>AMD's version of a 1080ti both in launch price and /v/ performance (but it's 3 years later with Vega 64's same godawful power consumption) >OnLive AMD Edition >more /v/ >3700x, effectively an overclocked 2700x but with 33% less power consumption >the end
Well that was fucking disappointing. I will say the big increase in workstation performance and 16gb VRAM on the Radeon Se7en great for A/V production, but I doubt any of you /v/ crossboarders give a shit about that
Thankfully the 9900k is still shit, so AMD can salvage this by pricing the 3700x low enough to dab on the $$00k.
>making assumptions about a cpu you know nothing about yet >making assumptions about a gpu you know nothing about yet
Cameron Robinson
Wasn't there some "25% more performance for the same power" point on the presentation of the GPU
Joseph Myers
t. shills they didn't even give it a proper name, and why the fuck would they not benchmark their TOTL against intel's TOTL? literally just vomiting up what they told us. -VII = rtx 2080 = 1080 ti, and they said it's at the same power compared to Vega 64. -did you even watch the benchmark they did live? exactly. 'same power' consumption as a vega 64, and that card's power efficiency was already shit on release.
Dylan Mitchell
>Summary * new mobile chips * new drivers directly from AMD for said chips * some 20 new laptops instead of three(3) at the launch of the first mobile chips *AMD chromebooks *Threadripper is doing great *Epyc servers are doing great, still killing double xeons * Radeon VII = Vega +25%/+30% in games and +60% in openCL * Ryzen 3000 is not ready yet, but an early ES with 8c did still beat a stock 9900k * there is space for a second chiplet on the ryzen die * release date in mid 2019, more informations closer to release date
Xavier Taylor
>* some 20 new laptops instead of three(3) at the launch of the first mobile chips We'll see. I looked for those AMD laptops. It's like HP and Lenovo made one model with AMD APUs which was sold in a very limited amount of stores for a limited amount of time. The entire laptop market is Intel or Intel+NVidia. AMD's simply not there.
If they manage to change this then that's great. I'll believe it IF I see it. Intel not being able to deliver enough chips to OEMs could have helped. I do have the distinct impression that Intel has signed deals saying "you get a rebate, which is really the normal price, if you ship XX% of laptops with Intel chips and if you ship too many AMD systems you'lll pay a lot more than your competitors". Keep in mind that they've been investigated and convicted of similar deals numerous times in the past. So... I'll believe there's AMD laptops if I see it.
Logan Morgan
>they didn't even give it a proper name, and why the fuck would they not benchmark their TOTL against intel's TOTL?
Cause they didnt have to.
Landon Stewart
>Radeon Vega VII >Se7en >7nm Pottery
Benjamin Morris
Lmfao difference to nvidia is that you don't get CUDA and a shitty hardware encoder that can't compete with NVENC for streaming.
Jacob Reyes
>space for a second chiplet >high confidence 12-core AM4 leak
adoredtv confirmed
Brayden Watson
>why the fuck would they not benchmark their TOTL against intel's TOTL? Just think of it this way, if they can show their mid range slapping intels top tier It'll be a massive slap in the face to intel and a huge gain in the minds of consumers to gain mind share/market share.
tl;dr to flex on intel
Ian Peterson
seething amd turds
Levi Morris
>3700x Neither you or I actually know what product it actually is, but it stands to reason that it wasn't a totl consumer product, given that it was only 1 chiplet on it. AMD's most popular processor is the 2600, why not show their new midrange product (even if they haven't announced that yet) matching intel's fastest offering? It makes them look really good.
I like the 16GB VRAM but because theyre dogshit for writing compute software for, its a waste.
Jonathan Bell
> 16gb VRAM on the Radeon Se7en great for A/V production, but I doubt any of you /v/ crossboarders give a shit about that It is expensive because of this hbm-vram BUT the new consoles in a year or two will be 16 gb too so all cards with 8 gigs will be obsolete.
I'm on a 7970 and I'd like to find a decent upgrade in the >300 USD price range
Tyler Sanchez
What's wrong with 8 core gaming pcs and 2080/1080ti tier perf for a midrange pc going into 202x? High end stuff isn't even out yet stop being so dumb to hate on it. Arcturus and mcm gpus soon
Aiden Price
Because it wasn't ready yet maybe /v/? 8 cores is probably all they could get stable enough to demo, there's clearly die space for another 8 cores
Brody Fisher
They exist already, no waiting required.
Matthew Gonzalez
and the Se7en is coming on the SEVENTH of February
I was waiting for new gpu and what a disappointment this was. I bought vega 56 right after conference finished hecause Im tired of just waiting. I am not spending 700 fucking dollars msrp on gpu. There will be no price drop too as this gpu targets very high end market instead of competing with older cards.
Josiah Johnson
>can't get 8-cores stable >imply they can get 8 more cores in 6 months
KEK
Ian Bailey
For SEVEN hundred bucks Oh wait
Jace Watson
Where were you when Intel-Aviv was completely btfo?
>3700x, effectively an overclocked 2700x but with 33% less power consumption
[citation needed]
Anthony Peterson
Is this the strategy?
Dont have chips entirely ready so try and leak it out even slower and build up hype for the real launch?
Leo Lewis
Why the fuck would they bother to use their top tier cpus when a fucking engineering sample of a ryzen 5 can beat the i9 at the same clock, wih lower power consumption? >6 shoah gorillion watts! Amiright guys? Hahah! Autistic meme >3700x It was a ryzen 5 engineering sample but whatever helps you sleep at night. >hurr, 1080ti And why does the 2080 exist? Hmm? The radeon 7 literally has double the vram of 2080 and still costs 100 bucks less. While the 2080 uses gimmicky aliasing to fool autists into thinking it is actually 4k, the radeon 7 can, without kike-ing, actually render textures. Come back to shill when nvidia actually supports real rtx. Not the gimmicky bullshit that has to be turned off for the cards to match their predecessors.
Christopher Robinson
>1080ti... it's 3 years later Need to fire your research team, Pajeet.
Adrian Gomez
Well it's actually true. In gaming it delivers 1080ti performance with 1080ti power consumption for 1080ti price.
Ryan Walker
>Missing the point Remember when the 1080ti dropped in early 2016? I sure don't.
Lincoln Allen
Everyone is misunderstanding for some reason what AMD are trying to do with Radeon VII 7nm is understandably not cheap and even though AMD has MI60 for good dies and MI50 in case they have defects, the yields are not good so AMD has extra supply of MI50 defective dies that they can't allocate for enterprise sales. So they just add some gaming centric support features on it and roll it out to compete with the 2080 price which is where it is performing currently and not undercut it to make money back. Plus, board partners and AIBs won't need to wait for Navi which is not ready and the card will help alleviate their concerns about slowing Vega sales and this wait. AMD know they won't sell gangbusters on the card. It's probably not going to be a ton of cards either, since I don't think 7nm has terrible yields but they are positioning it so it is a high end gaming/prosumer card which should with the way they are targeting the card sell off the excess defective dies of MI50 they can't allocate. To me, the 16 GB of HBM2 and the possibility of having non-gimped FP64 precision will make it a nice sell to prosumers who can't pay for MI50/have no need for the cost especially since there hasn't been a 1/2 double precision card since Hawaii.
>3600x, effectively an overclocked 2700x but with 33% less power consumption And what has Intel been doing for the past 6-7 years now?
Jason Taylor
Did they really truly beat Intel in single core perf?
>wccftech.com/amd-zen-2-7nm-cpu-13-percent-ipc-increase-rumor/ 2700X at 4.3 scores 174 in CB15 single core >174 * 1.13% IPC = 197 pts Assuming 4.8 GHz boost for 3600X >197 * (4.8/4.3) = 197 * 1.12% = 220 pts 220 pts is exactly what 9900K scores but at 5.0 GHz which we're not sure 3600X will be able to overclock to.
Thoughts?
Mason Powell
Well it is quite possible...
James Price
What's so hard about that? They have 64 cores on Zen working fine, not like Intel's 10nm fuckup, this is a working process they're cutting down and tuning
Jackson Long
Nice Reddit image, my friend.
Isaac Butler
We already knew Vega 64 could've had 4 stacks of HBM2 with full memory performance but they decide to release this a 17 months later.
A mid-range, early engineering sample of mid-range Ryzen 3000 beat a 9900k at multicore and must necessarily have come within a couple of percentage points of it in single core. Clock speeds always go up on final versions, and there will be a 12-core R7 and a 16-core TO above the mid-range R5 that will beat a 9900k in single core and multicore.
Intel has no answer (comet lake will have to give up single core for the 2 extra cores) until 2021 at the earliest as they will need both a new architecture and to unfucked 10nm.
And as soon as AMD hits 3nm, they're going to double core counts again :^)