Should I laugh or should I cry ? This is my result on a GTX1080 which was supposed to be slower than Vega: http://www.3dmark.com/fs/10146244 Graphics Score - 24031 Yes, it is a bit overclocked... but still... dang, I have that card for a year, and 1080Ti's are scoring WAY, WAY higher (Up to ~30000 graphics score)

I don't know what AMD was thinking with this release. They wanted a card like a Titan, but I guess the release is rushed so the gaming performance is terrible with the current "gaming" drivers, so Raja basically says "not for gaming" but they release a gaming mode for it anyway.. They should have either waited for the drivers to mature before releasing the card, or waited to release the gaming driver only until it was ready. AMD's community takes pride in the fact that AMD is more willing to interact with users. I really think they should take advantage of that and explain the situation, why the gaming performance is low, what they are going to do to fix it, timeline for the fixes, etc. I think they would received a way better response then not even sending review samples out and sending mixed messages about the target market for the card itself.

I don't know what AMD was thinking with this release. They wanted a card like a Titan, but I guess the release is rushed so the gaming performance is terrible with the current "gaming" drivers, so Raja basically says "not for gaming" but they release a gaming mode for it anyway.. They should have either waited for the drivers to mature before releasing the card, or waited to release the gaming driver only until it was ready. AMD's community takes pride in the fact that AMD is more willing to interact with users. I really think they should take advantage of that and explain the situation, why the gaming performance is low, what they are going to do to fix it, timeline for the fixes, etc. I think they would received a way better response then not even sending review samples out and sending mixed messages about the target market for the card itself.

Why does it really matter? The price of this is what, $1500? Anyways, it's way over a 1080ti and we know the Ti is faster and it only costs $650. Surely nobody will buy less for more(gaming wise)

Where's the Only Intruder guy now who had a lot to tell me on some other news post? I didn't bother to reply to him in that thread. Well, here's proof of the performance. All AMD is good for is providing you with a heater free of cost in colder countries. Good luck in hoping to see a 50% performance gain with driver optimization alone.

Why does it really matter? The price of this is what, $1500? Anyways, it's way over a 1080ti and we know the Ti is faster and it only costs $650. Surely nobody will buy less for more(gaming wise)

The AC edition is $1000 the WC edition is $1500, they dropped the prices yesterday. I guess it matters because the same hardware is being utilized in the RX variant (minus 8GB of HBM) as far as we know. So if this card has no legs with the gaming drivers, the RX variant will probably be similar.

And if they do not release a Vega in June, the wiseguys will shout "They still don't have any Vega cards but in July-August".

Seems infinitely better then releasing a card that underperforms, with an initial driver release of known issues that's a mile long. On top of not even providing hardware to reviewers and controlling the story.

Should I laugh or should I cry ? This is my result on a GTX1080 which was supposed to be slower than Vega: http://www.3dmark.com/fs/10146244 Graphics Score - 24031 )

WTF How can ur score so much high.My 1080 with i7 6700k(oc) score 5423. this cant be that much of difference between i7 6700k and i7 6800k difference of 14000 score? can anyone explain it this legit?

I could repeat the mantra that I'm actually certain is true, that the drivers aren't even close to ready, but then why even release them with a driver like that AMD? :S

Should I laugh or should I cry ? This is my result on a GTX1080 which was supposed to be slower than Vega: http://www.3dmark.com/fs/10146244 Graphics Score - 24031 Yes, it is a bit overclocked... but still... dang, I have that card for a year, and 1080Ti's are scoring WAY, WAY higher (Up to ~30000 graphics score)

Becouse you look at gaming benchmarks with preview drivers when you are interested in a professional GPU like a Quadro or Tesla..

To all above, card isn't for gaming. Full stop.

Becouse you look at gaming benchmarks with preview drivers when you are interested in a professional GPU like a Quadro or Tesla..

That's a good point, but even if the point is 100% correct, it just highlights the current state of the AMD gaming driver for the card, and the idiotic decision to launch it with it.

Very disappointed in this news. Is AMD going to be able to sell the gaming focused card at $450? His fire strike score looks normal. Post yours.

My bad i have score in Fire strike Ultra thats why score is lower.Heres my score for both firstrike:- http://www.3dmark.com/fs/11182818 http://www.3dmark.com/3dm/20739977?

That's a good point, but even if the point is 100% correct, it just highlights the current state of the AMD gaming driver for the card, and the idiotic decision to launch it with it.

What does it matter is the current state of of drivers for professional application. Non professional application performance do not matter at all. Moreover on such cards some genral "gaming" optimizations are disabled.

To all above, card isn't for gaming. Full stop.

Then why give it a gaming mode and say you can switch between the two? Why compare it to a Titan XP, which people constantly call a "gaming" card? At the very least, as PRMinister said, this shows the state of the gaming drivers for the RX variant. Which as I said, as far as we know, is this card minus 8GB of HBM.

What does it matter is the current state of of drivers for professional application. Non professional application performance do not matter at all. Moreover on such cards some genral "gaming" optimizations are disabled.

The current state of drivers for professional applications is worse than Quadro cards at the ~same price with half the TDP. Not to mention the known issues list includes a number of professional application issues. It just seems like they released this card because they were afraid of going passed the 2H date like BREAL said. Which is a huge mistake imo, because first impressions are everything. At the very least they could have gave it to reviewers along with a guide that explained why it's in this state - at least then they could somewhat control the damage and explain it. Now it's just going to be speculated that RX Vega will perform equivalently and people will just go buy a Ti or whatever.

Currently it seems to fair worse then Fury X in a sense. So everything is wrong.

Currently it seems to fair worse then Fury X in a sense. So everything is wrong.

I think in the end the drivers will shake out and the performance will be relatively fine. I don't think these results are indicative of the actual hardware. It's just a really bad way to release a new series, especially one as hyped as Vega is.

What does it matter is the current state of of drivers for professional application. Non professional application performance do not matter at all. Moreover on such cards some genral "gaming" optimizations are disabled.

If that was the case, why release with gaming drivers at all?

I think in the end the drivers will shake out and the performance will be relatively fine. I don't think these results are indicative of the actual hardware. It's just a really bad way to release a new series, especially one as hyped as Vega is.

Me too, to be honest, but I would prefer it to be a bit more delayed than bad.

All this comments. Just wait for the gaming version of this GPU, maybe it will be clocked higher, also it seems there are issues with the drivers, so like any AMD GPU, performance will rise, a lot with future drivers. I should know.

All this comments. Just wait for the gaming version of this GPU, maybe it will be clocked higher, also it seems there are issues with the drivers, so like any AMD GPU, performance will rise, a lot with future drivers. I should know.

You're failing to address the main issue which is that this GPU was released with gaming drivers.

Where's the Only Intruder guy now who had a lot to tell me on some other news post? I didn't bother to reply to him in that thread. Well, here's proof of the performance. All AMD is good for is providing you with a heater free of cost in colder countries. Good luck in hoping to see a 50% performance gain with driver optimization alone.

What's with the attitude? I said not to make snap conclusions because we didn't have any proper numbers or benchmarks. Now we actually have numbers popping up and we can say something about it. [spoiler]

And how have you come to that assessment? Can you show me the source of your information please? It's like the Fury X, it actually has more power than the 980 ti but it turns out that under particularly serial workloads, the GPU tends to be starved of work and performs badly. From what we know of Vega, most of the bottlenecks in Fiji have been ironed out and if polaris performance is anything to judge by, it should perform rather well or at the very least, comparatively to the 1080 range. It's still too early to draw conclusions, all we know from this news statement is that Vega needs a lot of power, it doesn't tell us how efficiently Vega is going to use that power, if it's like the Fury X, then yes, it quite possibly could be disappointing and conversely if it shows similar performance scaling to polaris, we might be in for a treat. Tl;dr wait for reviews before making snap conclusions.

[/spoiler] And as you can see in my post, I kept an open mind, because we only had snippets of information. You really need to get out of this attitude of being hostile towards others, I've seen you make posts like this before, particularly towards the user, Undying, it's impolite and disrespectful. Anyway, on topic, it certainly doesn't look good but apparently, the card is throttling because it's being starved of power (Onna states in his thread that the PSU used is 550W can anyone confirm this?). There are two things we can derive from this though, 1. Vega is not very good or 2. something is very wrong with software or hardware configuration. Maybe it really is a case that this revision is not meant for gaming but for compute purposes which if we believe AMD's statements then yes that would make sense but how do we know AMD haven't been saying that, just as a means to pass off the performance we've had a glimpse of. I'd rather prefer to see actual proper benchmarks anyway from the likes of Guru3d and Anandtech. Personally I reckon we still have to wait and see, hopefully the consumer version of Vega will indeed live up to expectations but if it doesn't, well I guess the graphics card industry is not going to look good in the future.