They can't because they need four stacks for that increase in bandwidth which enables the extra performance.
-
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.
Poll: Will you be buying a Radeon VII?
- Thread starter Gregster
- Start date
Soldato
Hard to decide without any real info. But likely no, I have a GTX 1080Ti and a G-Sync monitor.
Permabanned
16GB is extremely compelling for virtual machine host
Price wise its not looking too appealing but I will wait to see how the performance looks. After seeing Nvidia's ridiculous pricing for the 20XX series, I'm pretty much done with them and prefer to support their competitors to make the market more competitive, so maybe this could be a good alternative.
But why can't they use 4x2? Or doesn't it come in that size? Speaking of which, how do yields work with RAM? If you get a 4gb module with defects, can it be used as a 2gb module like they do with cut down CPUs?
Will be upgrading to a 1440p 32" Freesync 2 monitor very soon. Currently have a Vega 56 which i will run on the new monitor and see how that plays out. This Radeon VII has me interested so will wait for reviews and benchmarks before i commit to a purchase, if i like what i see of course.
Man of Honour
It is a nice looking card.
I don't know enough about it to say, it's just that with Vega 64 they used 2*4GB and yet Vega is bandwidth starved. There must be reason why they didn't use 4*2GB.
Soldato
It is a nice looking card.
Agreed, and that spec is epic.
1TB/s Mem Bandwidth !
16GB HBM !
I may have to double dip, split custody between Radeon VII and RTX 2080 xD
Either of these cards look great value in the face of the RTX 2080 Ti, basically double the price !
They seem to want to use 7nm purely to help with their inefficiency problems.
They needed to do something because the last AMD GPU with 1:2 FP64 support was Hawaii... AMD's market share in GPU HPC was very low, and they had nothing to respond to Nvidia's tensor cores. Vega20 adds a few math ops and low precision support that help a little but Volta is literally 5-10x faster at Deep learning in image net.
There is also a dedicated Turing card for deep learning inference, uses only 75w but is several times faster than a Vega20
It looks like the time spent on the cooler design was about 4 minutes: "A grey box, oh and make sure there's holes for the fans".
I like it, it's minimalist!
Some of us don't want colours, ridges, wings etc stuck all over it. I really liked the Vega 64 LC edition because of this.
It's got a lot going it, more for a productivity card than for gaming, £700 isn't a lot for what you can do with it as it's not solely a gaming card. I won't be buying one as I just got a Vega64 and I'm not salty as I got the V64 for a good price and it'll do me for years to come.
I'll be giving it a skip. Will carry on waiting for Navi, and stick with our RX580 and 1060s for a bit longer as they're doing the job. I'm hoping to jump to 1440p/144Hz if there's a suitable Navi card. Was optimistically hoping something would have been announced yesterday, but I realise that was stretching it a bit.
I doubt it, though I will wait to see the reviews first. Very keen to try an all AMD system later this year but I'd prefer to buy something better weighted towards value for money. I may end up getting the CPU/mobo and ram first, since my GPU still cuts the mustard. Then, further down the line I could drop an AMD GPU & new monitor in.