LinkBack URL
About LinkBacksRead more.And a card with the same name has been spotted in an HIS Digital shipping schedule.
Read more.And a card with the same name has been spotted in an HIS Digital shipping schedule.
Sounds plausible. A bit meh to be honest, a stop gap until 7nm vega/navi but I'm sure it'll be good value for those in the market at around the gtx 1070 mark.
Performance-wise it'll be more akin to the 1060, if it's 15% faster than the 580.Originally Posted by chj
if it's going to be akin to a 1060, it needs to be power efficient too...
Considering you can get an RX570 8GB for £160 delivered with 3 games,I think the equivalent GTX1060 3GB is a hard sell long-term.
Looking at the Hexus Turing reviews the RX580 is now consistently ahead of the GTX1060. Even the last set of RX570 reviews on Hexus had the card in the 4GB version trading blows with the GTX1060 6GB,and the GTX1060 was one of the better pre-overclocked cards and had more VRAM.
Another 15% would place this much closer to a GTX1070 than what people suspect.
Last edited by CAT-THE-FIFTH; 15-10-2018 at 04:01 PM.
The RX580 is a juicy old thing and a die shrink isn't going to perform miracles.
Little surprised to see AMD still working on Polaris in the midrange a year after Vega launched.
At launch, but optimisations on the 480 drivers pushed it past the 1060 a rather long time ago (as in before it became the 580 even).
http://www.hardwarecanucks.com/forum...review-23.html
I was hoping a x90 series of this range would be a dual gpu, i knew it wouldn't be but one can dream!
The 580 4GB currently outpaces the 1060 6GB in most games, so 15% faster would make it comfortably ahead of itPerformance-wise it'll be more akin to the 1060, if it's 15% faster than the 580.
Yea, that seems to be AMD's thing though: Release a product, then milk the crap out of it for 3-5 years, and give it 5 different names..
Another AMD launch and another ... why did I even think they might pull a rabbit out of the hat?
If this is just a little bit better than the 1060, well woopee do, over two years after the release by Nvidia of the 1060 AMD can now say 'we can beat that'. No wonder Nvidia feel they can screw the public on card prices, the competition is a joke.
Pretty sure this wasn't an Intel thread
The only people screwing themselves are enthusiasts and gamers who buy the cards at those prices and then excuse make for it. If they didn't then Nvidia would reduce them. An example was tech forums "justifying" the Titan series pricing due to "reasons" years ago,and then using it as an excuse to "have" to buy them at a high price. If people nowadays had any degree of self control,then companies would not be able to charge more,but they don't so pay beyond any price for these things,especially with cheap credit still a thing.
For example Gibbo over on OcUK said they already sold 500 Core i9 9900K CPUs on pre-order at £600,even though its £100 over the RRP(and its cheaper to import it from the US with all duties paid) and people know its over RRP. So all that inflated Intel CPU pricing(due to the shortages),well expect that to start affecting overall Intel CPU pricing longterm,and if AMD decides not to push too hard with 7NM Ryzen pricing,then expect AMD prices to also go up.
Even when AMD/ATI beat Nvidia in performance and time to market,they still bought Nvidia cards(think Fermi or the FX),so most of this is the fault of consumers. Valve told people not to buy the FX,and even showed graphs on how poorly it ran HL2. Not even GaBen had enough sway,and ATI sold less 9000 series cards than the FX and that was years ago. ATI would have probably folded if AMD hadn't bought them.
No wonder AMD decided that concentrating on CPUs made more sense in recent years,so expect things not to get any better. This is basically what they thought years ago and they said that.
The times of you expecting AMD/ATI to make your Nvidia cards massively cheaper is not going to happen anymore. So none of that HD4000 series moment. You better hope Intel actually is actually entering the graphics market aggressively.
Last edited by CAT-THE-FIFTH; 15-10-2018 at 06:54 PM.
Do you honestly think Intel is going to enter the graphics card market with anything other than overpriced goods? Honestly? AMD price their cards (with the exception of Vega) fairly reasonably at the moment. Same with Ryzen while they try to regain market share. Then what happens? Nvidia and Intel drop pricing to try to be competitive? Doubtful as they both have overheads to cover due to the sheer size of them.
Longer term, the prices will increase from all three of them. They'll be able to justify it due to people having no degree of self control, as you already mentioned.
You should know that Polaris also packs A LOT more Teraflops than GTX 1060... which will be more power consuming... and the 12nmLP node is still designed for low clocks and mobile parts.The RX580 is a juicy old thing and a die shrink isn't going to perform miracles.
Little surprised to see AMD still working on Polaris in the midrange a year after Vega launched.
NV used a node designed for high performance and efficiency (a polar opposite of GLOFO node), which allowed NV to also clocks Pascal 30% higher... which strangely enough didn't exactly produce that much of a difference since Polaris and Vega are either similar, same or better than their NV counterparts with those lower clocks (Suggesting that Polaris and Vega have quite powerful architectures and IPC... but they are bottlenecked by the manuf. process).
Also, AMD usually ships their GPU's from factory with 1.2V set as default... that's a massive increase in power consumption right there.
Undervolting the GPU's works and brings down power consumption far more in line with NV counterparts... but people seemingly forget this.
Also, the 7nm process is a polar opposite of the GLOFO one. TSMC 7nm is designed for performance and efficiency, so the node bottlenecks will be gone.
Though AMD should really work on shipping GPU's from factory with lower voltages... but part of that was due to AMD wanting to maximise the number of dies that actually work.
I'm thinking that since TSMC has more experience at making GPU's... failure rates should be lower, so voltages could go down from factory.
Intel has done contra-revenue with Atom which cost billions of USD,so there is the chance they might try to play a bit dirty by giving major financial incentives,and for them even a few hundred million to get marketshare is really nothing. Look at how much they spent on McAfee and how much they lost. The main issue is whether their consumer graphics card is going to be really any good at gaming,ie,is it more for compute stuff. If it isn't really gaming focused,then we will be again relying on AMD,and sadly there is a reason why AMD does not want to be the "budget brand" when it comes to graphics cards at least.
Imagine instead of pushing those early jumps to new nodes before Nvidia,and spending all that R and D on the HD4000,HD5000,HD6000,HD7000 and R9 200 series,they had spent that on CPU R and D instead?? Say instead of Bulldozer we had a far better successor to the K10?? AMD would be a much bigger company than it is now - when it comes to CPUs,they have far less reputation issues than with their graphics cards. How many times has ATI/AMD matched Nvidia or even beaten them at the high end,yet its never really helped sales?
The moment they don't even beat Intel convincingly in all workloads with a new CPU,which had its own set of teething issues at launch,their revenue has just skyrocketed.
The problem is why AMD GPUs do have issues with power consumption,etc since AMD pulls dual uses with their GPUs for gaming and commercial stuff,hence Vega is probably poor utilised for what it is(Pascal does not even do FP16 that well for example and its hardly used in many games either) and has a huge chip. HBM2 makes more sense for non-gaming desktop cards. Once Maxwell came along AMD were screwed - Nvidia split their lines entirely,with one set of GPU launches never really seeing gaming equivalents,so you have an issue of AMD being jack of all trades,master of none.
Now,the cunning ploy with Turing is that - all those tensor and RT cores,are not only there for gaming - Nvidia wants to get into the commercial VFX industry:
https://twitter.com/nvidiadesign/sta...006528?lang=en
So instead of making 3 classes of GPUs,one for AI,one for RT,and one for consumer gaming workloads,which will cost a ton of money,Nvidia has made sure they can quietly shoehorn all those features into games,ie,RTX,DLSS,etc.
Hence,Turing is Nvidia now going back to one size fits all GPUs,but making sure all that added functionality has a usage. However,unlike AMD they can push for this - AMD is too thinly spread. R and D has to be spent on BOTH CPUs and GPUs.
For me I hope AMD stop going after the high end consumer market now. If they are to make big GPUs,make it specialised to the commercial markets first,since they can be integrated as part of the servers as a package and work with the vendors. When it comes to gaming I hope they stick to the midrange,and engineer something that can be sold profitable at that level,and can help keep them the consoles.
In some ways we are seeing this already - the first 7NM GPU is a Vega one made for AI. Navi will be apparently more a value orientated GPU.
Now,AMD is doing the keynote at CES 2019 though,so it could be quite possible they have something up their sleeves.
There are currently 1 users browsing this thread. (0 members and 1 guests)
Posting Permissions
