Return to

The VEGA 56 / 64 Cards Thread! General Discussion



AMD did a demo in Budapest...........why does this concern me?

Love buildzoids stuff, only guy to actually overclock and review my 970M pro3 motherboard, when I was about to give up I kept telling myself "Buildzoid got 4.6 outta this piece of crap so it must be possible!"


Lovin it, AMD using hawt girlz to market Vega. THIS is what drives sales boys.



Would probably help if the girls were hotter than the cards though.


I havent put much research into it but on todays Pauls Hardware show they where speculating a vega card with 1080ish performance for $400-500USD. I have NFI what the pricing will be.

As I said before it comes out price competitive with a 1080 it means;
1. I good value buy WIN
2. Nvidia lowers 1080 pricing WIN
3. Its a lame duck and we all lose as consumers.

AMD's roadmap has NAVI slated next year to compete top to bottom on the GPU stack. Adored TV did explain this in a video posted a while ago that VEGA was not going to be able to take on 1080TI and Titan but AMD plan too. I hope the cash flows in to get them there and prices drop for customers no matter what color they are RED or GREEN.
Repost with time


So WCCF did an article today, that an aircooled vega reached "22,230 graphics score in FireStrike 1.1". Are they at least 50% trustworthy? They say 22000 but the picture shows 17447 score in Firestrike 1.1. I would love it to be between 1080 and 1080Ti and maybe even powerhungry, so that miners don't buy them all out.


... we all would.


Or better. That'd be sick.

(and if the FP32 figures indicated anything it should at least have the potential to do so, whether AMD can get the drivers to the point where it's fully utilizing its' potential is yet to be seen, and seems unlikely looking at past products)


Let me preface my bullshit by saying that I’m not a hardware expert within even Tolkien’s stretched imagination, so feel free to correct me where I’m wrong.

Past evidence would seem to indicate that AMD cards pick up somewhere between 10 and 15 percent in performance over their lifespans through driver improvements when compared to their nVidia competitors (from the same generation). We have already seen measurable, if not eye-catching, improvements in the Polaris line of cards from their initial launch. With the amount of hype leading up to this launch (largely perpetuated by AMD), a certain level of performance will need to be met by AMD at launch (roughly 1080 level at a similar price from what I’ve been hearing) or face a serious backlash from the consumers. Would it be a stretch to assume that the driver team from AMD has been focusing solely on Vega’s performance? I don’t think so.

With that in mind, I expect 1080 class performance at launch with subsequent patches letting it split the difference between the 1080 and 1080ti. I don’t think they are going to be able to fully recognize all of the performance the Vega chip can provide because, in AMD’s own words, it is very big change to the GCN architecture.

I think that in 18 months to 2 years after its launch, it will be beating (That might be too strong… how about competing with) the 1080ti. I might be reaching a little too much, but looking at the on-paper specs of the card, it easily competes at the 1080ti/TitanBS level. The biggest challenge the card faces is driver support, which is not something that AMD has excelled at. They have made strides in that area, but I think that hoping to have a 1080ti killer out of the gate is absurd.

Will AMD be able to produce the GPU/HBM/interposer package cheap enough for it to be a compelling option? No idea. It’s a big chip, the HBM system is not easy, or cheap, to produce. I don’t worry about the performance as much as I do the cost. I think it’s 2kliksphilip who says that there is no such thing as a bad GPU, only a bad price.


On paper, everything you say is possible. The compute figures indicate its potential to be the fastest gpu short of volta, but that’s the catch. 18 months to 2 years will be far, far into the Volta product cycle. We’re already deep into the Pascal product cycle. Vega looks to be a worst Fury IMO. A massive GPU with monster paper specs, that isn’t fully utilized until after Nvidia has released a counter and obsoleted AMD’s offering.


I couldn’t agree more. One of my goals was to explain a potential reason for all of the (over) hyping, and I thought the paper specs do lead to that kind of conclusion.

I also want people to think about the launch as an investment. I know the reasons “future proofing” is stupid, but you should only by a card of this caliber and price if you intend to hold onto it for a while (say, five years ± 1 year) (or you have way more disposable income than me so could you please send me $50? lol).


Oh no, let’s remember, people on these forums were saying “wait for Vega” prior to ANY information coming out. The fanboy hype was over the top and started way, way too early. People did the same thing for zen, but that one panned out.


If it drops pricing we win if not you win while everyone loses. I guess I had hope. What is it now 1 week from a launch paper or real.


I’m no moron, I’m rooting for AMD like everyone else. But I definitely was not among those telling people 8 months ago that they should wait for Vega. I want competition in the market just as much as the next guy. Advised my brother to go rx 480 and the build I did prior to that had a 390, I like AMD stuff. But there was some pretty big proponents of waiting for Vega and it turns out that betting on a product months in advance which we knew nothing about didn’t pan out this time. All I can do now is cross my fingers and hope AMD is getting great yields, figures out the drivers, and can actually turn a profit selling the thing.


I think the biggest mistake that AMD could make would be in my opinion that they market and price VEGA as HEDT GPU for 800+€/$. if AMD is kinda smart I hope that they compete with the 1080 = VEGA at a smaller price and over time release better GPUs and or better drivers.

the main concern here is that AMD has to get a lot of things right at the same time while handling a new architecture and on top of that they have to get Devs on board.

I think for now and for the Siggraph “launch” whatever that may manifest itself is some kind of launch, that we see some okay to good price/performance to compete with the 1080 and in future, we can hope to get somewhere into HEDT GPUs from two manufacturers.

as long as AMD doesn’t fuck up this overhyped prototype of their new architecture…


Well they have new shader tech so optimization may help a lot, but I will gladly buy a cheaper 1080 equivalent from AMD, just to help competition. I own a 1060 right now, it is great, but I dislike the NVIDIA strategy (logging in for updates, data collection, passthrough disabled, proprietary tech, …). Once new consoles get Freesync 2, G-sync will have a hard time.
I also hope the Vega will have good machine learning applications. They need to compete with nvidia also on that front.

PS: I really don’t envy the AMD driver dev team right now…


So a couple of years ago I remember seeing a nVidia road map that showed gddr & system memory unification sometime in the future and I told myself “I’m not going to buy a new graphics card until that technology is released”. And guess what AMD beat nVidia to the punch and vega has that feature I’ve been waiting 2 years for.
That’s why I’ve been waiting for vega to launch and I’ll probably buy it on launch day if I can.


so elephant in the room, the rx vega tour blind side by side demo was done vs a 1080ti.
apparently hardocp people recreated it with amd and did fresh install of 10 and nv drivers themself to make sure no funny business.

end result out of 10 people side by side at 1440p, 6 couldn’t tell a diff, 1 picked the ti and 3 picked the vega.

Whole lot of screeching about vega being utter fail, blind test it’s = or better than an independently setup 1080ti.

Yes I know they didn’t reveal muh fps, personally I don’t give a rats ass about some marketers best case scenario jittery ass max numbers from either side.

I’m willing to sit back and wait the few weeks to see how it plays out with pricing rather than bleed green and screech about 1070/1080fe beating vega (which they clearly won’t from a playability viewpoint), max fps spikes mean nothing, minimums, color accuracy and consistency are the real things to look at.


Pictures of Vega are circulating on Twitter, here are some links:

Water-cooled :arrow_up:
Air-cooled :arrow_down:

Looks like they did the press brief tonight, and we’ll probably see info in the next day or two. Coolio. Hardware Caucks said on twitter they have price and approx performance figures. Will be interesting to see what AMD managed to squeeze out with Drivers. Fingers crossed :crossed_fingers:



And Video: