Editorial Intel's 28-core HEDT Processor a Panic Reaction to 32-core Threadripper

[Vya Domus]

Interesting how no one calls out the media and news sites for posting brain dead articles and headlines such as "Intel's new amazing 5Ghz 28 core CPU". Intel , AMD , Nvidia do all they can to put on a show but ultimately it's the press that manages to exalt and validate their retarded presentations and claims. Even Anandtech reported this piece of news exactly the same way , I typically had higher expectations of them as they put out very in depth reviews where they get extremely technical.

I know most of these sites knew very well this was a BS presentation but went for it regardless as to present dramatic headlines to draw the attention of people but nonetheless I am reminded once again of their quality.

No offense but anyone who believed that the cpu was not overclocked to the max on stage is mentally challenged.

Well , that must include the press as well.

 

[GlacierNine]

Interesting how no one calls out the media and news sites for posting brain dead articles and headlines such as "Intel's new amazing 5Ghz 28 core CPU". .

I addressed this before - It's in Anandtech and Tom's interests, to run the most attention-grabbing headline based on the info they're given. Being more critical of announcements is often left on the back burner in favour of driving traffic, aka the lifeblood of the site.

 

[Hood]

I hope this is nothing less than pure sarcasm... otherwise, I hope you enjoy your kool-aid... you know, to help you kool your parts after Intel is done with them

Yes, it's sarcasm. It was in response to eidairaman1's comment about dishonesty, because why expect honesty from a corporation? Lying for dollars is pretty much their job description, whether they make PC hardware or underarm deodorant. It's all about saying whatever separates customers from cash, just short of actionable lawsuits. It's one of the main flaws in the free enterprise system - too much honesty will cause you to lose money, while your competitors lie their asses off and take it to the bank. This leads people to the mindset that lying is a virtue - it results in more cash, how can that be a bad thing?

 

[btarunr]

Updated as an editorial.

 

[champsilva]

The real question is

Why someone believe this wasnt overclocked?

Was very clear 2.7ghz in the cinebench

Also 5912 is probably some WC overclock.

 

[GlacierNine]

The real question is

Why someone believe this wasnt overclocked?

Was very clear 2.7ghz in the cinebench

Also 5912 is probably some WC overclock.

Doesn't Cinebench report the base clock in that area, not the boost clock? I could have sworn my 6700K shows as a 4GHz part in cinebench.

It would also make a bit of sense to doubt *that* number, given that the 8180 has a 2.5GHz base clock, which makes the base clock reported there, only 200MHz more than the server part, and also 800MHz *lower* than the 8180's boost clock of 3.5GHz.

 

[bug]

Interesting how no one calls out the media and news sites for posting brain dead articles and headlines such as "Intel's new amazing 5Ghz 28 core CPU". Intel , AMD , Nvidia do all they can to put on a show but ultimately it's the press that manages to exalt and validate their retarded presentations and claims. Even Anandtech reported this piece of news exactly the same way , I typically had higher expectations of them as they put out very in depth reviews where they get extremely technical.

I know most of these sites knew very well this was a BS presentation but went for it regardless as to present dramatic headlines to draw the attention of people but nonetheless I am reminded once again of their quality.



Well , that must include the press as well.

Go subscribe to every site you read, then they won't need to use clickbait and tabloid-style articles

 

[efikkan]

Come on guys, this isn't even newsworthy. Anyone who claims Intel "mislead" anyone is just clearly showing how low they will sink, just to take any blow at Intel. There is no substance in this story, this is just another made up hysteria. Grow up!

Take a look in the mirror, and imagine if the shoe was on the other foot. Where were all of you when AMD obviously lied to us by claiming Polaris would improve efficiency by 2.5× and be twice as efficient as Maxwell… I could have dug up another couple dozen examples, but everyone should get the point; It's time to get some perspective and stop complaining about everything AMD's competition does.

 

[jabbadap]

The real question is

Why someone believe this wasnt overclocked?

Was very clear 2.7ghz in the cinebench

Also 5912 is probably some WC overclock.

Probably yeah, Linus ran Cinebench r15 on Rog Dominus at the Computex(cpu score 6100), which was with custom water loop.

Spoiler: Linustechtips@computex

 

[GlacierNine]

Come on guys, this isn't even newsworthy. Anyone who claims Intel "mislead" anyone is just clearly showing how low they will sink, just to take any blow at Intel. There is no substance in this story, this is just another made up hysteria. Grow up!

Take a look in the mirror, and imagine if the shoe was on the other foot. Where were all of you when AMD obviously lied to us by claiming Polaris would improve efficiency by 2.5× and be twice as efficient as Maxwell… I could have dug up another couple dozen examples, but everyone should get the point; It's time to get some perspective and stop complaining about everything AMD's competition does.

I point you to the earlier reply I made on the topic, explaining the motivations and corroborating evidence that all point towards the conclusion Intel knowingly misled consumers via this demo.

They include:

1 - Q: Why use a chiller instead of LN2 for an overclocking demo? A: Because it can be hidden under the desk and the system can be made to look close to an open loop, whereas LN2 would instantly give the game away.

2 - Q: Is it reasonable for someone stood on stage, next to a chilled system with a noisy compressor running, to "Forget" to mention to the public that the system was overclocked, while at the same time repeatedly trumpeting the figure "5GHz!" A: No.

3 - Q: Is it reasonable for the custom Gigabyte board in question to have a VRM capable of handling over 2KW, unless this was very much planned from the beginning to be an overclocked demo? A: No.

4 - Q: Given the answer to question 3, is it reasonable that literally no part of the demo, in terms of scripted or promotional material, mentioned the word "overclocked" at all, given the overclocking was clearly planned months in advance in order to give Gigabyte time to design and fabricate that motherboard? A: No.

5 - Is it reasonable that for the launch of a new Halo product for the entire Intel consumer segment, there was no Q&A session? A: Maybe, but it's not usual, and it clearly works to their advantage in terms of plausible deniability for the omission - If a journo were to ask "Is this demo representative of clockspeeds at launch" for example, then Intel's entire ruse is scuppered, they can no longer hope to profit from the turn of events we're seeing right now.

6 - Given the clear confusion this has caused, is it reasonable for Intel to be providing piecemeal, individual answers to individual tech publications, regarding the "forgetting" to mention the overclock? A: No. If Intel weren't intending to profit from the confusion they've created, they would have made a much more visible public statement about this. Instead, what they're doing is *only addressing this concern, via the individual sites that have reported on the issue.* Sites that have made no mention of this demo being "misleading", do not seem to have been contacted by Intel to help them clarify their articles.

 

[Air]

People arguing that Intel's deception was not good enough, so it's the people/press fault that they believed them and Intel is not to be blamed. Trying to misslead consumers is ok if it's not done too well.

Amazing.

 

[Patriot]

In Finland. Normal fuse in rooms here is 10A, but 16A fused lines are really common for kitchen, laundry rooms etc. So I could easily find free 16A line for one if needed.

That's a 10A cable, needs a c19/20 cable for over 1600w. c13 cant handle it.

 

[deu]

The issue is not whether the tech savvy should have believed that was realistic. - They shouldn't, but as I explained before, there's an incentive in terms of site traffic, for the tech press to report the most exciting-looking, dramatic headlines, often to the point that those headlines aren't checked for accuracy, or are essentially just regurgitations of company press releases making wild claims.

The issue is more that Intel know this, and made clear efforts to both focus on the 5GHz number, and also to specifically avoid using LN2 cooling that would have been picked up on more readily - instead opting for a comparitively rare form of extreme cooling that just so happens to look much more like a conventional open loop liquid cooler.

Add onto that the extremely conveniently unverifiable excuse of "We forgot to mention" it was overclocked? I don't for a minute believe that this wasn't planned in advance. Nobody fucks up a detail that important and then just passes it off with "we forgot" - That's the kind of omission that when a company has genuinely fucked up and didn't mean to, you get press releases and statements. Not just private communications to whistle-blowing media outlets stating "we forgot".

That's a very specific thing here too - Intel aren't producing or mailing out a broad statement about this, or a large document. They're literally just privately messaging the sites that have taken issue with this and allowing them to relay in a very limited way to the smallest possible audience, their excuse. Most average users who have already picked up on the 28 core 5GHz story, will never see these retractions or updates - and Intel have a clear vested interest in keeping it that way.

So forgetting to say 'overclocked' at the scene is what I can 'forgive'. Depending on how long intel waited to correct this informations livelyhood on the techsites and forums is where the there could be an isssue. Sorry to say but half of the people on those stages should not be there due to their cringeworthy delievery. So I think we agree I just focused on the specific part of forgetting to mention that the cpu at the scene running at double beforeseen clockspeed with compressorcooling WAS indeed a overclocked clock!

 

[jabbadap]

That's a 10A cable, needs a c19/20 cable for over 1600w. c13 cant handle it.

Uhm I think you mean that power supply on that video, or? I agree, that does look like a 10 Amps IEC c13 cable(well the spec says it could handle 250V@10A=2500W), not IEC C21 required for 16 Amps systems.

 

[TXST Guardian]

The fact that they had to use a chiller to achieve this performance is hilarious. AMD has their heat issues , but sheesh! If Intel can make those clocks on a heatsink and fan setup I will be highly impressed. I would love to see AMD's new 32 core run in comparison to Intel's 28 core.

 

[Patriot]

Uhm I think you mean that power supply on that video, or? I agree, that does look like a 10 Amps IEC c13 cable(well the spec says it could handle 250V@10A=2500W), not IEC C21 required for 16 Amps systems.

bit of both... I live in server world and 1600w is the last psu that can take c13, 2200w is above the 208v/2008w threshold for 10A cables.
So the demo used a C19, and the 3300w psu would be a firecracker on a c13. I forget some places use 250v... that said because power supplies can take a variance on input voltages... most ATX 1600w and up PSUs use C19.

 

[John Naylor]

I have one big question ... who cares about cores ? I just don't understand the focus here. What is even that 5th core gpoing to do for 99.9% of us ? Why is the media flooded with "Ooh Ooh, this will change everything ... look at all the cores". It's like a Hollwood movie that uses a technical procedures drawn from every chapter in a film school textbook... who cares that you utilized 17 different camera techniques if the viewer has no idea what the movie was about nor cared for any of the chracters. The car that crosses the finish line 1st get the purse, you don't get a special prize for having the most cylinfders or cubic inches.

If I'm building a gaming box, this is what matters ...
https://tpucdn.com/reviews/AMD/Ryzen_7_2700X/images/perfrel_1280_720.png

If Im building a video editing box, this is what matters (well in 2017... have not seen 2018 test yet)
https://www.pugetsystems.com/pic_disp.php?id=42852&width=800&height=800

 

[newtekie1]

a comparitively rare form of extreme cooling that just so happens to look much more like a conventional open loop liquid cooler.

Did you see the setup? It looks pretty much nothing like an open loop liquid cooler. But I don't think Intel even let the press see the demo unit, because if Intel had it would have been very obvious immediately that it wasn't running a normal cooler. Even if they had used one of the systems where the entire refrigeration unit is enclosed in the case, think OCZ's Cryo-Z, it is still pretty obvious that you are using refrigerant based phase change cooling.

Any time a presenter gives you a demo of new hardware, but doesn't actually let you see the actual new hardware that is running the demo, you should be very suspicious.

 

[GlacierNine]

Did you see the setup? It looks pretty much nothing like an open loop liquid cooler. But I don't think Intel even let the press see the demo unit, because if Intel had it would have been very obvious immediately that it wasn't running a normal cooler. Even if they had used one of the systems where the entire refrigeration unit is enclosed in the case, think OCZ's Cryo-Z, it is still pretty obvious that you are using refrigerant based phase change cooling.

Any time a presenter gives you a demo of new hardware, but doesn't actually let you see the actual new hardware that is running the demo, you should be very suspicious.

Yes I did see the hardware, but while it was onstage, it would have been difficult enough to spot that insulation and that pack of cables, that Intel could have banked on people in the audience assuming an open loop. This would have been further planted in peoples minds by the fact the system DID have open loop hardware clearly visible, such as the tube res.

I've said all along - Intel did not have to fool everyone forever here. They only needed to generate a single days headlines, safe in the knowledge that later clarification would not reach NEARLY the same number of eyes and ears as the original, misleading information. For that, a chiller sufficed as subtle "enough", an LN2 pot would have given the game away immediately.

 

[Air]

Yes, it's sarcasm. It was in response to eidairaman1's comment about dishonesty, because why expect honesty from a corporation? Lying for dollars is pretty much their job description, whether they make PC hardware or underarm deodorant. It's all about saying whatever separates customers from cash, just short of actionable lawsuits. It's one of the main flaws in the free enterprise system - too much honesty will cause you to lose money, while your competitors lie their asses off and take it to the bank. This leads people to the mindset that lying is a virtue - it results in more cash, how can that be a bad thing?

Some companies do push the "lying for dollars" way further than others, in all sectors of the economy. Let's not pretend they are all the same.

 

[las]

No, they showcased a 28 core CPU @5GHz & nowhere did they say it was OCed, with extreme cooling. This is lying by omission IMO, like many others, not to mention there was no Q&A session so none of the journos could catch them in the act!

They showed all the core clocks and they didn't hide the machine on stage. The machine even had window. You could easily see that the cooling was way above average.

AMD's 32C/64T must have been much slower since AMD pulled the CB result. Would be fun to see stock vs stock tho, and oc vs oc.

 

[Patriot]

They showed all the core clocks and they didn't hide the machine on stage. The machine even had window. You could easily see that the cooling was way above average.

AMD's 32C/64T must have been much slower since AMD pulled the CB result. Would be fun to see stock vs stock tho, and oc vs oc.

Yet most tech sites didn't have pictures of the on stage box but only the wc'd one... so everyone reading and not watching saw a closed loop with intel's 5ghz 3kw numbers

 

[Rahmat Sofyan]

Fantasy vs Reality ..

"Knowing full well that AMD would be showing a 32-core Threadripper CPU during Computex 2018, Intel created a diversion to steal some of AMD’s thunder a day early, showing unreal Cinebench numbers from a suspecting 28-core 5GHz CPU that is not a real product."

@techspot.com

 

[Imsochobo]

Of course it was dishonest, that's how marketing works. I'm more interested in knowing who believed that a current-gen 28-core could do 5GHz stock - especially when the demo machine required 1700W of cooling capacity.

Either way, some fanboy idiots will buy this, just like the idiots who bought the FX-9590.

The level of idiocy required to buy the fx-9590 even as the most die hard amd fan was high.....

Well my 230V line with 16A fuse would drive that easily, but that looks hoax.

You could run the 1700w cooler + the 1600 W psu for the cpu + another psu for a 1080ti and you'd totally max out that fuse for one system

What country are you in? Here in England, a standard wall socket supplies 240V at 13A. You can go higher obviously, but then we're talking about special mains installations like they have in factories etc.
I'd have believed that on two mains inputs connected to different sockets, but not on one. Something's not right here, video or no video.

In Finland. Normal fuse in rooms here is 10A, but 16A fused lines are really common for kitchen, laundry rooms etc. So I could easily find free 16A line for one if needed.

In scandinavia 10\15(16),22 is common I believe, I got 20 fuses @ 16amps at 250v(248v measured) in Norway.

That's a 10A cable, needs a c19/20 cable for over 1600w. c13 cant handle it.

C13 is more than capable of 3000watts it's rated at 15 amps for C13, 15*240=3600watts but In USA you cannot do a 2000W psu because 15*110 for instance = 1650 watts max for C13.
In any 240V system a C13 is sufficient for pretty anything you can throw at it unless you daisy chain through it.

 

[mouacyk]

This thread needs a water chiller.