AMD on the Road: takes Radeon RX Vega to the Gamers

Why would AMD release a card to compete with the 1080 which is ~20% faster than the previous generation 980Ti when they already had the Fury X to compete with that (and it already competed well, disregarding VRAM limitations - and now competes better)?

Is Vega going to be 20-30% faster than a Fury X? They might as well not release a card at all.

It doesn't make sense to me for AMD to be targeting this card against the 1080. Not at all.

 

We don't know anything for sure about Vega. But read somewhere that the FE has the same chip as the gaming Vega will have. If that's true, then means the only difference will be bios and drivers. And that means is very unlikely you will see the massive jump in speed people are really hoping.

But then again, doesn't make sense the early year Vega previous appereances, where performance shown matched or exceded the 1080, and 6 months (or more) later, and still showing same or worse performance. It doesn't make any sense at all.

 

They arent. People who dont understand the HPC market are just talking a bunch of nonsense basically.

 

Explain that to me, late to the game. Or others saying that level of performance, which AMD doesn't offer now and having a 1440p to 4k capable gaming card with Freesync will be a great thing, is a failure.

The 1080ti was released this year correct? It's NVIDIAs top of the line GPU, who's price was by NVIDIA standards slashed based purely of speculation about Vega, which is the only reason why 90% of those lauding it are even talking about it as a viable mainstream buy, which it is not.

Based on you comment, you would think an $800 gaming card was the top selling GPU and is what most gamers buy or can afford. So if AMD provides a better price/performance (including freesync) package, it's a bust because it doesn't eclipse their one top card, that few can afford for gaming, the ti?

Explain that reasoning to me ...

 

Irrelevant. Nvidia's net income is $1.67 billion. Surely, they're not struggling to pay off marketing, manufacturing, and R&D.

Source:
http://investor.nvidia.com/financials-statements.cfm

 

Magnitudes more than AMD? That just reinforces my point. You can't call Vega a fail by any means IMO.

 

From a gamer perspective, yeah, it's reasonably competitive compared to what Nvidia is offering. From a technological/financial perspective of both companies, it's a problem.

Ti launched this year - but the architecture it's built on is over a year old. Nothing prevented Nvidia from launching a Ti sized card last year - other then them wanting to reap sweet margins on their Founders Edition cards. In real world, Ti is nearly 40% more effective perf/w. To make matters worse, Nvidia is already shipping GV100 cards to select partners. Theoretically, Nvidia could ship a GV104 with the same perf level of a 1080 - Nvidia claims a 50% perf/w increase on Volta compare to Pascal, making it a ~120w card that would perform at the same level of the Vega FE at ~300w.

Obviously the price would be set in such a way that FE would still be reasonably competitive - but the entire point is that it gives Nvidia an advantage in options. They don't need GV104, instead they can focus everything on growing an entirely new market where they can sell cards at 10K a pop and companies won't bat an eye on buying them. That money goes back into R&D, which leads to better products and more advantage, if they choose to do it, which they seem to be.

In the mean time AMD is forced to take massive hardware risks that lower margins and increase the time to market. Their options are limited by comparison. And from that perspective it's a problem, I don't know about "fail" but definitely a problem.

And I get people sit here and say "why did they even build Vega, it makes no sense" or "why didn't AMD go x route" but what they fail to realize is that AMD had to have made those decisions nearly two years ago, if not longer. That's not an easy decision, especially if on paper Vega looked like it was going to be a win for them. But in the end those decisions become harder when the products you are shipping show up ~6 months after competition and you're forced to sell them at lower prices than what you wanted.

I also think that AMD's marketing doesn't help in the least bit. Vega FE is like a Titan that isn't really good at anything. People say it's not a gamer card, but it's not a workstation card as it's not certified and AFAIK it doesn't have HPC drivers yet either. And if RX Vega was going to be substantially better, AMD could easily clear the air with a single tweet, but they haven't yet. So either it's not that much better, or AMD's marketing department is missing the dozen or so posts I see each week about people saying "i'm not waiting for Vega anymore and buying a Ti".

So again, while I don't think it's a "fail" it's certainly not a win. If anything it's looking more like a repeat of the Fury X vs 980Ti but slightly worse. And I don't think that what AMD intended.

 

Success in R&D, to a very limited extent, is linked to expenditure. Just compare Intel and AMD budgets for R&D and current processors. Inventors of transistors, VLSI circuits and many other groundbreaking technologies often had very limited budgets. When we evaluate the effectiveness of R&D, we always need to take into account the quotient, cost/effect. In this category AMD wins with both Intel and NV.
The question remains whether AMD's budget is sufficient to enable new technologies to be discovered. In my opinion, completely. NV boasted of issuing 4 billion on Pascal. Did the technology used in Pascal knock us on our knees?...

 

AMD will be over a year later then Nvidia in delivering a high end GPU to the market (referring to the GTX 1080). That's obviously bad for us as consumers and even worse for AMD since they have not had a single option to offer for those interested in more then a mid range gaming rig (Rx 480 is great at 1080p, not so much for higher resolutions). In order to make up for being so far behind time wise they need to deliver a GPU that substantially beats the GTX 1080.... in order to beat the GTX 1080 by what I consider a substantial amount AMD needs to deliver a product that roughly matches GTX 1080 Ti performance. I really want AMD to deliver that product, but I don't think it looks like they will.

When I say they're late to the game it's just stating the simple obvious fact that they are over a year behind Nvidia and therefore late om offering anything for the consumer to consider buying in the high end segment (let alone enthusiasts). That doesn't mean I dislike AMD or their products, and I am rooting for them to give the market something to get excited about.