclock 12400 cpus to 5ghz +using bclock . Now that i would buy, but alas i am sure Intel will just release bios update to stop it in its tracks if it takes off!
But the AMD better than Intel scenario, is all pretty much unheard of since way back in 2004 or so!
 

spongiemaster

Admirable
Dec 12, 2019
2,273
1,277
7,560
But...you're the one with your basic fact that is wrong. You say that Intel doesn't use nm in their process labeling, and then also label Intel 7 as 10nm. What exactly makes it 10nm? There're no measurements that are 10nm in that process. Density, perf/watt, and node size are certainly similar to TSMC 7nm. So what is 10nm? Do you correct articles on Zen Mobile 6000 and say "Acktually, TSMC 6nm is TSMC's 7nm process that just has some improvements"?
What Intel used to call 10nm Enhanced Superfin, which is what Sapphire Rapids and Alder Lake use, is now called Intel 7. The previous labels Intel used included "nm" after the number, even though no significant structure within the transistor actually matched that nm rating. Intel's current naming scheme no longer includes "nm" after the number. This article originally called Sapphire Rapid's node Intel 7nm, which is something Intel has never called it. It was 10nm, it is now Intel 7, it was never Intel 7nm.
 
observation - over the last 3-4 years all i see is Intel copy AMD in quite a few things. First they bag them for it, then they copy it! Intel are no longer the technology leader ! Latest they are trying to copy the Ryzen 3d v cache!
Everything AMD has implemented in their designs in the last 20 years has been taken from some other company doing it earlier, if not first. 3D V-cache as you mentioned, is just another application of chip stacking. And if you're implying that Foveros was what Intel was "copying" from AMD, Intel announced it in 2018 https://www.theverge.com/2018/12/12/18137401/intel-foveros-3d-chip-stacking-10nm-roadmap-future
 

guru7of9

Reputable
Jun 1, 2018
58
7
4,545
Everything AMD has implemented in their designs in the last 20 years has been taken from some other company doing it earlier, if not first. 3D V-cache as you mentioned, is just another application of chip stacking. And if you're implying that Foveros was what Intel was "copying" from AMD, Intel announced it in 2018 https://www.theverge.com/2018/12/12/18137401/intel-foveros-3d-chip-stacking-10nm-roadmap-future
You are missing the point! Intel is the Technology leader and and is trailing the vastly smaller AMD, which is not expected at all!
As far as your link goes where are all the intel products that come with 3d chip stacking ? Dont seem to remember any that have come out ?
 

spongiemaster

Admirable
Dec 12, 2019
2,273
1,277
7,560
You are missing the point! Intel is the Technology leader and and is trailing the vastly smaller AMD, which is not expected at all!
A technology leader in what? CPU design? They still have that. Hardware implementation? They still have that. Process manufacturing? They're one of the few companies that can do 7nm (which despite Intel labeling it as 10nm, the transistor density is comparable to TSMC's 7nm)

Oh, and AMD can't do the last part. They're dependent on another company to do that for them.
 
  • Like
Reactions: shady28

JayNor

Reputable
May 31, 2019
426
85
4,760
Intel's use of EMIB will raise assembly costs, but it is odd that AMD's use of chiplets has been repeatedly sited for the fab yield advantage, while the same advantage is dismissed for the 56 core SPR chips.

EMIB interconnects are also enabling four in-package HBM2E stacks on a version of SPR, adding 32 memory channels that will undoubtedly produce performance gains in ai/hpc applications. Jeff McVeigh referred to its performance as "astonishing" in the sc21 keynote.