So no compelling arguments then? Another year, another round of people trying to discredit speculation and disappointment from past patterns based on instinctual hope rather than sense.
So no compelling arguments then? Another year, another round of people trying to discredit speculation and disappointment from past patterns based on instinctual hope rather than sense.ill simply bookmark this
Nvidia man... Their aibs must be getting quite irritated at Nvidia eating those first mover sales.nvidia store only for now. I mixed up thinking it was 10am and checked back a few minutes ago and it was too late
So no compelling arguments then? Another year, another round of people trying to discredit speculation and disappointment from past patterns based on instinctual hope rather than sense.
you are simply talkin out of your ass what do you want me to argument? your "arguments" are some guesses on older products. according to you amd would have never been able to compete with intel ever again because bulldozer sucked. now amds cpus are competing and even with lower power consumption. what do you want me to argument? vega is the first iteration of gcn with tile based rasterization. its the same thing that made maxwell/pascal powerefficient.
you are simply talkin out of your ass what do you want me to argument? your "arguments" are some guesses on older products. according to you amd would have never been able to compete with intel ever again because bulldozer sucked. now amds cpus are competing and even with lower power consumption. what do you want me to argument? vega is the first iteration of gcn with tile based rasterization. its the same thing that made maxwell/pascal powerefficient.
No, my argument is more akin to "look at the performance figures AMD gave us for Ryzen, they're fairly accurate representations of Ryzen's best-case scenarios just like Polaris and Fiji's marketing material and really those of every hardware company ever; there's no reason for AMD's marketers to show us performance of a product that's still drastically improving because it would hurt the hype, but also no reason for them to hold back on their representative data and choose something other than a favorable workload".
Except now Ryzen launched and is just one more in a long list of examples demonstrating exactly what I'm saying. Ignore it all you want, but AMD have already deliberately shown us Vega's performance. That's a fact, period.
ryzens first estimates by amd were ~40% but later they even beat that. hows that? amd beat their own estimates.
the most funny part is that you are talking as if there is only one vega. what type of vega did they show us? early eng sample? maybe even cut down? how do you know what type of gpu theyve showed us? all of your so called speculations are nothing more than hot air.
Grats to those who get pre-orders! But dang, I overslept. Fighting a cold and took medicine so I was out. Wife called tried to wake me up.
Guess I'll have to wait for second batch (or hopefully retail has it date of release).
Oh well. At least my cold is subsiding.
The Doom demo was big Vega; the only Vega AMD have publically shown and tested.
Which they've stated were using Fiji drivers. It was also done many months before launch, so if everything was ready like you are assuming, why wouldn't they have launched it already?
I mean seriously you want to use the Doom demo where they had the whole case tapped shut with barely supported drivers as the best case win for AMD? It was ~10% faster than an OC'd 1080 even with those limitations.
You don't use "Fiji drivers" with a GPU of a completely different architecture no matter what some marketing guy tells you nor do you willingly give out performance numbers if they're simply liable to ruin interest. If AMD actually did that, you should be questioning the basic competence of whoever was in charge of the event.
People want to see the GPUs so they are releasing tidbits of info with big disclaimers of "this is not final performance numbers".
Amazing that people will say that AMD releases with poor drivers constantly, but when talking about demos 4+ months prior to launch they are happy stating that drivers are finalized...
Tidbits? Sure. Numberless demo? Okay.
Bacon1 said:Guess we'll agree to disagree, we'll see if performance is under 1080 like you claim it will be in a few months.
Nice, my order didn't say when it was shipping.Pre-Orders Up!
https://www.nvidia.com/en-us/geforce/products/10series/geforce-gtx-1080-ti
Ships March 10
He said about GTX 1080 performance in the AMD demo, while you mentioned ~10% faster than OCed GTX 1080 with the 'whole case tapped shut with barely supported drivers'. Yes, we'll see who's right in a few months.
He said about GTX 1080 performance in the AMD demo, while you mentioned ~10% faster than OCed GTX 1080 with the 'whole case tapped shut with barely supported drivers'.
Yes, we'll see who's right in a few months.
I don't claim it's under the 1080, just around (certainly think it's closer to the 1080 than either the 1070 or 1080 Ti, AIB to AIB). +/- 5 to 10% in either direction is effectively the same level and there hasn't been a direct comparison yet, just ballparks in the same area. So I wouldn't try to pinpoint a specific performance distinction like weaker or stronger than a 1080 which a few % from minor clock or driver changes can decide.Guess we'll agree to disagree, we'll see if performance is under 1080 like you claim it will be in a few months.
I don't claim it's under the 1080, just around (certainly think it's closer to the 1080 than either the 1070 or 1080 Ti, AIB to AIB). +/- 5 to 10% in either direction is effectively the same level and there hasn't been a direct comparison yet, just ballparks in the same area. So I wouldn't try to pinpoint a specific performance distinction like weaker or stronger than a 1080 which a few % from minor clock or driver changes can decide.
I don't know how heavily Doom favors AMD anymore after all the patches and stuff either, only that it does. And AMD's improvements these past couple years seem targeted at those spots where Nvidia-favoring games punish Radeons, so typical per-game variance might be tightened up a bit.
Yes. AMD being in an even worse position relative to Nvidia this year relative to the past wouldn't be a new thing at all. They've been lagging further and further behind since around the time Maxwell first launched. Arguably even earlier than that when you consider how long GK110 had free reign on the market before Hawaii launched. Now it's been nearly a year since Pascal's launch and AMD haven't even announced a 1070+ competitor yet while the 1080 Ti is about to launch and the Titan XP launched 7 whole months ago. GP100, GP102, GP104 - all 3 of these Pascal chips originally released last year across various markets and have absolutely nothing from AMD that could challenge any of them. Unprecedented level of inactivity from Team Red. Heck, when you look at aftermarket models/max OCs all around, the Fury X was actually very close to being closer to a 980 than a 980 Ti.So you believe the top end Vega GPU will be even WORSE than Fury X relative to the competition?
Probably not, no. Excluding the fact that Doom and Vulkan were pretty much coded for AMD/from the ashes of Mantle, their driver team would have made Doom their number one optimization priority from the start so it could be ready to demo as nicely as possible. Management would have made sure of it. These companies don't just haphazardly operate on whims and impulse ya know. They weren't packing up for the show when they suddenly decided to demo Doom and show the unlocked FPS while they were at it. It's all considered and planned for beforehand just like the Hitman and Battlefront demos for Polaris. If their current drivers weren't sufficiently up to par with what they knew Vega could do, they wouldn't have presented the demo as they did. People get paid to think about and plan these actions out very carefully.Edit: Nvidia just launched a DX12 driver that boosts performance. I guess AMD can't release any driver that boosts Vega performance between Demo and now.
Oh well... not like AMD isn't historically massively dependent on getting good drivers down to be able to compete with Nvidia who has FAR more polished drivers (performance wise this is all about performance when talking about drivers).
Don't expect consumer Volta in 2017. It's a 2018 product, just look at what Micron is saying about GDDR6 availability. That's when you should expect Volta.
Is this^ common knowledge? I had thought NVidia was releasing Volta around the holiday season in 2017? Did something change, or has it always been a 2018 release?