Discussion in 'Frontpage news' started by Hilbert Hagedoorn, Jul 10, 2019.
None of it is sarcasm, not sure why you would think so.
I was hasty in my original inspeaction, I later found out that is it for Mobile as it clearly has no PCI E 3.0 X16. But that only adds to discredit this leak, not to add more credibility to it.
"HT for everybody" would be a welcome thing. I hate how they've made it exclusive to i9 (which is even pricier than i7, yet i7 costs the same here as it did when it had HT).
What are the chances this is a slide intel made with some errors on purpose just to spot which of the press leaks secret materials and info?
This chance exists. It is done in industries more than people think. Because leakers get hit much later without knowing why.
I'm arguing with the admin of an italian kinda famous website who swears they're real because he's spoken with his sources, etc...I think he's full of S, but anyway he might be right. And he's also saying the 9900KS won't be on the market this october, but instead these will be.
No idea if i'm allowed but i'll do it anyway.
Please how do i use this function? This forum got so much haters / nonsensers that i feel bad for it. I would like to clean up my personal view of the thread.
I do not see benefit of those chips over current ones. (As I wrote before.)
What people save on CPUs, they'll have to pay on MBs, since no backwards compatibility. And likely no upgrade path to next generation as there will be DDR5 and yet another socket.
That's no good for people who are hard stuck on getting intel only. (At least if they want to upgrade every year or even two.)
Anyway, for consumers, this is bad time to buy intel for that reason.
I wonder how popular those chips will be in 2 months if true. On other hand, what else should intel do? Their 10nm seems to still have clock trouble while power draw was not the best. And unless that's solved, they can't release it against Zen2. (As they need that clock advantage now.)
And they'll have to fix both clock and power against Zen2+ or Zen3 whichever comes around time of release. (If Zen2+ comes at all, but it makes sense.)
Weirdest part on this leak are prices. Yes, intel acknowledged AMD for 1st time in many years as competition. (While Raja had some darn stupid comments on AMD having nothing in their future after being removed.)
But this "leak" has them competing by both increasing core count and reducing prices. That's big acknowledge to AMD and validation too. Which will send more people towards AMD. nVidia does marketing "right" way, they always berate AMD's products as inferior in all aspects. (While it is pathetic, it works on people.)
Therefore I am more on side that it is not truthful leak, I do not deny possibility that it may be real and entertain idea that it is for sake of thinking about implications.
Comet Lake is just another re-fresh from the "broken" architecture with all the security vulnerabilities intel is suffering for quite a long while by now.
While some of these were fixed via software and firmware updates (also not entirely, e.g. Spectre got fixes to only a few types of it) it still remains vulnerable and due to the architecture it is not possible to make hardware fixes on those either.
So wether the leak is true or not, the referred CPU lineup will still be affected by the security vulnerabilities as much as the current lineup is.
Which is also a good reason to safely assume that the leaked CPU lineup might get additional problems upon launch, being pushed too fast with an already broken architecture. If odds are true and these CPUs really going to be released even this year, then I don't want to know how Quality-Check will be neglected at intel's end.
I can sense a return-wave incoming.
Well honestly if i buy a PC in a few months, i'll get something good enough i'd keep at least 3 years, but possibly 4-5. Anyway i honestly think this all here is a bunch of BS, there's no time to do everything in 3 months, if these were coming out in october we would've heard something from intel back in april if not march, anyway i think we'll only see 9900KS and mobile chips from intel this year, nothing else, we'll see next CPUs probably Q2 next year as that new roadmap says, which might be fake too, but i guess the time is around that probably, unless they're able to release 10nm CPUs and it'll happen probably Q3/Q4, meanwhile they could drop the prices on series 9000.
Since when Trolls allowed to stay unbanned on Guru3D?
Sadly its never going to happen, go check the MP600 benchmarks, the issue is that 3D NAND is slow as hell, 980GB when uncached and using 8 channel controller.
The only way to go faster is more NAND chips and more channels per controller and the industry is going to the opposite direction, of bigger NAND chips so less NAND chips per SSD needed.
Unless they invent z new type of NAND or go back to SLC, we going to stay with current speeds unless its big drives that have more then 8 channels per controller.
And the only boost we going to see is during the DRAM or SLC layer cache part, just like current PCIe Gen 4 SSD drives.
Turns out its not quite real, it will use LGA1200.
The biggest part is, no release until 2020
edit: ah, I see @oxidized already mentioned it.
No PCIe 4.0, no buy.
There is huge difference between hyperbole and trolling. At least for emotionally uninvolved.
= = = =
Btw. that roadmap has rather funny code names: Glacial Falls + Cascade Lake (Feels like global warming.)
Then there is interesting notion:
Going from Coffee Lake-S 8C/16T 95W to Comet Lake-S 10C/20T 125W means, that for 25% more cores there will be 32% higher TDP.
There are ways around these limitations by using better materials, vertical 3D stacking and heterogenuous integration of memory controllers and memory modules.
There are emerging NVRAM technologies based on resistive (ReRAM), magnetoresistive (MRAM) and phase-change (PCM) materials which could be used to supplement or even replace NAND flash - such as Intel Optane (3D XPoint), patent protected until 2031.
There are new APIs, filesystems and firmware strategies designed to work around flash memory properties like large write block size.
Cheaper DRAM to allow larger buffers, return of battery/supercondenser memory backup, RAID5/50 and 6/60 volumes made from 4-8 smaller SSD drives... lots of possibilities.
It will be years for them to emerge and none of the companies are donation-ware, Intel is sucking the life out of Optane with insane price [A believer here got day one 900p 480gb Optane with that free spaceship cd-key that I sold, got back 150USD]. have you seen Optane cards? The PCIe card has ton of chips, thats how they get the speed, I juts looked at mine with flashlight its 4 chips in 3 rows on each side so at minimum 24 chips for 480Gb [I may of missed some].
Right now they pushing QLC which is even slower.
And we stuck at 2Tb drives for some reason, only Samsung has commercial 4TB drives and they overpriced aka cheaper to buy x2 2Tb but even cheaper to buy x4 1Tb, price has to be like on HDDs, bigger = cheaper per TB, in the SSD world its bigger = More expensive per TB