Monday, September 18th 2017

Intel to Bring 8-core/16-thread CPUs to the Mainstream Desktop Platform in 2018

Having hit a wall with fab process-assisted micro-architecture advancements, and facing an unexpectedly competitive AMD, there's only one direction left for Intel's product development over generations, core-counts. The company is on the verge of introducing 6-core/12-thread "Coffee Lake" processors to the mainstream-desktop (MSDT) platform in a few weeks from now; and if leaks by a popular BTO desktop-replacement/mobile-workstation manufacturer Eurocom is to be believed, the company could increase the core counts a second time in 2018, by introducing an 8-core/16-thread MSDT part.

A Eurocom representative, posting on NotebookReview forums, hinted at the possibility that the upcoming Intel Z390 Express chipset, which hits motherboards in 2018, could exclusively support 8-core/16-thread processors, which come out in the second half of 2018. The representative revealed this in context of the company skipping the Z370 Express chipset, as it lacks support for those upcoming 8-core/16-thread chips. In addition to support for new processors and possibly next-generation "Ice Lake" processors, the Z390 chipset adds several new features over the Z370, including a better onboard audio solution, integrated WLAN, and SDIO controller.
Sources: NotebookReview, VideoCardz
Add your own comment

54 Comments on Intel to Bring 8-core/16-thread CPUs to the Mainstream Desktop Platform in 2018

#26
dj-electric
Alright, i get it than.

Now, i suspect that H2 of 2018 might turn into october-december. That's quite a long wait. A year actually.

I refuse to believe that intel will make a new series obsolete after less than 12 months (please exclude the i7 7800X, its a miserable excuse of a CPU)
Posted on Reply
#27
Ebo
AS I see it, Intel slept in the hour and regarded AMD as the underdog, then came Ryzen and after TR CPU's. Now they need to do something, and so far it hasent changed anything, AMD still has the momentum and the upperhand in the server market which is where the money really is.

Right now its like a panic selution opon another without a clear way of moving forward and give people new tech, it feels like giving "grandpaar" a viagra, get hard and move on.

This is from an Intel owner.
Posted on Reply
#28
dwade
Intel panicking. LOL just stop. A panicking Intel would be selling their 18 cores for $999. All I see is AMD selling their half-baked 16 cores for $999 because it can't compete with their best.
Posted on Reply
#29
R0H1T
dwadeMost people only want AMD to be on Intel and Nvidia's radar so they can buy their stuff for cheaper.
Yeah I hate them, then there are those who pretend that without AMD Intel (or Nvidia) would shower with mainstream hexa, octa ,deca, dodeca cores without charging both kidneys & a full grown liver :rolleyes:
dwadeIntel panicking. LOL just stop. A panicking Intel would be selling their 18 cores for $999. All I see is AMD selling their half-baked 16 cores for $999 because it can't compete with their best.
OR release an 18 core with TIM slightly better than toothpaste so it's either being rushed (by the competition) or being greedy as f*** to not care about your top end customers, I'd say both o_O
Posted on Reply
#30
Vya Domus
dwadeAll I see is AMD selling their half-baked 16 cores for $999 because it can't compete with their best.
You might want to check upon that once more , that "half-baked 16 core CPU" is currently the most powerful consumer level CPU money can buy.
Posted on Reply
#31
dwade
Vya DomusYou might want to check upon that once more , that "half-baked 16 core CPU" is currently the most powerful consumer level CPU money can buy.
1950x sacrifices client PC performance in order to cater to a niche audience, where a 1600x is faster than a 1950x in gaming. You'll see why they force you to use Game Mode (lol) just to somewhat close the gap between their weaker parts. A 7920x is also better all around so calling the 1950x the "most powerful consumer level CPU" is a fallacy. It's a half-dead server chip with compromising client PC performance.
Posted on Reply
#32
Vya Domus
You're being a tad ignorant to say the least.

TR and Skylake-X are both "half dead server chips" and also they aren't made for gaming. A 1600X is faster than a 1950X in games the same way a 7700K is faster than a 7900X.

These are workstation class CPUs for HEDT platforms , that IS a niche market.

To me the fact that you brought up the matter of gaming performance tells me how out of context you are. In my opinion the only things that are fallacies are your points.

Also , what the hell is client PC performance in this context?
Posted on Reply
#33
Parn
The need to change motherboard for a new CPU is getting more and more frequent. This is a new strategy of milking the consumers by Intel.

During the Core 2 era we had 3 and 4 series chipsets capable of supporting all existing LGA775 CPUs. Then moving onto the Core ix generations the chipsets supported at least two generations of CPUs with some nice new features added every update such as USB3, PCIe 3.0, DMI 3.0 and M.2/U.2 NVMe. Now what do we have? A rebadged Z270 as Z370 that would only work for less than a year...
Posted on Reply
#34
dwade
Vya DomusYou're being a tad ignorant to say the least.

TR and Skylake-X are both "half dead server chips" and also they aren't made for gaming. A 1600X is faster than a 1950X in games the same way a 7700K is faster than a 7900X.

These are workstation class CPUs for HEDT platforms , that IS a niche market.

To me the fact that you brought up the matter of gaming performance tells me how out of context you are. In my opinion the only things that are fallacies are your points.

Also , what the hell is client PC performance in this context?
Seems u dont even know what ur talking about. You said best consumer CPU money can buy. That's already a lie because 1950x isn't anywhere close to that.
Posted on Reply
#35
Vya Domus
dwadeSeems u dont even know what ur talking about. You said best consumer CPU money can buy. That's already a lie because 1950x isn't anywhere close to that.
Show me some benchmarks so you can back up your claim that it "isn't anywhere close".



You are telling me a 7920X with just 2 extra cores not only catches up to the 1950X but out performs it by far , "all around" ? Oh boy :laugh:

Also , you haven't answered me , what is "client PC performance" ? I am really curious to have someone that clearly knows a lot to explain to me what that is. I mean you just made that metric up. Maybe you have some experimental testing methodology which we don't know about.

I have a hunch you are a massive troll with no knowledge on the subject. But go ahead , surprise me.
Posted on Reply
#36
evernessince
This is right around when Zen 2 will release so it's no wonder Intel will have to up their core count.

en.wikipedia.org/wiki/Zen_2

AMD will likely match or exceed Intel's IPC and with the transition to 7nm it will also increase it's lead in power efficiency. Don't know what AMD will do with the extra die space though, perhaps increase the spacing or robustness of the logic on the chip so they can get higher frequencies. That would really be the trifecta, as they could match Intel's frequencies with better power consumption and better IPC.

The question then becomes, is Intel going to introduce IPC improvements? Can they even squeeze more out of their current architecture? How does Intel propose cooling these higher core count CPUs? Their current stock coolers are not enough. Are they finally going to move back to using solder?
Posted on Reply
#37
HisDivineOrder
evernessinceThis is right around when Zen 2 will release so it's no wonder Intel will have to up their core count.

en.wikipedia.org/wiki/Zen_2

AMD will likely match or exceed Intel's IPC and with the transition to 7nm it will also increase it's lead in power efficiency. Don't know what AMD will do with the extra die space though, perhaps increase the spacing or robustness of the logic on the chip so they can get higher frequencies. That would really be the trifecta, as they could match Intel's frequencies with better power consumption and better IPC.

The question then becomes, is Intel going to introduce IPC improvements? Can they even squeeze more out of their current architecture? How does Intel propose cooling these higher core count CPUs? Their current stock coolers are not enough. Are they finally going to move back to using solder?
I think you're dreaming if you think AMD is going to increase the IPC again in a significant way just after releasing a new architecture. I expect minor gains, improved memory compatibility, and enhanced clockspeed possibly. Possibly.

All of which would be awesome, by the by. I welcome that happening. Thing is, I don't believe that last one (clockspeed) is as likely as you think because I know that most of it relies upon the fab doing it and AMD has a known history of being screwed by new processes and fab's in general.
Posted on Reply
#38
Dave65
ZoneDymoThank you AMD.
Exactly what I was thinking!
Posted on Reply
#39
Scrizz
Vya DomusShow me some benchmarks so you can back up your claim that it "isn't anywhere close".



You are telling me a 7920X with just 2 extra cores not only catches up to the 1950X but out performs it by far , "all around" ? Oh boy :laugh:

Also , you haven't answered me , what is "client PC performance" ? I am really curious to have someone that clearly knows a lot to explain to me what that is. I mean you just made that metric up. Maybe you have some experimental testing methodology which we don't know about.

I have a hunch you are a massive troll with no knowledge on the subject. But go ahead , surprise me.
If we're going to start throwing in Over-clocked results,



:nutkick::toast:
Posted on Reply
#40
Vya Domus
Take a better look at that chart , you can find stock results too.
Posted on Reply
#41
btarunr
Editor & Senior Moderator
ScrizzIf we're going to start throwing in Over-clocked results,



:nutkick::toast:
Those are very good numbers for the 1950X TBH.
Posted on Reply
#42
evernessince
HisDivineOrderI think you're dreaming if you think AMD is going to increase the IPC again in a significant way just after releasing a new architecture. I expect minor gains, improved memory compatibility, and enhanced clockspeed possibly. Possibly.

All of which would be awesome, by the by. I welcome that happening. Thing is, I don't believe that last one (clockspeed) is as likely as you think because I know that most of it relies upon the fab doing it and AMD has a known history of being screwed by new processes and fab's in general.
AMD in 8% behind in IPC. Lisa sue already said they expect a 15% increase in IPC with Zen 2. I can put 2 and 2 together thank you very much.

"I think you're dreaming if you think AMD is going to increase the IPC again in a significant way just after releasing a new architecture."

Um, that's exactly when you'll see the biggest gain. AMD themselves have already stated that it will be easy to fix the low hanging fruit in Zen. Remember Intel's first i7? then remember the 2000 series? Yeah, that was a huge release.
Posted on Reply
#43
R0H1T
HisDivineOrderI think you're dreaming if you think AMD is going to increase the IPC again in a significant way just after releasing a new architecture. I expect minor gains, improved memory compatibility, and enhanced clockspeed possibly. Possibly.

All of which would be awesome, by the by. I welcome that happening. Thing is, I don't believe that last one (clockspeed) is as likely as you think because I know that most of it relies upon the fab doing it and AMD has a known history of being screwed by new processes and fab's in general.
And what would you define as "significant way" btw? AMD exceeded their IPC target for Zen, they also exceeded the initial clock speed targets that many were expecting, although the OC headroom is pretty low.
The next iteration should be much better, there's very little reason to be so pessimistic.
evernessinceAMD in 8% behind in IPC. Lisa sue already said they expect a 15% increase in IPC with Zen 2. I can put 2 and 2 together thank you very much.

"I think you're dreaming if you think AMD is going to increase the IPC again in a significant way just after releasing a new architecture."

Um, that's exactly when you'll see the biggest gain. AMD themselves have already stated that it will be easy to fix the low hanging fruit in Zen. Remember Intel's first i7? then remember the 2000 series? Yeah, that was a huge release.
The second (major) iteration was Nehalem not SB, both of them derived from the same conroe. The large gains in Nehalem were due to HT & imc(?) while SB was also iteratively better due to clocks & further tweaking of the design.
Agree with the rest though, Zen 2 or even Zen+ (RR for desktop & laptops) should be better than the initial Zen chips.
Posted on Reply
#44
DeathtoGnomes
ScrizzIf we're going to start throwing in Over-clocked results,



:nutkick::toast:
where is this chart from? some trashcan?
Posted on Reply
#45
dj-electric
This is actually from the official results of the overclocking world league, HWbot.
Posted on Reply
#46
DeathtoGnomes
Dj-ElectriCThis is actually from the official results of the overclocking world league, HWbot.
Yea no way I could tell that without a link.
Posted on Reply
#47
GAR
ZoneDymoThank you AMD.
Yup, lazy Intel.....Milking the stupid 4 core 8 thread cpu for 6 years now......All my systems as of now are RyZen, no more intel for me.
Posted on Reply
#48
TheGuruStud
The stock results must be with some slow ass ram. I've already seen the 3k+ scores with decent ram. 2,700. ha.
Posted on Reply
#49
Scrizz
DeathtoGnomeswhere is this chart from? some trashcan?
The same place Anadtech pulled the 4122 TR score.....
hwbot.org/
Posted on Reply
#50
DeathtoGnomes
ScrizzThe same place Anadtech pulled the 4122 TR score.....
hwbot.org/
See thats why posting source links are important here, without them it just looks like another troll post. :shadedshu:
Posted on Reply
Add your own comment
Oct 18th, 2024 18:54 EDT change timezone

New Forum Posts

Popular Reviews

Controversial News Posts