Intel's 28 Core 5Ghz CPU is misleading at best

[asus killer]

Intel lost it's freaking mind, this must have been some brilliant idea by the marketing genius not to be outshined by TR2 announcement. I'm sure the engineers were not behind this insane idea. 

[Ezzy-525]

15 hours ago, InertiaSelling said:

GN has just released a video about this. It's called 28 cores of bullshit.

 

Loved Steve's take on this! 

[Peskanova]

well I'm a bit of newbie about tech expo's but I never had seen an old architecture on a expo without telling if its new or not or the sku or anything, just throw this and let people think what they want

[leadeater]

2 minutes ago, mr moose said:

So I don't know how it can be considered cheating or misleading if they are simply demonstrating a 28 core CPU without any real claims of what it is. 

During a press conference talking about current and future products at a trade show about current and future products I think Intel gets zero leeway on lack of disclosure, if this was at a side event at Computex during one of the normal OC events then it wouldn't need to be stated as it's already been done by context of the event at that point in time.

 

Intel spent over 50 minutes presenting and talking about real products coming to market or already on the market then goes in to this:

"I wanted to give you a preview of something we've been working on"

"I would like to present to you a new 28 core processor"

"When we have 28 cores running at 5 Ghz"

 

Then they proceed on more about how you can get the best of both worlds with 5Ghz and how amazing it is clearly implying anyone can achieve this, still no mention of overclocking.

 

"We are going to take that product to market in Q4 this year"

 

So would I be wrong based on the information presented that Intel is going to release a new 28 core 5Ghz CPU in Q4 2018?

 

Misleading doesn't only require giving of information it applies to the withholding of it, to mislead you only have to give a false impression or wrong idea and the news articles speaks volumes as to if this meet the criteria of being able to be called misleading.  

 

Now I'm not blind I can see the pipes running out the back of the computers and I instantly know something is fishy, for that reason and what I know Intel CPUs of that core count are actually capable of doing.

 

And as quoted above they did indeed make real claims, not false but still misleading.

[Thaldor]

11 minutes ago, Blademaster91 said:

Well because that is all it was, a showcase, like a concept car showing it could be done then later like a concept car the tech in some way gets into production. You don't see car companies revealing every detail on something that isn't final yet either.

As @leadeater said, at least car manufacturers come clear that the concept car is just a concept car that will probably never be manufactured, while Intel boasts how this new 28-core beast processor has been in development and will launch in some time.

 

They even went so far that in the showcase ASUS machine is well lit to show the watercooling while phase-cooled Gigabyte machine is showcased with Cinebench and only RGB-leds shine through the sidepanel and is so carefully placed that the phase-cooling lines are not shown to the audience giving an impression that the Gigabyte machine is only watercooled like the ASUS machine. In Pauls video it's even shown that Intel went as far as adding a reservoir to the Gigabyte machine to make it really look like watercooled. Pauls video is also interesting because it clearly shows what that processor is; It's Xeon and 28-cores and base coreclock is 2,7Ghz which would make it quite close to the Xeon Platinum 8180(M) which has 28-cores and 2,5Ghz, 6 memory channels and uses 3647 socket.

 

Oddly Intel also was running the Gigabyte machine with Windows 8, but that can be put aside as showing the best part of the product like you would bin the cream of the CPUs for OC competition or for the showcase, which is completly understandable. But going through a trouble of making the machine show as watercooled (image with a lot of WTF going on), like why to throw in some reservoir, pump and tape some mixed bag of pipes to a phase-cooled PC while also having on a showfloor real watercooled version that performs way worse (Linuses video with ASUS machine and CB15 score ~6100) and still amrket it as "upcoming new CPU" without even mentioning that the showcased one is extreme overclocked.

[asus killer]

55 minutes ago, Ezzy-525 said:

Loved Steve's take on this! 

"will review vacuum for change" 

[mr moose]

1 hour ago, leadeater said:

"We are going to take that product to market in Q4 this year"

 

So would I be wrong based on the information presented that Intel is going to release a new 28 core 5Ghz CPU in Q4 2018?

 

If they have no intention of bringing a 5Ghz 28 core product to market in q4,  then absolutely it is misleading.  Mind you that's an "if" based on a future event.  Assuming they aren't or can't is not misleading on their part, it is an assumption on ours.

[asus killer]

21 minutes ago, VegetableStu said:

*walks down stairs, hits head on low clearance, doesn't skip a word*

i laughed my ass off when he couldn't get is head straight because of the low ceiling. I guess he Airbnb the house from a midget 

[leadeater]

42 minutes ago, mr moose said:

If they have no intention of bringing a 5Ghz 28 core product to market in q4,  then absolutely it is misleading.  Mind you that's an "if" based on a future event.  Assuming they aren't or can't is not misleading on their part, it is an assumption on ours.

The misleading part is the conditions that clock was made under and giving the impression that was a product specification which it is not. They may have an intention of bringing a 28 core HEDT part to market but it will not carry a spec of 5GHz boost for all cores as shown on stage.

 

The presenter pointed directly at the computer and stated they are taking that product to market in Q4 this year. If it's not going to be as shown then you are misleading, unless they sell every CPU with a water chiller then I'll take back the misleading part. The water chiller not shown on stage or mentioned.

 

And it's not an assumption, they can't. It is impossible beyond literal magic to bring a 28 core 5Ghz all core boost stock configuration product to market on LGA3647 on any process node Intel has without extreme cooling, bloody important thing to say.

 

So they are either going to sell a CPU no normal person could actually run, not hinted at ever so misleading on that situation, or they are going to sell it with stock boost clocks lower than 5Ghz so misleading product specifications.

 

You can take your pick on what the product will actually be if ever released but no matter how it plays out that press conference and their demonstration was misleading.

[laminutederire]

12 minutes ago, mr moose said:

If they have no intention of bringing a 5Ghz 28 core product to market in q4,  then absolutely it is misleading.  Mind you that's an "if" based on a future event.  Assuming they aren't or can't is not misleading on their part, it is an assumption on ours.

Except if you say we're going to release it, then showcasing another product with ridiculous cooling to say that the future product will do that on air or normal cooling. It is technically lying. Because either it's the same product theyll launch, and it can't do it on normal cooling otherwise they'd have used it on normal cooling to showcase it. Or it's not then it's lying because it's not the same product they're going to launch.

It'd be like and saying they'll crush Nvidia with a 2.5GHz Vega 64 running on LN2, saying next Vega will be basically that. If they had a sample of Vega which could normally do that on air they'd have shown it...

[WereCat]

6 hours ago, asus killer said:

will the chillers have RGB? and will they sync with AURA?

 

6 hours ago, GoldenLag said:

can you chill with the RGB? RGB isnt that Cool

yes, but you must set it to BLUE, because RED is performance and GREEN is efficiency.

[Taf the Ghost]

1 hour ago, leadeater said:

The misleading part is the conditions that clock was made under and giving the impression that was a product specification which it is not. They may have an intention of bringing a 28 core HEDT part to market but it will not carry a spec of 5GHz boost for all cores as shown on stage.

 

The presenter pointed directly at the computer and stated they are taking that product to market in Q4 this year. If it's not going to be as shown then you are misleading, unless they sell every CPU with a water chiller then I'll take back the misleading part. The water chiller not shown on stage or mentioned.

 

And it's not an assumption, they can't. It is impossible beyond literal magic to bring a 28 core 5Ghz all core boost stock configuration product to market on LGA3647 on any process node Intel has without extreme cooling, bloody important thing to say.

 

So they are either going to sell a CPU no normal person could actually run, not hinted at ever so misleading on that situation, or they are going to sell it with stock boost clocks lower than 5Ghz so misleading product specifications.

 

You can take your pick on what the product will actually be if ever released but no matter how it plays out that press conference and their demonstration was misleading.

Der8aer had a demonstration of a i7-8086k running at 7.1 Ghz under LN2. No one would think that's a normal speed or use case. Intel pretty much did that for a product demonstration. 

 

While we're on the topic of Intel's next "HEDT" platform, how is this even a "Desktop" part? The ASUS board barely fit in a mATX form factor. 

[Carclis]

Assuming people even did this, would it even be possible for somebody in the US to run this type of setup? I mean the chiller is rated at 1200w and the CPU is almost certainly around 1000w, perhaps more... I feel like that would probably trip a circuit breaker.

[Taf the Ghost]

2 minutes ago, Carclis said:

Assuming people even did this, would it even be possible for somebody in the US to run this type of setup? I mean the chiller is rated at 1200w and the CPU is almost certainly around 1000w, perhaps more... I feel like that would probably trip a circuit breaker.

You'd probably need to replace a circuit breaker to handle it in many setups, but a decent 220V setup should be able to handle the load without much issue. It's just no one would bother.

[Misanthrope]

3 minutes ago, Carclis said:

Assuming people even did this, would it even be possible for somebody in the US to run this type of setup? I mean the chiller is rated at 1200w and the CPU is almost certainly around 1000w, perhaps more... I feel like that would probably trip a circuit breaker.

I think you can: I've seen houses with several old fashion window AC units and they manage.

 

However they do have a 220v circuit (I think that'd be 240v in America) but you could technically pay as much for a computer that manages to give you a couple percentages point extra on Cinebench in exchange for what you'd pay for cooling a fairly large house during a very hot summer in Texas.

[leadeater]

1 hour ago, Misanthrope said:

I think you can: I've seen houses with several old fashion window AC units and they manage.

 

However they do have a 220v circuit (I think that'd be 240v in America) but you could technically pay as much for a computer that manages to give you a couple percentages point extra on Cinebench in exchange for what you'd pay for cooling a fairly large house during a very hot summer in Texas.

Combined phases in US is 220V or 240V (they for what ever reason have 110V and 120V standards, weird), most other countries are 230V standard single phase but allow for 220V-240V due to a few that use those.

 

As for AC units they are not wired in to normal outlets (not ones up to electrical code anyway) unless they are low rated or portable ones. A standard wall mounted AC unit is 230V AC on a 15A socket.

 

Regular 230V 10A wall outlet will do 2300W, so the 28 core Intel system shown off would probably be unsafe on a single outlet however you'd just use two, so long as the circuit breaker is 16A and not 10A which is pretty standard.

 

US 120V 15A wall outlet will do 1800W (1650W if 110V) so even two outlets of those would work.

[Drak3]

1 hour ago, Carclis said:

Assuming people even did this, would it even be possible for somebody in the US to run this type of setup? I mean the chiller is rated at 1200w and the CPU is almost certainly around 1000w, perhaps more... I feel like that would probably trip a circuit breaker.

Typical house circuit runs 120v 15a to 20a, or 1800w to 2400w.

 

It's definitely doable, but it'll take multiple circuits.

[Fetzie]

8 hours ago, VegetableStu said:

(meanwhile, next door)

 

"sir, why again, did you say you needed a 480v residential hookup?"

"Electric cooling."

https://pcgamesn.com/cooler-master-thermoelectric-peltier-cooler

I'd hook the hot side of the heat exchanger up to my central heating system, might as well use the heat it produces.

[Carclis]

15 minutes ago, Drak3 said:

Typical house circuit runs 120v 15a to 20a, or 1800w to 2400w.

 

It's definitely doable, but it'll take multiple circuits.

That's pretty much what I was expecting. A stupid workaround for a stupid setup I suppose.

[vanished]

Honestly I wouldn't say it was a feeble attempt; I think it did the job pretty well.  Sure, it was shortly after revealed to be a bit of a sham, but at the time this got all the attention, and tbh I still haven't really seen anything about the new threadripper other than that it'll have 32 cores.  Sure the truth always comes out eventually and that leads people to be interested in the right thing, but in the mean time, this did do a great job at getting attention.

[BroliviaWilde]

Shocking Interview with Intel Engineer about 28-Core 5 GHz CPU

 

 

[S w a t s o n]

58 minutes ago, BroliviaWilde said:

Shocking Interview with Intel Engineer about 28-Core 5 GHz CPU

 

 

The true hero this thread needed

[pas008]

really?

crying about misleading cooling?

we all know all companies do this

 

intel amd and nvidia have mislead us

even lie

 

oh fury the overclockers dream

how about the volta killer

yes I named amd because I like to trigger  all fanboys and this thread seems full of them

 

 

 

[GoldenLag]

3 minutes ago, pas008 said:

really?

crying about misleading cooling?

we all know all companies do this

 

intel amd and nvidia have mislead us

even lie

 

oh fury the overclockers dream

how about the volta killer

yes I named amd because I like to trigger  all fanboys and this thread seems full of them

 

 

 

within the context and without the context intel failed to provide it makes sence to wine about misleading people. especially since this lead to many mainstream tech presses to misslead their readers due to lack of information.

[S w a t s o n]

6 minutes ago, pas008 said:

really?

crying about misleading cooling?

we all know all companies do this

 

intel amd and nvidia have mislead us

even lie

 

oh fury the overclockers dream

how about the volta killer

yes I named amd because I like to trigger  all fanboys and this thread seems full of them

it's not misleading cooling, it's a misleading product pitch as a whole.

"Poor Volta" was never a real thing. It's a fan created meme based on a poster in the background of a youtube video. Not quite the same thing as saying "Oh we have this 28 core cpu running at 5ghz, and btw available Q4."

All companies do misleading things but you can bet if AMD showed off a 5ghz, 32 core threadripper using a fucking water chiller and said "Available Q4", they would be getting shit on too.