>JewJizz on 18 core HCC i9's Why Intel?

File: 1483187550716.jpg (122KB, 1500x1054px) Image search: [Google]

122KB, 1500x1054px

>>62196955
>Why
>Intel
You just uttered an oxymoron.

>>62196955
Explain

>>62196978
new ultra expensive i9 CPUs once again have thermal interface material between the die (the CPU proper) and the IHS (heat spreader, the silver cap on top). This is bad because the TIM (JewJizz) is super fucking cheap, and causes poor temps because it doesn't conduct heat for shit.

The alternative to JewJizz is to solder the IHS to the die, which is far superior. AMD solders all their current CPUs, and we hope to fuck they don't switch to JewJizz.

>>62196955

that's a big die

i wonder what the yields are kek

>>62196962
No I believe an oxymoron has to be a metaphor as well.
Furthermore he wasn't contradicting himself.
Just answering his own question in the same sentence.

>>62197036
>that's a big die
>>62197036
>>62197036
>>62197036
>>62197036

>>62196955
Because even with shitty TIM they can still beat AMD

>>62197036
Yields are so bad they had to price the full die so high that it's basically unsellable.
This way they can claim it exists without selling a single one as other than a Xeon.

>>62197052
Yeah, in heat output and power consumption.

>>62197036
>cares about yields when you have 60% margins

File: 1499950820983.png (67KB, 200x144px) Image search: [Google]

67KB, 200x144px

>>62197103
>Yeah, in heat output
More like heat insulation. With AMD the heat can actually get out.

>>62196955

If you go with the yid, you must delid.

>>62196955
BUT WHY

>>62196955
JUDEN

>>62197051
4u

>>62197165
with more insulation you need higher temps on the source side to transfer the heat to the sink side. which leads to higher required voltage to operate in a stable way. which leads to higher power consumption. which leads to higher heat output.

>>62197036
Who cares what the yields are when you can sell all the mistakes as lower market products?

Why are they even trying to sell a $2000 18 core processor as a consumer product?

>>62197406
4 years ago, $550 is all that you need to get the second best i7(X9X0K). Now the best i9 is over $2000+, out of reach of many people who once could afford a top of the line consumer product. The cost of being an enthusiast is rapidly increasing.

>>62196955
So they fail after a few years, that's the only reason.
Remember, Intel can solder even their atoms, the excuse it's "too small" is 100% bullshit, they've done smaller.

They do it so that CPUs have a expected failure time that's built into the product (not that it will kill the CPU, but eventually the product will overheat to minimum the point it will throttle down and at worst shut off for thermal protection - making the user think the computer is broken or just slower than it really should be and time to upgrade)

>>62197461
What would a consumer do with an 18 core CPU? The consumer range should stop at 8 or 10 cores and anything beyond should be proper Xeons with ECC support, etc.

>>62197461
It's Ryz'ing everywhere - Ryzen is great performance-per-dollar, but to get the best out of it you need (currently) expensive high clockrate ram with ultra tight timings.
I can't really pin DRAM pricing on AMD - but the result is the same, no matter who you go with - the cost of enthusiast gear is increasing at an alarming rate.

>>62196962
that's not an oxymoron but you are a regular moron

Is virtualization disabled on these? Despite all the awful shit about this product, almost 50% faster 3d rendering than threadripper is hard to pass up. I would rather an unlocked 32 core epyc but AMD refuse to sell one and there's no epyc workstation boards for sale anywhere.

>>62196955
It also has an RFID tag on to store user information.

Even Tom's is giving them shit, and they've been Intel's biggest cucks since forever...

>High Temperature Differences Challenge Cooling Performance
>The reason that Skylake-X is so much harder to cool traces back to the thermal paste Intel chose to use instead of solder between the processor die and heat spreader. Although paste is cheaper, it's also less than ideal for cooling performance.
>Intel’s digital temperature sensors report reliable results from 35 to 40°C and up, prompting us to only include values above that range in our analysis. The difference between the water cooling block's temperature, which is held at a constant 20°C, and the CPU temperature reported by Intel's sensors shows just how bad of a choice thermal paste really was.
>We measured the CPU heat spreader’s temperature the same way we did when AMD launched Ryzen 7 1800X, by using a thin copper plate. The resulting curve shows very clearly that waste heat can't be dissipated quickly enough. A solution good enough for a thermal lightweight like Intel's Core i7-7700K just doesn’t work for Core i9-7900X.

>>62197612
tom's hardware aren't as big intel shills as everyone makes them out to be, in fact they gave skylake-x probably the worst review out of any of the 'mainstream' tech review faggot outlets out there.

>>62197534
It's not that much more expensive. DDR4-3000 is about where it starts to plateau in terms of performance gains. 16GB kits can be had for $150 or so.

>>62197568
>50% faster
Wut, sauce?

>>62197524
The same reason people went with the 3960X instead of the 3930K. To show off.
>>62197534
My perspective was that Ryzen is meant to being prices down, but the current 8-10 core i9s does not seem to be any more affordable at all. Instead Intel just stack more expensive SKUs on top of it just for the brand loyalty earnings, and tries to cut corners by switching entirely to TIM (originally the argument was the mainstream die was too long and narrow which can form micro-cracks, but TIM on the 2066 CPUs is frankly unacceptable) Not only that, with Ryzen's presence, Kaby Lake (and i5 2066) was a platform that is not justified for purchase at all. Hopefully coffee lake is priced reasonably.

>>62197042
>>62197548
Nobody cares, go back to >>>/lit/ where you belong, you tools.

>>62197659
It's not that 3200 ram is expensive, it's 3200 ram with good timings is probably where it starts getting expensive. But then those models probably also have rgb shit on them so who cares.

>>62197662
Ryzen did bring it down. I remember all the news outlets saying it's great now intel's 10core is now 40% and better performing, ignoring that the $1700 price before was just fucking insane.

Even the 8core 7820x isn't that far out of performance/price if you match it with the 1800x launch price. It's just the 1600 and 1700 are such incredible value there's not much point in looking at anything else.

>>62197783
True, timings get expensive, but I don't think they're going to be all that helpful other than squeezing some moar fps out of it.

File: 1336072054602.gif (7KB, 256x256px) Image search: [Google]

7KB, 256x256px

>>62197307
>with more insulation you need higher temps on the source side to transfer the heat to the sink side. which leads to higher required voltage to operate in a stable way.
Holy shit dude. What sort of crazy fantasy world do you live it.
Intel chips don't run hot because the heat needs to grow stronk so it can burst out through the IHS.
They're just leading heat away badly. Do you believe you need to heat up your CPU so it will run well?

>>62197768
/g/ isn't just for illiterate Pajeets who can't speak English.

>>62196955
Got a source on that image?

>>62197886
Found it
http://www.coolenjoy.net/bbs/27/1605931

>>62197845
resistance increases as temperature increases, this is EE 101 here

>>62197036
4u

>>62197768
>Nobody cares, go back to >>>/lit/ where you belong, you tools.
The current state of /g/

>>62197795
Pretty much. My stock 1700 with cheapo 3200 cl16 ram runs all my games at above 60fps and that's it for vidya really.

But cheap 8core was a god send for me. The biggest bottle neck for me after ram amount was always the cpu. Up to 40 minutes on compiling an open world game engine when opening various projects is biblical.

>>62198188
Does it not throttle because of TJ Max before that becomes an issue?

>>62198188
Not in semiconductors. The hotter they get the lower their internal resistance gets, which allows more current to flow, thus increasing power and heat. Too much heat or not enough cooling sets up a positive feedback loop resulting in thermal runaway and the component in question going up in flames.

>>62197768
I can see how Xiaomi shills have become so dominant with this board so full of pajeets like you.

>>62197659
>Wut, sauce?
Apparently 4.4ghz is possible on all 18 cores.
http://www.coolenjoy.net/bbs/27/1608222

>>62199698
That's is golden silicon for sure.

>>62196955
The 18 core exists solely so Intel can say "We have the biggest ePeen", even though nobody in their right mind is going to pay twice the price for ~10% more performance over Threadraper.

>>62199698
>Apparently 4.4ghz is possible on all 18 cores.
Everything is possible under the LN2.

>>62197036
and it's competing against poo in loo single die.

intel is so goddamn far ahead.

>>62199832
Are you, by any chance, functionally retarded?

>>62199843
7551p is *only* $2100.
And that's full fat 32 cores, 8 memory channels and 128 PCI-E lanes.

>>62199832
TR and Epyc are MCMs though

and the way they're priced, only retarded fanboys or clueless office drones will keep buying intel

>>62199859
Yeah, I missed that. By the time you hit $2000 you're looking at server prices, for sure. And the 18 core is still a gimped Xeon with no ECC, making it useless for workstation tasks.

>>62199892
>making it useless for workstation tasks
It's not exactly useless, but compared to 7551p it's clearly uncompetetive.

>>62197009
Who would they do this?

>>62199986
To save a couple bucks per processor in materials costs. The glob of thermal paste is a couple cents whereas the gold and indium needed to properly solder it together is a couple bucks.

>>62199858
single dies are better and difficult

>>62200000
Numbers of truth.
I hope ryzen will kill these practices

>>62197051
nice

File: 4L_r7IFp68s.jpg (68KB, 626x461px) Image search: [Google]

68KB, 626x461px

>>62200000

File: large t2 usfire.jpg (117KB, 1600x674px) Image search: [Google]

117KB, 1600x674px

>>62196955
>>62199698
>4.4ghz is possible on all 18 cores
Can't wait.

>>62200006
Ok, so you're retarded. Thanks for letting us know.

>>62197147
>implying anyone is actually going to buy this thing

>>62196955
HAHAHHAHAHHAHAHAHAHHAHAH HOLY FUCK HAHAHHAHAHAHAH

>>62200000
Witnessed

File: heil_mein_quads.jpg (16KB, 300x472px) Image search: [Google]

16KB, 300x472px

>>62200000

>>62199698
If someone is using this as a serious render station it's not going to be OCed. Stability > ePeen.

>>62197662
>Hopefully coffee lake is priced reasonably.
Pricing leaks put the 6 core 12 thread at ~$390.

>>62196955
> 2 kiloshekels
> no ECC

What did Intel mean by this?

>>62200177
>If someone is using this as a serious render station it's not going to be OCed. Stability > ePeen.
Not really, overclocked i7 + 1080ti is most common thing you will see with 3d users for workstations. Even when gpu rendering usually its 4x 1080ti, no one touches "professional" cards.

>>62200210
YOU WANT ECC? BUY XEONS

>>62200352
I'm going to assume this is only because Ryzen is relatively new .. nobody should be buying i7's for rendering these days.

And anyway, OCing a quad core i7 is a lot less of an issue than OCing a 18 core i9.

>>62200000
>a couple bucks
Way less.
But add penny to penny and you will be able to bribe enough people to never actually have to pay the enormous fines you were sentenced to.

>>62199698
>Apparently 4.4ghz is possible on all 18 cores.
I'm not even going to open that link but I hope is real so we can have some real housefires

>>62200917
Intel did this last time with the i7-7900X. They seeded golden silicon to a few OCers, I remember some guy hit 4.5 GHz on a shitty air cooler and Intel faggots were screaming left and right that AMD was doomed, this was the ultimate CPU, etc. We all know the rest of the story. It comes out, and you need a custom loop to keep it under control.

>>62196955
im just going to liquid metal my 18 core

>>62197052
>Because even with shitty TIM they can still beat AMD
Pay attention to the TIM please, and not the benchmarks.

>>62197516
Soldering on these chips incur some loses, thefore additional yield lose.
It's not the solder costing itself it's the cost on an already troubled yield.

Corning im is shitty but very stable under thermal stress, this shit is rated +5 years under daily abuse.

>>62201727
They certainly don't have yield issues with their normal 4 core consumer chips and they use the same shitty TIM. I think its just that they haven't had competition for more than half a decade and got away with saving a cent per SKU. Until now.

>>62197009
>AMD solders all their current CPUs
No they don't, only the zenver1 chips and I would wager that the zenver1 APU's will once again have TIM'd IHS, like most (not all, some of the high clock chips have soldered IHS) of the Bulldozer variants on FM2 socket.

Since Intel moved to TIM'd IHS on these 'high performance platform' chips too, I would say changes for soldered IHS across the board in any near future is less than zero and thus Intel have really fucked up. And they already fucked up with the super thin PCB on the 14nm chips (as we can see, the small manufacture doesn't 'need' thin PCB, since the zenver1 chips have PCB just as thick as on the AM3 chips).

File: 1200px-Transistor_Count_and_Moore's_Law_-_2011.svg.png (48KB, 1200x1079px) Image search: [Google]

48KB, 1200x1079px

>>62196955
WHY DONT THEY JUST MAKE BIGGER CPUS TO GET BACK TO MOORES LAW

WHY NOT MAKE THEM BIGGER SO THEY HAVE MORE TRANSISTORS AND CAN DO CALCULATIONS FASTER??? REEE

>>62197662
The market for such 10-18 cores either amd ir Intel is really tiny. Considering both ryzen and Intel 6-8 cores on mainstream expect the hedt platform to get even less market. Intel was already pushing Gaming angle on hedt just hard and instead deliver a platform that is miserable on games.

>>62201796
For a bean counter any yield damage including soldering is bad. With Tim they simply remove that from equation, even more now the need to shave margins everywhere

18 core will be only 20% faster in most real-world workloads. 16 core about 15% faster.

>>62201984
Latency probably.

>>62201984
Pollack's law means single core performance only grows at roughly the square root of transistor count, sweetie.

>>62196955
>inb4 apple will be their only customer

>>62196955
>Why Intel?
Why? Intel.

>>62200000
probably a buck max intel is probably .001 think of how cheap and easy bulk they can do

>>62200000
Check'd

File: 1475383281221.jpg (75KB, 816x816px) Image search: [Google]

75KB, 816x816px

>>62200000

>>62200000
Check em.

File: umineko jews.png (707KB, 768x768px) Image search: [Google]

707KB, 768x768px

>>62200749
>>62202554
>charge 2000$ for the processor
>jew out and fuck up your customer to save 2$
They could just stop putting the lid all together at this point, but people would stop breaking processors by trying to delid. Let's make PCB even thinner and put more capacitors at the top instead.

>>62202820
they buy a whole roll of metal sheets, and make their own lids soon pot iron levels

File: maxresdefault.jpg (180KB, 1920x1080px) Image search: [Google]

180KB, 1920x1080px

>>62196955

tfw got the last good intel soldered chip

>>62202933
5820k?

>>62202989
6850k, pretty good temps.

>>62202933
last good any chip from intel was sandy bridge though

>>62202996
Oh, that's a nice one, as long as you got one for cheap.

>>62203018
$299 at microcenter

File: 1503777021722.jpg (125KB, 1280x720px) Image search: [Google]

125KB, 1280x720px

>>62203020
Nice

>>62202820
>They could just stop putting the lid all together at this point
But that makes for better heat transfer, and then they won't overheat under the insulating lid.

File: c4e46361f034042813d096ff4c983854.jpg (725KB, 426x1500px) Image search: [Google]

725KB, 426x1500px

>>62196955
People will buy it. People will even shill it for free on the net. So why not?

Companies tend to milk and fuck over customers when they feel they have reached the peak of profitability/ability to grow/innovate/improve the product.

https://www.nextplatform.com/2017/09/01/huge-premium-intel-charging-skylake-xeons/

>>62203812
Intel's super expensive HEDT shit has historically sold very little .. 0.04% marketshare for i7-6950X ($1700) on Userbenchmark, and that was before there was any real competition

>>62200000
A reasonable post for quints.

>>62196955
Because the last time they did this was a few years ago so they hope people won't remember, in a few years they'll release a new cpu line with a soldered heatspreader and that will be a marketing headline just like last time.

Rinse and repeat.

>>62199951
Reminds me of the Pentium 4 HT extreme edition from the mid 00's.

>>62200000
You must be so proud of yourself.