Soldato
I've had enough of waiting for Vega. I'm waiting for Navi now.
You just made me laugh so hard with that comment
I was kinda thinking the same thing a couple of weeks ago tbh
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.
I've had enough of waiting for Vega. I'm waiting for Navi now.
It can be used for some AI work, namely deep learning due to the FP16 double-rate performance but the lack of FP64 is still an issue for general machine learning algorithms. For deep learning Vega has a lot of potential but the critical aspect is software support. Most of the DL tools out there don't even support OpenCL, they all support CUDA or at the very least CUDA is the main priority and gets the first features. The CUDA integration also achieves massive efficiency increases t that could largely mitigate the FP16 advnatge of Vega when compared to lower end Pascal units. The big players in DL are all happily using GP100 GPUS, or simply using Amazon AWS GPU instances. So there isn't a big market there. I do a little Deep learnign at work 9although msotly use alternatively ML archietcures), if we did a little more then we might very well buy a Vega GPU for development but using Amazon AWS works out fine for us.Vega FE can be used for AI work, it also has plenty of pro use cases where certification is not required. The certification is only really needed in big companies where the work being done can be considered mission critical etc.
more details, great. No actual release at Siggraph then, launch might be in another month and then availability a month later. Makes ense if Vega really did reqyuire a respin to fix some issues.tweet by ayymd ceo: Getting ready to ship @AMDRyzen #Threadripper... super excited! More details on @Radeon Vega and Threadripper at #SIGGRAPH2017.
can she be any more clear about the status of vega's paper launch?
It has to beat a GTX1080FE by a reasonable amount though,or at least be around aftermarket GTX1080 level,otherwise I am not sure unless they pitch this under the price of a GTX1080,whether it will gain any traction as a card.
Edit!!
If you look at the Fury X,the TPU launch review shows the reference GTX980TI was upto 10% faster at qHD and 4K and that was not helped by the 4GB of VRAM for certain games I suspect.
If this can barely match a GTX1080,it will be relatively worse especially as it is on a new node,as the GTX1080TI FE is between 20% to 30% faster than a GTX1080 FE:
https://www.techpowerup.com/reviews/Zotac/GeForce_GTX_1080_Ti_Amp_Extreme/31.html
Well it does look like a bargain compared to Vega FE Water
It has to beat a GTX1080FE by a reasonable amount though,or at least be around aftermarket GTX1080 level,otherwise I am not sure unless they pitch this under the price of a GTX1080,whether it will gain any traction as a card.
Edit!!
If you look at the Fury X,the TPU launch review shows the reference GTX980TI was upto 10% faster at qHD and 4K and that was not helped by the 4GB of VRAM for certain games I suspect.
If this can barely match a GTX1080,it will be relatively worse especially as it is on a new node,as the GTX1080TI FE is between 20% to 30% faster than a GTX1080 FE:
https://www.techpowerup.com/reviews/Zotac/GeForce_GTX_1080_Ti_Amp_Extreme/31.html
I've had enough of waiting for Vega. I'm waiting for Navi now.
The way this "learn more about vega" stuff is dragging ... i may end up with a bloody phd by september
somehow those obnoxious prices quoted by nordichw are looking acceptable if it can cut the wait..
i hope the swedish have good healthcare programs for 25% vat
It has to beat a GTX1080FE by a reasonable amount though,or at least be around aftermarket GTX1080 level,otherwise I am not sure unless they pitch this under the price of a GTX1080,whether it will gain any traction as a card.
Edit!!
If you look at the Fury X,the TPU launch review shows the reference GTX980TI was upto 10% faster at qHD and 4K and that was not helped by the 4GB of VRAM for certain games I suspect.
If this can barely match a GTX1080,it will be relatively worse especially as it is on a new node,as the GTX1080TI FE is between 20% to 30% faster than a GTX1080 FE:
https://www.techpowerup.com/reviews/Zotac/GeForce_GTX_1080_Ti_Amp_Extreme/31.html
Agree but i can still not get my head round them shrinking to 14nm and it using + 300w and only be a tad slower than the 1080
seems to be a power sucking chip ..there has to be more ...not long to go to find out
Polaris is on 14nm yet a 580 uses more power than a 1080, whilst the 1080 outperforms it by 60-70% in games. AMD just can't seem to get GCN near Pascal levels of efficiency.
I know its not Vega ...but lets all have a break from it ...
I know its not Vega ...but lets all have a break from it ...
That guy sounds more like someone doing an impersonation of someone Scottish, the way he puts emphasis on words with t or d does my nut.
Not your potato?
Yeah, it's shocking if true. No one got close to that price in the Shankly SweepstakeThat it does
Pricing's gonna kill it for me. :C
me too
Agree but i can still not get my head round them shrinking to 14nm and it using + 300w and only be a tad slower than the 1080
seems to be a power sucking chip ..there has to be more ...not long to go to find out