Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.

koyoot

macrumors 603
Jun 5, 2012
5,939
1,853

Also a quote on them about previous news they did, when they were going to drop Intel in favor of ARM:
"Readers of our blog might remember our excitement around ARM processors. We even ported the entirety of our software stack to run on ARM, just as it does with x86, and have been maintaining that ever since even though it calls for slightly more work for our software engineering teams. We did this leading up to the launch of Qualcomm’s Centriq server CPU, which eventually got shuttered. While none of the off-the-shelf ARM CPUs available this moment are interesting to us, we remain optimistic about high core count offerings launching in 2020 and beyond, and look forward to a day when our servers are a mix of x86 (Intel and AMD) and ARM.

I think this pretty much sums current state of technology. ARM may be future, and in specific scenarios, but looking at x86-64 next gen products(AMD Milan) its fair to say that both architectures will work in tandem, not be a replacement for each other.

However, I wish ARM products will start offering something genuinely new in terms of performance.

I would love to see DIY products, with High-Performance, just to tinker around with them.
 

deconstruct60

macrumors G5
Mar 10, 2009
12,493
4,053
....
Guess which fruity-sounding OEM beginning with A will most definitely NOT be passing on that saving to the consumer (when the W series inevitably drops as well).

:D

The Xeon W 2200 series already saw a price cut. The bigger issue there is whether/when Apple will actually do something with those. Pretty likely that won't see an overall system price reduction for the iMac Pro but a core count bump up. ( e.g., start at 10 cores instead of 8). That would passing along value gain more so than price cut.

The Xeon W 3200 series in the Mac Pro ... probably not going to see a price cut any time soon. Intel brought these Xeon SP adjustments as a binning additions to that line up. The W 3200 series is already binned pretty high. Don't typically get multiple adjustments in this space with this few variants in the line up. If they have a W 3300 series then perhaps there. ( or if the W 3300 has a release slide out problem. I suspect it is easier though to get "Ice Lake" variants to cap out at 26-28 cores than it is for higher core counts targets. SP they have bigger volume release problems, so bigger cuts now. )

[ The only corner case for the Mac Pro would be the $3K , "> 1TB" tax on the top two CPU core counts. There was an alternative models that Apple skipped that didn't have the tax. if Intel deletes those and reprices the "max memory" models without the tax then perhaps. That won't bring any correction for the majority of the Mac Pro line up though. If Intel slaps a new product number on them that would be the 'fig leaf' Apple would need to change to the price.

Threadripper itself isn't a major threat to the "> 1TB" Xeon W 3200 series for folks who actually need that level of RAM capacity. It can't. The larger problem for Apple is that there is already an Intel option that doesn't have the tax. It is the subsegment who only need 500-700 GB (or less) coupled to the 20+ core count where Apple's choice is already a problem and Threadripper alternatives just make it even less competitive. Intel really doesn't need two options with same core count and clocks, just differentiated on max Memory.

I wouldn't be surprised though for Apple to keep their ~$900 markup the max mem tax brought then even after a move to a cheaper SKU, but pass along the $3K "savings'. ]



]
 

AustinIllini

macrumors G5
Oct 20, 2011
12,699
10,567
Austin, TX
On the smaller machines, the biggest threat to Intel on Mac are the A-series chips.

I would be far more worried about that if I were Intel than threadripper.
 

Ulfric

macrumors regular
Apr 4, 2018
160
124

After seeing those results I may change my stance about current state of ARM, and start believing in it, again ;).

Now I want something for consumer market with DIY, off the shelf parts, and software to play with it.

If you really wanna see the real capability of ARM Chips for Exascale & HPC workloads then you will find A64FX from Fujitsu quite interesting; BTW they are already operational.

 

koyoot

macrumors 603
Jun 5, 2012
5,939
1,853
If you really wanna see the real capability of ARM Chips for Exascale & HPC workloads then you will find A64FX from Fujitsu quite interesting; BTW they are already operational.

What I want to see is mITX boards with CPU(with RAM soldered also, if it would have sufficient bandwidth) soldered in them, and with PCIe 4.0 sockets, and M.2 for DIY markets, that I can take off the shelf, put into my computer, connect dGPU, my own SSD, and install all of the software I want to use, and that includes games.

Only then I will consider ARM a viable option, 100%. So far it might still be nieche.
 

Mago

macrumors 68030
Aug 16, 2011
2,789
912
Beyond the Thunderdome

After seeing those results I may change my stance about current state of ARM, and start believing in it, again ;).

Now I want something for consumer market with DIY, off the shelf parts, and software to play with it.
Gcc vs gcc o gcc vs llvm no avx/avx2 when they comeback comparing vs avx2 then I'll consider it.

Meanwhile it's an very good target platform for non optimized binaries (like in GoLang)
 

mattspace

macrumors 68040
Jun 5, 2013
3,344
2,975
Australia
What I want to see is mITX boards with CPU(with RAM soldered also, if it would have sufficient bandwidth) soldered in them,

... why?

soldered in them

seriously... why?


dear god, why?

image.png
 
  • Like
Reactions: seek3r

koyoot

macrumors 603
Jun 5, 2012
5,939
1,853

Apparently Intel just killed off 10 nm server chips.

7 nm process will not come before 2021 for mobile chips, and 2022 for server chips.

Anyone else wants to argue that sticking to Intel is good idea?
 

Kpjoslee

macrumors 6502
Sep 11, 2007
417
269

Apparently Intel just killed off 10 nm server chips.

7 nm process will not come before 2021 for mobile chips, and 2022 for server chips.

Anyone else wants to argue that sticking to Intel is good idea?

Intel has just canceled a major server program. This one is a major failure and a crushing retreat from their main market. ====> Apparently Intel just killed off 10 nm server chips.

???

Premature conclusion from vague statement that lacks details. Considering Charlie usually comes off with strong clickbait title and what happens after is quite different from what his title implies, every time. I wouldn't trust it as it is.
Wait until actual news comes out. Semiaccurate is not a valid news source, that is a fact.
 
Last edited:

koyoot

macrumors 603
Jun 5, 2012
5,939
1,853
Intel has just canceled a major server program. This one is a major failure and a crushing retreat from their main market. ====> Apparently Intel just killed off 10 nm server chips.

???

Premature conclusion from vague statement that lacks details. Considering Charlie usually comes off with strong clickbait title and what happens after is quite different from what his title implies, every time. I wouldn't trust it as it is.
Wait until actual news comes out. Semiaccurate is not a valid news source, that is a fact.
You have to pay 100$ for student access, and 1000$ for professional access, per year. This is paywalled under 1000$, professional access.

Access that high-level people from around the industry buy.

Its not premature conclusion. It is actually in this analysis, that is behind the paywall. You may not trust this. But his moles when it goes to Intel are pretty accurate.

I guess he was dead wrong about 10 nm process, all along, from 2013, when he first started saying that it won't happen.

P.S. I always like when people do not consider SemiAccurate not good source, when the information they provide they do not like ;).

Connection to Threadripper and the Mac Pro? None.
Well, considering that for at least 2-3 years there is no upgrade path for Mac Pro, and next year, everybody will have brand New EPYC CPUs in their workstations because Intel will be still on 14 nn process, what is the point of sticking to Intel for next upgrade?

Especially when from competition you will be able to get more for less?
 
  • Like
Reactions: ssgbryan and Ulfric

Kpjoslee

macrumors 6502
Sep 11, 2007
417
269
You have to pay 100$ for student access, and 1000$ for professional access, per year. This is paywalled under 1000$, professional access.

Access that high-level people from around the industry buy.

Its not premature conclusion. It is actually in this analysis, that is behind the paywall. You may not trust this. But his moles when it goes to Intel are pretty accurate.

I guess he was dead wrong about 10 nm process, all along, from 2013, when he first started saying that it won't happen.

P.S. I always like when people do not consider SemiAccurate not good source, when the information they provide they do not like ;).

So basically, you consider Semiaccurate as a good source because you like the information they provide. ?
Come on buddy, that is like relying on tabloids for accurate news.

Not worth talking about it further until actual news with more substance comes out.
 

koyoot

macrumors 603
Jun 5, 2012
5,939
1,853
So basically, you consider Semiaccurate as a good source because you like the information they provide. ?
Come on buddy, that is like relying on tabloids for accurate news.

Not worth talking about it further until actual news with more substance comes out.
No, I consider SemiAccurate good source, because more often then not Charlie is dead right with his info, and analysis.

Have you taken out that 1000 bucks from your pockets for the article, already? If not, you won't get the news.
 
  • Like
Reactions: Ulfric

Kpjoslee

macrumors 6502
Sep 11, 2007
417
269
No, I consider SemiAccurate good source, because more often then not Charlie is dead right with his info, and analysis.

Have you taken out that 1000 bucks from your pockets for the article, already? If not, you won't get the news.

Not gonna try to convince you on that. I know it is hard to convince people to snap out of the cult.
My condolences if you actually paid $1000 for "professional" sub.
 

koyoot

macrumors 603
Jun 5, 2012
5,939
1,853
Not gonna try to convince you on that. I know it is hard to convince people to snap out of the cult.
My condolences if you actually paid $1000 for "professional" sub.
Tell that to people like Bob Swan(Intel CEO).

Those who pay 1000$ per year are people who are heavily investing in tech stocks.
 
  • Like
Reactions: Ulfric

Kpjoslee

macrumors 6502
Sep 11, 2007
417
269
Tell that to people like Bob Swan(Intel CEO).

Those who pay 1000$ per year are people who are heavily investing in tech stocks.

Even more reason to not to trust it.
Well, true or not true, we will find out in a few.
Stay safe from coronavirus!?
 

Ulfric

macrumors regular
Apr 4, 2018
160
124
For Anyone investing in Semiconductor stocks (NVDA, INTC, AMD, Micron) Semiacurate is a great source for analysis & that $1000 or $100 for students are well worth it.

Charlie predicted the problem Intl will face with 10 nm, Charlie was the first who said Intel had nothing to go after AMD in HPC sector. He was the first one stating Intel is trying to force out from Modem business. When it comes to Intel he was often times right.
 

koyoot

macrumors 603
Jun 5, 2012
5,939
1,853

Kpjoslee

macrumors 6502
Sep 11, 2007
417
269

Two follow-ups to latest news.

I don't want to spoil anything so I will write only this: its worse than you think.


This is the real scoop. Now stop spreading the fud without actually mentioning the real story or linking the news people can't read without subscription.

P.S. It is not Intel killing off 10nm server chips. ;)
 
  • Like
Reactions: AidenShaw

Mago

macrumors 68030
Aug 16, 2011
2,789
912
Beyond the Thunderdome
Marvell (no that who publish comics and so-so movies), announced a 96 core 240w ThunderX3 (they bought cavium), I'll normally won't give much attention as there are a bunch of ARM processors for Servers announced or in development which looks promising but still seems half-baked, this ThunderX3 while still seems half baked (considering it's memory interface and # of pcie lines) it incorporate something I didn't see in decades besides in GPU: smt-4, and while it's not that efficient as expected (threaded performance gain while meaningful, not really impressive) I consider it's an meaningful development as smt-4 could be an good antidote to the lack of simd(avx) instructions so it's performance comes very close to that of avx-512 optimized code an a plus is it is done in gcc or STD llvm Something which helps languages that doesn't support simd yet as go and kotlin native.
 

deconstruct60

macrumors G5
Mar 10, 2009
12,493
4,053

This is the real scoop. Now stop spreading the fud without actually mentioning the real story or linking the news people can't read without subscription.

P.S. It is not Intel killing off 10nm server chips. ;)

from the article.

"....
  • Intel remains on track for delivery of 10nm Ice Lake CPUs later this year.
...."

Yeah, illustrates that while Semiaccurate's "raw" info may be mostly good , the analysis ( "fitting info into big picture' , scope , ramifications ) are quite often Charlie seeing what he wants to see as opposed to what is really there. He'll probably still be beating the "10nm can't possibly work at all with any possible adjustment "drum well into 2021-22. ... in part because he loves saying "I told you so ... I was right on date xx.yy.zz " .

In the Xeon W space, there is not necessarily a good reason why Intel had to chase AMD into the double die , super large package zone. If Intel went with two 14 cores , 10nm dies they'd still be able to field something with 28 core range that the Mac Pro has now (and gets real work done for people) and keep the 10nm yield rates up. How the Xeon W Ice Lake are priced probably matters just as much to Apple as the the max core count available. If the yield held up could do 28 cores on single die like they do now and use about the same socket as they do now.

Two socket CooperLake was going to be a huge overlap with two socket ice Lake for a more than reasonable number of core counts. Neither one was probably going to top AMD's max core count per socket. But as been pointed out macOS isn't particularly a super high demand zone for absolute maximum core count as the sole primary selection driver. Cooper Lake int the two socket models was a hedge on not figuring out how to adjust well enough on the ice Lake iteration. If they don't need it because the adjust worked reasonably well then don't need that platform.


Intel probably will sell fewer server chips, but utter and complete collapse? ... probably not. They are probably going to step back from the sell everything to everybody approach. They still cover a large and broad spectrum of solutions.

The other factor is that these all these > 2 socket systems are shrinking.


https://www.nextplatform.com/2019/04/24/why-single-socket-servers-could-rule-the-future/


There is increasingly a narrow group that uses them. Google / FB / Baidu / etc that have low latency network data input ML inference workloads that need bfloat16 ... Intel is more than competitive with all of AMD's current offerings. Decent chance next ones too once weave in software stack optimizations. So Intel left the platform for the group that asked for it in the first place. Probably not going to loose most of those customers for that workload.

Did Intel ever think that Cooper Lake was going to be super competitive outside of the folks focused on bfloat16 . "Analysis" that presumes that seems a bit whacked. It was never designed in the first place to be primary option outside of that space for single/dual sockets workload since not much different than the cascade lake solutions.
 
Last edited:
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.