Preview: AMD VEGA GPU Architecture
Click here to post a comment for Preview: AMD VEGA GPU Architecture on our message forum
GeniusPr0
Can't wait to get it
Goiur
Thats all? 😛uke2:
cowie
^hopefully we all will not be saying that when it does launch
Endymion
So .. WHEN?
negachampa
Really, still over 6 months away for the high-end VEGA? No wonder Nvidia didn't announce the 1080Ti at CES. They are just waiting for AMD to catch-up and as soon as they do the 1080Ti will be released.
Seems AMD has basically fell a generation behind. Not to soon after VEGA is out, Nvidia will be on to their next gen (end of 2017) and AMD will once again have no answer.
I was really looking forward to seeing something announced at CES for new GPUs.
malitze
As always great article Hilbert. Surely makes sense to go enterprise first if it's more profitable, even though I'd like to seem some consumer stuff sooner. There is a lot of catching up to do in the enterprise area.
Noisiv
mattm4
cryohellinc
This is quite sad really, by the time they release Nvidia will have 1080Ti coming right after it. And then several months later they will release 1180... at 1000$ per card. GG
Undying
Stormyandcold
OK Boss, doesn't this also mean that all Vega reveals so far have been using an enterprise-class gpu and not a consumer version as has been expected? I assume this because I get the impression that the consumer version is not finalised yet.
flashmozzg
Someone needs to proofreed these articles. Stumbled upon many weird "typos" like
Solfaur
Hilbert Hagedoorn
Administrator
SirDremor
Asgardi
negachampa
What is even more sad is by the time high-end VEGA is released the Titan X would have been out for over a year and AMD still probably will not be able to compete !
I can't see how they could ever catch-up at this point.
Zooke
Loophole35
FranciscoCL
No 1080ti and no VEGA launh dates.
So, if VEGA is really launched at the end of H1, maybe nVidia will never launch the 1080ti and save those chips for a future GTX1180 (crippled GP102, 8 GB GDDR5X@256 bit) and new Titan (full GP102 with 3840 CUDA cores and better clocks, 12 GB DDR5X@384 bits), something similar to Kepler days.