Yep agreed, and hopefully Vega will be able to highlight this. I just think there's proven titles where it does make a difference so I'd rather play it safe if there's going to be an 8Gb version of Vega. All depends on costs as well though Not an unlimited budget.As stated before, VRAM usage doesn't mean a whole bunch. Some games take up a lot more when its available with no actual performance difference.
Could be, but I still think it has more to do with GFs process not being up to snuff and AMD holding out for Samsung LPC/LPU and making it there.
They wouldn't have changed the wafer supply agreement to allow for such if they weren't going to use it and GF is going to be at capacity with AMDs other products all using GF so they'd need to move something to a third party. It makes the most sense to do it for the highest margin parts that need the fewest wafers from a cost perspective.
If you're expecting Vega at Computex, you will likely be disappointed. Nothing in the agenda really points to Vega.
100% of the "clues" people are grasping at can 100% be explained by RX 5xx series products.
Computex is one of the biggest tech tradeshows in the world. It would be astonishing if there were no Vega.
yes please
Prepare to be astonished, then. People were saying the same thing around the time the Radeon RX 480 launched, that for sure it would be a hard launch at Computex:
http://www.pcworld.com/article/3073...laris-gpu-updates-and-bristol-ridge-apus.html
That turned out to be a bust.
Launch dates of the past three GPUs from AMD:
"Polaris" AKA Radeon 400 series: June 29, 2016
"Fiji" AKA Radeon Fury series: June 24, 2015
"Grenada" AKA Radeon 3xx series: June 18, 2015
You may hear additional details about Vega via AMD livestream at Computex. You may even see a few carefully curated, canned demos and be promised that it is arriving soon™. I would not expect a hard launch of Vega before mid June because that's just setting yourself up for disappointment.
I am hopeful for Vega. I want Vega to be great so that the industry we ALL love can get back to being healthy. Nvidia is milking the market hard due to lack of competition and AMD has been absent from the high end for a year now.
Hopefully Raja and his team have created something truly great with Vega. Just to put some historical perspective on what we may expect..
Apple hired away Raja Koduri in 2009. The last architecture to release that Raja most likely saw to completion was the launch of the Radeon 5870 http://www.anandtech.com/show/2841. In comparison to Fermi, this was the card to get for that generation.
Raja rejoined AMD from Apple in 2013 http://www.anandtech.com/show/6907/the-king-is-back-raja-koduri-leaves-apple-returns-to-amd
It will be just a bit over 4 years since Raja rejoined AMD that Vega will launch to market. Is 4 years enough to see a GPU go from drawing board to final product? Few here or anywhere doubt Raja's engineering brilliance.. Has he created Vega from scratch? (or distantly derived and heavily modified GCN?)
Source?Navi is Raja's first GPU from scratch, but he has heavily influenced Vega.
Source?
Source?
It's been said quite a few places, but this was the first result on google so;
http://www.theinquirer.net/inquirer/news/3009635/amd-to-show-off-vega-and-navi-gpu-tech-on-16-may
Prepare to be astonished, then. People were saying the same thing around the time the Radeon RX 480 launched, that for sure it would be a hard launch at Computex:
http://www.pcworld.com/article/3073...laris-gpu-updates-and-bristol-ridge-apus.html
That turned out to be a bust.
Launch dates of the past three GPUs from AMD:
"Polaris" AKA Radeon 400 series: June 29, 2016
"Fiji" AKA Radeon Fury series: June 24, 2015
"Grenada" AKA Radeon 3xx series: June 18, 2015
You may hear additional details about Vega via AMD livestream at Computex. You may even see a few carefully curated, canned demos and be promised that it is arriving soon™. I would not expect a hard launch of Vega before mid June because that's just setting yourself up for disappointment.
It seems that Vega will be the first GPU to have Infinity Fabric for an interconnect. We'll see if they are able to take the various bottlenecks and inefficiencies present in a chip the size of Fiji and resolve them with IF and a focused set of enhancements in trouble areas.Not sure if it there is one, sebbi made a post about how Vega is really GCN 5, but that seems to contradict what Lisa Su said about Vega being a ground up design.
Exactly. First I have heard of Navi being the first instead of Vega was WCCFTech and its associated echo chambers. Zero remarks from AMD or anyone close to AMD about that. The only indication we have is that Raja has been there over 4 years which I believe should be plenty of time for him to have led the design of Vega and some comments from Lisa Su stating that Vega is a ground up design.Not sure if it there is one, sebbi made a post about how Vega is really GCN 5, but that seems to contradict what Lisa Su said about Vega being a ground up design.
They cite WCCFTech as a source... very large grains of salt required on accuracy.