A Facebook post from AMD Poland has started rumours flying around the interwebs as the implication is we will see it arrive this quarter; though perhaps taking a Facebook post as a verified source may lead to disappointment. However, there is a bit more evidence than just the post which Digital Trends has displayed, we have seen CompuBench results of a mysterious AMD GPU which is very likely to be Vega. It will have 64 compute units which translates into 4096 stream processors unless something very strange is going on. The benchmarks also list two frequencies 1GHz and 1.2GHz which indicates computing performance of 8.2 TFLOPS and 9.8 TFLOPS respectively which puts it in the neighbourhood of the GTX 1080's 8.9 TFLOPS. That will not necessarily directly translate into gaming performance but does indicate that AMD has a nice surprise in store for us.
Don't forget to add a bit of salt to your rumour consumption and keep your eyes peeled for more information.
"Of course, it’s entirely possible that someone at AMD Poland spoke out of turn, and so this information should be considered unverified at this point. But there’s other information that seems to confirm an imminent release, such as a Vega GPU showing in CompuBench benchmark results — something that usually happens shortly before a new component is released."
Here is some more Tech News from around the web:
- AMD: Vega graphics cards are coming this quarter @ The Inquirer
- Hackers uncork experimental Linux-targeting malware @ The Register
- US ISP Goes Down As Two Malware Families Go To War Over Its Modems @ Slashdot
- For what it's worth, the Windows 10 Mobile Creators Update is rolling out @ The Inquirer
- Another ZX Spectrum modern reboot crowdfunder pops up @ The Register
- Microsoft cracks open patch mega-bundles for biz admins, will separate security, stability fixes @ The Register
- COUGAR Armor Gaming Chair @ Benchmark Reviews
- We review the “world’s best standing desk” – the Autonomous SmartDesk 2 @ BabelTechReviews
If those numbers are right I
If those numbers are right I am hoping it is a misdirect by AMD and this card is actually running a lot faster than what was listed and has more performance than what was listed in the chart because it looks to be so far behind the 1080 Ti and with all the hype of Vega it would be a pretty big stain for AMD to come up this short of the expected mark and only be able to maybe match a year old released product and not the Nvidia high end product that was just released a month ago. AMD has done the misdirect in the past so here is hoping they are doing that now.
I actually am planning on buying a Vega based card just because I think they will be better for the long term with DX12 based games than Nvidia current line up and probably Volta as well.
I could be wrong and maybe this is actually considered a good thing..lol
Indeed, if these #s are
Indeed, if these #s are correct, it’ll be a horrible blow to any of the momentum Ryzen has brought to AMD. The stats for the card listed aren’t much better than a 1070. And in some tests the 1070 crushes it.
It’d be swell if AMD would have another ‘tech day’. If nothing else to at least, ‘manage expectations’.
Who cares it’ll be a
Who cares it’ll be a disappointment like everything else AMD has come out with in the last decade.
Assertions like this make you
Assertions like this make you look foolish and only serve to invalidate every other assertion you might wish to be taken seriously.
Last decade? Nope, not
Last decade? Nope, not everything was that bad from them this last decade. And don’t forget they are fighting against two financially stronger companies. Well, Intel is more than just financially stronger. Think your self playing a game of chess against a much stronger player, where you start the game with only half of the pieces. And think of you managing to survive in that game at least a dozen moves without losing. That’s what AMD managed to do this last decade. And if Ryzen manages to create enough income for them, you can expect much less disappointments in the future.
AMD hasn’t been that far off
AMD hasn’t been that far off of par even at their worst in the past decade. Their real failures have been in the adoption of support for their products by the industry. Especially gaming. It can be real hard to go for AMD when every game you see is optimized for and uses all the fancy flash shit for NVIDIA and ignores the AMD side. But that isn’t exactly AMD’s “fault”. It’s just the nature of the market right now where they have so little adoption. It’s sort of a catch-22 where adoption may not occur until devs give as much attention to AMD as they do NVIDIA stuff but they may not do that until there is more adoption.
I personally expect something
I personally expect something around the GTX 1080. Probably faster than the stock GTX 1080 but with limited overclocking headroom with AMD already having to push it higher to beat the 1080. If they don’t come up with a miraculous version of GCN, we shouldn’t expect much more. Probably a Fury X with a frequency at around 1.3GHz and an extra 5-10% performance because of architectural improvements. Power consumption will be close to 200W probably. And only 8GB of VRAM.
If Ryzen produces enough income for them, maybe AMD will be in a position to threaten Nvidia in 2019-2020. But with empty pockets, we only can ask them to stay competitive in the mid range market and offer limited at least competition in the high end.
For those hoping to see strong competition in the high end, they should realize that every time they happily attack AMD, or others in forums suggesting AMD hardware(people do this a lot, even in cases where AMD’s hardware is better for the needs of the person asking for advice), they just help Nvidia and Intel to maintain their monopoly. People should just be patient. Ryzen is better than expected. Wait a couple of years and we might get a better than expected GPU.
I’d say it would be quite
I’d say it would be quite disappointing if all they manage is something that competes with the 1080 (and 1080 v2 or whatever it should be called, with faster memory).
AMD have played up expectations a lot, and plenty of people stand to be very disappointed.
Ryzen was a welcome addition to the CPU market in general, but not really interesting to me other than as competition to push Intel to do better.
AMD is taking much more time
AMD is taking much more time to come up with a product, because of financial difficulties. The more it takes for a rumored product to come out to the market, the more people are hyping what it will bring. And the more time it takes them to finish that product, the more and faster products the competition is announcing. And as competitor’s products become faster, the expectations of fans get from reasonable, to unreasonable levels. So, AMD fans who where expecting a Vega that will beat the GTX 1080 10 months ago, where talking about Vega beating 1080 Ti two months ago, and now they expect Vega to be as fast or faster than Titan Xp. If Nvidia was announcing GTX 2080 tomorrow and that card was 30% faster than Titan Xp, some people would be talking about Vega being as fast as the GTX 2080.
In the end, I don’t think that all this hype is AMD’s fault. it’s more or less people’s fault. People who spend all day on the internet arguing about who’s company’s product is better, create all this hype and come to unreasonable expectations about Vega’s performance. AMD only described Vega as the brightest star, and it is going to be compared to AMD’s other products. But going against 1080 Ti is just too optimistic. Way too optimistic. It’s still GCN, it’s still 4096 stream processors, it’s still GlobalFoundries’ 14nm.
Because of this hype-train
Because of this hype-train avalanche, I think it would have (now that they’ve confirmed Vega isn’t delayed and definitely this quarter), been better to delay Vega until after Naples or the rumored HEDT 16 core Ryzen. Get something out that’s demonstrably better than the competition’s offerings w/superior margins.
It is unfortunate, at this point, that even if they squeeze out a part that is == to a 1080, at half the price, they’re going to get mocked/maligned. “AMD Disappoints” is a likely headline. As you said, there’s so much hype that it’s really not possible for them to ‘win’. Especially given the similarity wrt to price/perf w/Ryzen achieved the same sentiment for many analysts and gamers. The binary, ‘It wasn’t #1 in everything so it sucks in all ways.’ view is annoying and severely myopic.
Naples seems like a much better release for them. Strike now, while Intel has no answer, and get back some of that 25%, IIRC, server market they’ve effectively completely lost in the last 11ish years.
Naples is their most
Naples is their most important product, much more important than Ryzen and Vega. That’s why we have seen Zen in the desktop platform first, so they have a huge user base and enough time to test and optimize the architecture and the new platform. The HEDT models probably will be the final test before pushing Naples to servers.
With so many years Intel and Nvidia dominating the high end, it’s no surprise that many people have become fanboys of those two companies. It’s the same as in politics or sports. People become usually fans of the biggest political party or the strongest team. They need to be in the winning side. So, no surprise seeing people saying ‘It wasn’t #1 in everything so it sucks in all ways.’. At least the press has changed it’s attitude this last year. It’s not happily attacking AMD for no reason, while covering most of the bad stuff coming from Nvidia or Intel. Articles are more balanced compared to the past. It took them a while, but they finally realized that without AMD, half of the tech sites having PCs as the main subjust, would close.
These numbers are the same
These numbers are the same ones that were taken back 4 months ago when AMD didn’t even have a working graphics driver for Vega. They were using Fiji drivers that had zero clue on how to run Vega architecture. No one has even seen any official data numbers using drivers specifically built for Vega yet.