Well, shit. How much did they shrink their Vega? Does anyone have any comparisons with the previous Vega chip? Sad that there's no confirmed date for gaymen, but the enterprise might be happy with the gains(the few of them at least).
Vega on 7nm
Isaac Allen
Austin Campbell
Videocardz says 40%
videocardz.com
Isaiah Turner
Pretty sure that's smaller than Lisa's areola.
Mason Foster
L-lewd.
Jacob Mitchell
Oh my how lewd
Luke Turner
~300mm2 is not bad at all, it would put them right there with the likes of the 1080. The problem is that Nvidia themselves haven't shrunken to 7nm, so they would still be very far behind.
Caleb Martin
>dub
Carson Hernandez
But doesn't Vega have hbm2 on die ?
Wyatt Murphy
How did it work again? Do they include the HBM size? It wouldn't make much sense since you can clearly see how small the middle die looks.
Dylan Long
It should out perform the 1080ti, they actually manage a 35% performance uplift.
They also touted a 2x increase in power efficiency, though they'd only need to improve by a factor of .7 to trade blows.
GCN isn't ever going to be as energy efficient in its current form when crunching general gaming workloads, but thats still a pretty tremendous improvement.
No.
HBM is not on die memory, its on package. Die size and package sizes are completely different things.
Alexander Johnson
Also how do you measure a rectangular chip in mm squared ?
Oliver Morales
Are you kidding?
Charles Wright
Dumb trips.
Carson Perez
brb retrieving my well worn dunce cap
Ayden Diaz
How'd you get those upside down 'w's?
Nolan Rogers
i live in Australia
Noah Kelly
Hudson Thompson
I wonder if AMD will run any 7nm GPUs at GloFo later in the year, or if they'll stick to TSMC for volume on all their GPU designs. It'd be interesting to see how much the two differ.
Eli Perez
But is it smaller than her clit?
Jacob Garcia
size of a breakfast link
Hunter Russell
/ourmom/
Jose King
When nvidia, bitmain etc. will use this node too and there is too much competition.
Nicholas Lewis
Doesn't work like that. If TSMC fails to provide adequate volume you can just take the masks to another foundry and have them run it. There is no direct porting just because they have the same industry marketing name.
Complex ICs are not foundry agnostic designs. They're designed along the guidelines and design rules of a singular given process. Everything launching in 2019 and 1H2020 is already set in stone.
Brayden Allen
??????
Brayden Brooks
Nigga just look at the center die, it looks smaller than previous Vega and Fiji.
Ryder Rogers
I wonder what performance it would have if it had the same die size as 14nm vega
Nathaniel Roberts
It does not matter. Nobody but miners will be able to buy them anyhow.
David Cook
>tfw just bought a 22nm cpu for 350 bucks
it's an 18c ES Xeon @3.6ghz boost, did I fuck up ?
William Ortiz
let's just say that the money you saved on the cpu you'll spend on electricity
Hunter Wood
meh I live in Canada, it's free :^)
good purchase then ? It was a v3 btw
Robert Torres
It's an HPC die, the only thing that matters is DPFP perf.
Brody Campbell
well, depends, what do you do with your computer?
Camden Torres
I bought a dual cpu Mobo too, planning to add another one in 4 years or so
I gaym, and do a lot of virtualization, encoding and blender
Also i love all the Xeon features since I plan to try out gpu pass through
I know it's gold for the rest but do you think the gayming will be sub-par or okay?
William Myers
well, some people would say it's extremely subpar, but I'd say it's okay given the amount of MT performance you're getting in exchange, but you need an acceptable gpu too
Jose Richardson
Let's say I get a 1080 or a ti, do you know how many frames I'd lose approximately compared to an i7 at the same price ? can't be more than 10 right ?
Cameron Peterson
I don't have the exact model number of your cpu, but I'd say a gpu that strong would get seriously bottlenecked
Eli Morgan
E5 2680v3
no way it gets bottlenecked, seriously ?
Aaron Watson
2 of them should be enough though 1 wouldn't be enough to power a 1080ti
Robert Martinez
>E5 2680v3
nah not powerful enough for just one and also uses +150 TDP on stock 2 of them doesn't make sense given the highest GPU you could run would be a Titan V or last gen Quadro.
Benjamin Collins
youtu.be
>a v0 wins/matches in gaming
>v3 not powerful enough
user please ... are you sure of what you're saying
Gabriel Hernandez
a bunch of weak cores will bottleneck it no matter what, for gaming you want faster cores
Leo Gutierrez
it can't.
because they're putting more ram
Eli Nguyen
just increase the size of the package then
Jayden Cooper
is this TV?
player?
all inone is here
satch.tv
Hudson Brooks
>AMD Radeon RX using 7nm coming to gamers.
e3 starts, people start to talk about consoles.
easy leap to ps5, using vega 7nm.
kinda makes sense, seen a couple already. Timelines though vary wildly. Some going as far to saying ps5 will come tomorrow, most settling for the soon™ up to q4 2018 and others relagating it to 2019.
John Clark
OC'd Xeon vs i9
your really going full retard now eh?
Mason Thomas
not sure about that. reminder, btw: 10% of AMDs total revenue last quarter came from miners. That's no small amount. It's actually a lot. HBM and GDDR5/GDDR6 memory prices also contribute to higher GPU prices but all those chips going directly to miners have been a big factor. Now there's ASIC chips for quite a few popular coin algorithms. Both AMD and NVidia are starting to face some real competition.
Liam Anderson
That's definitely the plan, but they might just use 14nm because of the time.
Still have not bought a Switch because of its garbage Tegra, not waiting for AMD was unfortunate, but they sure have won money.
Connor King
please explain ?
John Wright
Only if your motherboard has room for another one
otherwise you dun goofed
William Hill
Lol. Good one.