• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

AMD VEGA confirmed for 2017 H1

Status
Not open for further replies.
Behave yourself Tommy lol....maybe on desktop idle ;)
hence why every air cooled card sounds like a jet engine. ;)

I'm sitting with (they tell me)an 'over engineered' 1070 Amp Extreme, the fan doesn't kick in until it hits 69c, it generally hovers about 65c, fan hardly ever kicks in-if at all in the mammoth Primo case that used to house a wc'ed 290X CrossFire system running 0rpm fans@desktop.:p

Not suggesting all high end Nvidia cards won't sound like a jet engine as I'm only on the (historically)mid range 1070 Pascal chip which is already blindingly fast but doesn't require jet engine coolers, but quiet can be bought.:)
 
What is the point for selling 4GB high end card in 2017?

The only logic is that they are selling it against GTX 1070 and most people are not stupid to buy a 8GB card over 4GB.
Because more versions means more money for them. Just a wild stab in the air that.
 
It needs developer support. HBC require coding from developers and it means that every game require coding for HBC just like Prey ,which will have VEGA features.

It is a bit more complex than that in regard to how and what it does and doesn't work with - however as per the examples they found that games actually were using around 2GB (in my experience often around 2.5GB) in future that will increase so 4GB isn't a lot of headroom for a true high end card and higher resolutions.
 
It needs developer support. HBC require coding from developers and it means that every game require coding for HBC just like Prey ,which will have VEGA features.
no it doesnt, it's just an option you turn on or off from the driver control panel, Deus EX wasn't coded for it, they already said so at capsaicine event, the real question remaining is , does it impact performance negatively compared to when you could have extra packed ram, if so how much.
if there is no negative impact, or a negligable one, the amount of memory then becomes irrelevant, just X number stacks to set the bandwidth.
 
4Gb Vega strikes me as a 8800gts 320mb sort of card.

If it can bring out a performance and price difference of 8800GTS VS 7900GTX of its time it'll be very decent really.

Honestly hope the 4GB is only relegated to the small very cut down Vega chips though. Like a 460 vs 480 type of deal.

HBM 2 being limited to around 8GB max per Vega die isn't great, but it's decent; 4GB for a highend card is a no no now.
 
It needs developer support. HBC require coding from developers and it means that every game require coding for HBC just like Prey ,which will have VEGA features.

I heard the complete opposite. It's an out of the box feature, because it is hardware managed, not software managed.
So in Dx and OGL it works without support, in DX12 and Vulkan the devs need just need to enable it (few lines in the code) again because it is hardware managed, they don't need to write long optimization codes. Just enable it, and (in theory) the hardware will do the rest.

Edit: Just seen AlamoX mentioned it already.
 
It does not matter how much VRAM a card has, performance is what matters. Would you refuse to get a card if it had 1GB of VRAM and outperformed everything else? As long VEGA performs well at a good price I will get one.
 
It does not matter how much VRAM a card has, performance is what matters. Would you refuse to get a card if it had 1GB of VRAM and outperformed everything else? As long VEGA performs well at a good price I will get one.

Of course it matters as we know games need a certain amount of VRAM to run. 1GB would simply be impossible. 4GB wouldn't be enough for most at higher resolutions. As already mentioned 4GB is 4GB, 8GB is 8GB... that speaks for itself. VEGA isn't going to be troubling the 1080Ti, it's going to be in and around 1080 performance. I don't doubt it will be a decent performing card, but the price is going to be the key. I don't think it's going to be a bargain though, it will simply be priced accordingly and finally give AMD fans (and those running Freesync monitors) an option at the high end.
 
I will put 10 bags of haribo down that at least one of the vega cards will be 16Gb... I thought that was the whole point in them pushing towards HBM2...
 
I thought the whole point of the HBCC was that huge amounts of on-board RAM were not necessary? Where new and innovative technology is concerned, bigger numbers do not always equal "moar!!11!"
 
Last edited:
Of course it matters as we know games need a certain amount of VRAM to run. 1GB would simply be impossible. 4GB wouldn't be enough for most at higher resolutions. As already mentioned 4GB is 4GB, 8GB is 8GB... that speaks for itself. VEGA isn't going to be troubling the 1080Ti, it's going to be in and around 1080 performance. I don't doubt it will be a decent performing card, but the price is going to be the key. I don't think it's going to be a bargain though, it will simply be priced accordingly and finally give AMD fans (and those running Freesync monitors) an option at the high end.
i dont know if i explained it badly, but Vega is moving to a new way of handling memory with HBCC, it's not wise to compare classic method to another one we never experienced before, it might turn out different where bandwidth is much more important than size, and AMD touched this subject many times, saying that they will need to educate ppl that bandwidth is more important than how much memory is packed ( again not in a classic GPU, but for an HBCC GPU ).
i do believe Vega wont be priced over 500$ ( my guess 329$ and 479$ ) again koduri spoke many times for the fact that AMD wants a 4k capable PC for about 1 thousand $.
 
Too funny some of the comments. No point replying to them really. Not long left now, we will see what's what soon :D
 
yes some people will find it funny because they are in denial just like they were for Fury X and Polaris clocked speed.

If anyone bought the rubbish that the fury x was an "overclockers dream" then they were mad... it was pretty obvious that it wouldn't over clock due to them all being under water and 3rd party rebranded.. oh and that the only way to get one on air was to get the scaled down fury.
 
If anyone bought the rubbish that the fury x was an "overclockers dream" then they were mad... it was pretty obvious that it wouldn't over clock due to them all being under water and 3rd party rebranded.. oh and that the only way to get one on air was to get the scaled down fury.

True.
At least with the GTX1080 we could do a 36% OC when watercooled.
 
yes some people will find it funny because they are in denial just like they were for Fury X and Polaris clocked speed.

Polaris clockspeed is probably not really AMD's "fault" as such - broadly similar devices produced on Intel, TSMC and Samsung's equivalent(ish) processes manage 1500-1800MHz or a little higher in the case of nVidia GPUs.
 
Polaris clockspeed is probably not really AMD's "fault" as such - broadly similar devices produced on Intel, TSMC and Samsung's equivalent(ish) processes manage 1500-1800MHz or a little higher in the case of nVidia GPUs.
Nvidia achieved high clocks through a lot of critical path optimizations sacrificing some IPC in the process.
 
Polaris clockspeed is probably not really AMD's "fault" as such - broadly similar devices produced on Intel, TSMC and Samsung's equivalent(ish) processes manage 1500-1800MHz or a little higher in the case of nVidia GPUs.
Don't try and defend AMD, it was not GloFo's fault at all. AMD, they are trash and sooner they go bankrupt the better for us all!

Nvidia FTW, best company, never do anything wrong, always best products and they never lie to their customers, they have our best interests at heart!

/sarcasm :p
 
Status
Not open for further replies.
Back
Top Bottom