Amongst the X299 disaster they silently launched this. Is it

>>61458977
Better for what? All of them have incredibly low clock speeds and would be shit on by cheaper Ryzen chips for desktop use. The cheap "Bronze" ones are sub-2GHz.

>>61459100
there have been some 3.5+GHz E3s and E5s in the past

>>61459257
What does that have to do with the current lineup? A 3.6 base/3.7 turbo quad core Xeon Platinum costs $7007. The Gold version costs $1221. The Silver and Bronze lineups don't have a single chip with a base clock over 2.6GHz.

>>61458977
Just get a broadwell.

>>61459413
>A 3.6 base/3.7 turbo quad core Xeon Platinum costs $7007.
Thats for a E7 series chip you retard when that other anon was talking about E3/E5 chips.

>>61459983
There is no "E" anything you stupid fucking cocksucker. The chips are segregated by Platinum/Gold/Silver/Bronze tiers, all of which I fucking covered in my post. Here's the full fucking lineup:

http://www.anandtech.com/show/11544/intel-skylake-ep-vs-amd-epyc-7000-cpu-battle-of-the-decade/7

Feel free to pick one out of it that you feel would be a good deal for desktop usage. Or just go back to /v/, you moron.

>>61458977
for the price of one 1950X you can buy the xeon silver/bronze 4116 (12c/24t 2.1GHz base clock, 3.0 turbo)

it's just not worth it

Xeons are on LGA 3647 socket, there's no Xeons on LGA 2066 socket

>>61460043
>http://www.anandtech.com/show/11544/intel-skylake-ep-vs-amd-epyc-7000-cpu-battle-of-the-decade/7
Holy fuck that's gotta be like 100 SKUs

I bet they just produce one and sell the binned abortions at arbitrary prices

>>61460325
53 or 56 actually

and yes, there must be at most 15 different ones, then they just turn off certain features so they can sell 4 different versions for different prices

>>61458977
Whatever dipshit they put in charge of this shit show deserves the rope, what the fuck are they doing?

Not only are they shitting all over the Xeon brand with this tacky metal tier shit but they're needlessly complicating an already incomprehensible lineup even further.

Is there something else going on? Why the fuck is Intel utterly losing their shit just because AMD bothered to glue a couple moderately competitive Ryzen dies together? Were they really this unprepared?

>>61460325
They have three different dies for the different core counts (iirc LCC is up to 10, HCC is up to 18 and XCC up to 28) and OmniPath comes with an extra connector to do whatever the hell it does, but yeah most of it is just turning shit on and off while adjusting the clock frequency to make decoding the product stack as needlessly annoying as possible.

>>61461128
This entire lineup was made before Ryzen was even released. What they thought would happen is they'd be cementing their market position, but instead they look like fools.

>>61461128
SKL-SP was delayed so many times it's hilarious.
Fuck, Apache Pass support and some on-package thingies are stated for Cascade Lake (Purley Refresh).
Thee lineup was made in the days before AMD teased Naples.

>Intel managed to make 58 SKUs from 3 dies
The fuck.
HOW.

>>61461581

an experiment to see if their silent major shareholder is alive because he's not reacting so far

>>61461581
They thought they'd be applying extreme artificial market segmentation to wring every last cent out of their customers. "Intel: Because you don't really have any other choice". Which would've worked, except Epyc appeared.

I don't know if I'm retarded or what, but the Xeon SKUs made no sense to me whatsoever before this. Even if it it is retarded, I can without looking up part numbers probably know that a Platinum Xeon is better than a Silver one.

>>61461656
There was *rumors* about AMD cooking MCM design since what, mid-2015?
Intel had all the time in the world to clench their anuses.
Looks like Intel's current management is beyond retarded.

File: Xeon.png (45KB, 672x593px) Image search: [Google]

45KB, 672x593px

>>61461693

Nevermind, this is fucking retarded. Having a 28 core part in the same Platinum echelon as a 4 core part. Someone must've been drunk when they devised this garbage.

>>61461795
That 4 core part is for cache-sensetive workloads.
It's still XCC silicon, Intel would NEVER sell it cheap.

>>61461813
>muh cache sensitive workloads
meanwhile in real life they have to sell the bottom of the barrel bins that have almost no cores working