wccftech.com
>During a presentation at HPE Cast 2019, AMD revealed that their 3rd Generation Zen 3 based EPYC CPUs codenamed ‘Milan’, would offer better performance per watt than Intel’s 10nm Xeon chips.
During a presentation at HPE Cast 2019...
Landon Russell
Other urls found in this thread:
cdn.wccftech.com
s21.q4cdn.com
fuse.wikichip.org
twitter.com
Alexander Parker
fuck you goy
Aaron Mitchell
>7nm+ node
>allows for 20% more transistors than Zen 2’s 7nm
>also delivers 10% better efficiency
So how many cores are we going to see on Milan, something around 75-80?
Colton Morris
Same core count until Genoa at least. Genoa might still be the same, though it is increasing memory channels to 10.
Zen3 is 19h family, and Papermaster said it will be focused primarily on energy efficiency.
All prior Zen iterations are 17h family, so Zen3 represents the biggest generational change yet.
William Ross
>Zen3 represents the biggest generational change yet
Nah, just some really nice energy efficiency gains
Juan Bennett
Its the architecture delivering those virtually all of those efficiency gains, not process. If Zen3 was a small arch tweak compared to Zen2 then it would still be in 17h instead of its own new family 19h.
Cooper Richardson
lol HPE needs to shut their mouth if they ever want to get reasonable prices on Intel chips when they're back on top
Lincoln Rivera
>10% better efficiency
10% better overclocks?
>4.5 ghz + 10% = 4.95%
We can DRREAM
Evan Davis
7n+ is actually 15% efficient at ISO clock or 10% better performance at ISO power. TSMC updated the profile a few weeks ago during a presentation
Jayden Hughes
I wish a nigger would. I got HPE stock and they are dragging my portfolio through the streets of SF right now.
Jeremiah Hernandez
AMD claims their 7nm+ chips that don't exist yet will beat Intel's 10nm Chips that also don't exist yet.
Ya ok
Camden Gray
Isn't EPYC2 already better?
Jackson Howard
And chips are being made right now Intel's are never
Jacob White
Milan has already sampled to some customers, so those chips definitely exist. Lrn2 risk production and early tape out.
Noah Hall
Leo Cox
Dominic Watson
what's it to.you, smuck. G-d damn you stupid pieces of shit.fucknamd fuck amd fuck amd fuck amd fuck amd fuxk amd fucka mamx
Jack Davis
It's funny, in the past AMD always compared itself to generation-behind Intel or low/mid-range current generation Intel products while marketing. Now they're comparing themselves to future Intel top products while Intel does the opposite.
Jaxson Gray
>Our unfinished samples beat Intel's 10nm chips that don't exist.
Is AMD taking a page from the Russian Shitposters' book?
Joseph Wood
They definitely exist, Intel's own slides say they've been shipping samples for months in the 2nd link of They've allowed photographs of them since last year
fuse.wikichip.org
It's not hard to imagine one of the companies that had the ability to get them ran some tests and was allowed to report the results to AMD, assuming Intel's NDA is sufficiently dumb or one of the dozen antitrust lawsuits allow it regardless
Henry Miller
>But Jow Forums told me 10nm was going to be ready in 2021 at best.
Hunter Howard
Sunny Cove seems to be a step in the right direction. Seems like they got their head out of their ass's and made a new arch.
Fuck, it's been like 10 years of Core arches.