Indeed. AMD's only realistic approach (due to low R&D priority) was to make a compute card that can do double duty as a gaming card, hence Vega. I don't know if this is confirmed but there have been several stories about Vega developers being snatched for Navi instead (since that's where the real money is for AMD). The fruits of Zen and Navi in consoles will hopefully help get AMD back into the gaming market but it won't happen overnight.
-
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 7nm GPU News and Rumours 2018/2019
- Thread starter Beren
- Start date
So you are all adamant that AMD wouldn't use a new node (7nm) to gain small boost with an old chip (Vega 64).
Well in my defence I give exhibit A: the RX590 a old chip (RX480) made on a new node (12nm) to gain a small boost.
Also what else are they going to bring to market, they have nothing ready to replace the currant Vega 64 with, Navi is supposedly going to replace the Polaris line which is now on its third revamp all while a perfectly good faster 7nm Vega just sits there unused but for professional use.
Maybe I'm wrong only time will tell.
Well in my defence I give exhibit A: the RX590 a old chip (RX480) made on a new node (12nm) to gain a small boost.
Also what else are they going to bring to market, they have nothing ready to replace the currant Vega 64 with, Navi is supposedly going to replace the Polaris line which is now on its third revamp all while a perfectly good faster 7nm Vega just sits there unused but for professional use.
Maybe I'm wrong only time will tell.
They might but I doubt it'll turn up any time soon.So you are all adamant that AMD wouldn't use a new node (7nm) to gain small boost with an old chip (Vega 64).
Well in my defence I give exhibit A: the RX590 a old chip (RX480) made on a new node (12nm) to gain a small boost.
Also what else are they going to bring to market, they have nothing ready to replace the currant Vega 64 with, Navi is supposedly going to replace the Polaris line which is now on its third revamp all while a perfectly good faster 7nm Vega just sits there unused but for professional use.
Maybe I'm wrong only time will tell.
Soldato
Also what else are they going to bring to market, they have nothing ready to replace the currant Vega 64 with, Navi is supposedly going to replace the Polaris line which is now on its third revamp all while a perfectly good faster 7nm Vega just sits there unused but for professional use.
Maybe I'm wrong only time will tell.
Well that's a tricky one and open to interpretation. Nobody ever said Navi is replacing Polaris, what has been said is Navi is targeting the mid-range. But what that actually means is open to interpretation. Also bear in mind that Navi is supposedly aimed at GTX 1080 performance, which is where RX Vega 64 sits already.
Common interpretation: Polaris is mid-range, Navi targets mid-range, ergo Navi replaces Polaris. Fine. But if Navi is now GTX 1080 performance, and Vega 10 is already GTX 1080 performance, what happens then? This is where a 7nm gaming Vega could happen, but even if Vega 20 can be cut down, can work with 2 stacks of HBM and offers 25% more performance, Christ knows how much it will cost. I'd say double the price of Navi for only an extra 25% performance? And that's still not GTX 1080 Ti performance? Doesn't make any sense to me.
My interpretation (as posted before): RX Vega performance is mid-range, Navi performance targets mid-range, ergo Navi replaces RX Vega. That then shifts Polaris 30 to the entry-level bracket because, as I debated with FoxEye, AMD will still need a product at the sub £250 price point. This is why I'm surprised Polaris 30 is only a new RX 500, rather than the basis of a new RX 600 line that's 15% faster across the range.
You never know, we may end up with both of these things happening. Hell, if it is viable to bring Vega 20 to gamers to cover the 1080 Ti performance bracket at a viable price point then bring it on! I doubt anybody is going to argue.
Soldato
So you are all adamant that AMD wouldn't use a new node (7nm) to gain small boost with an old chip (Vega 64).
Well in my defence I give exhibit A: the RX590 a old chip (RX480) made on a new node (12nm) to gain a small boost.
Also what else are they going to bring to market, they have nothing ready to replace the currant Vega 64 with, Navi is supposedly going to replace the Polaris line which is now on its third revamp all while a perfectly good faster 7nm Vega just sits there unused but for professional use.
Maybe I'm wrong only time will tell.
We do not know anything about Navi tbh. I wouldn't be surprised AMD pulling a fast one and delivering something great.
Vega is designed with compute in mind, so I don't see it delivering £1200 GPU to the consumers.
Soldato
Its all in the name?
Navi.. The cheaper part?
Vega.. The more expensive part?
Arcturus..
Not so sure about this one, although some seem to think this is a double star.
Dual chip GPU perhaps?
Navi.. The cheaper part?
Vega.. The more expensive part?
Arcturus..
Not so sure about this one, although some seem to think this is a double star.
Dual chip GPU perhaps?
Soldato
Soldato
So you are all adamant that AMD wouldn't use a new node (7nm) to gain small boost with an old chip (Vega 64).
Well in my defence I give exhibit A: the RX590 a old chip (RX480) made on a new node (12nm) to gain a small boost.
Also what else are they going to bring to market, they have nothing ready to replace the currant Vega 64 with, Navi is supposedly going to replace the Polaris line which is now on its third revamp all while a perfectly good faster 7nm Vega just sits there unused but for professional use.
Maybe I'm wrong only time will tell.
It's not a "new" node though, it's tweaked transistor geometry and marketing fluff. I recall seeing a quote from an engineer stating GF 14 to 12nm was basically copy paste... They have a quota to maintain with GF, the RX590 was probably the least horrible way to fulfil this
It's not a "new" node though, it's tweaked transistor geometry and marketing fluff. I recall seeing a quote from an engineer stating GF 14 to 12nm was basically copy paste... They have a quota to maintain with GF, the RX590 was probably the least horrible way to fulfil this
This is especially true because the dimensions of Polaris 30 are identical to 20. There is been zero change to layout to take advantage of the improved density.
Its not even clear if the new 12nm process has any real power advantages, as the 590 now sucks 225w. The 590 is really just an overclocked 580.
Soldato
Soldato
I have officially lost my mind
Having a sound mind is overrated
Soldato
Yep the 12nm isn't new, but saying that it is new to AMD as the 590 is the first chip they have made on it, but the 7nm is undoubtedly new, new to AMD, new to the world GPU wise, hence why AMD can say quite correctly the MI60 is the worlds first 7nm chip.
Soldato
and they would blow there power budget even more if a 512bit GDDR5/x controllers and chips where used.
Well, AMD weren't obliged to do it, it is just the next evolutionary step: https://www.amd.com/en/technologies/hbm
Revolutionary HBM breaks the processing bottleneck
HBM is a new type of CPU/GPU memory (“RAM”) that vertically stacks memory chips, like floors in a skyscraper. In doing so, it shortens your information commute. Those towers connect to the CPU or GPU through an ultra-fast interconnect called the “interposer.” Several stacks of HBM are plugged into the interposer alongside a CPU or GPU, and that assembled module connects to a circuit board.
Though these HBM stacks are not physically integrated with the CPU or GPU, they are so closely and quickly connected via the interposer that HBM’s characteristics are nearly indistinguishable from on-chip integrated RAM.
Power Efficiency
GDDR5 has served the industry well these past seven years, and many gigabytes of this memory technology are used on virtually every high-performance graphics card to date.
But as graphics chips grow faster, their appetite for fast delivery of information (“bandwidth”) continues to increase. GDDR5’s ability to satisfy those bandwidth demands is beginning to wane as the technology reaches the limits of its specification. Each additional gigabyte per second of bandwidth is beginning to consume too much power to be a wise, efficient, or cost-effective decision for designers or consumers. Taken to its logical conclusion, GDDR5 could easily begin to stall the continued performance growth of graphics chips. HBM resets the clock on memory power efficiency, offering >3X the bandwidth per watt of GDDR5.1
Smaller Form Factors
Beyond performance and power efficiency, HBM is also revolutionary in its ability to save space on a product. As gamers increasingly expect smaller and more powerful PCs, the elimination of bulky GDDR5 chips in favor of HBM can enable devices with exciting new form factors that pack a punch in a smaller size. Compared to GDDR5, HBM can fit the same amount of memory in 94% less space!2
With HBM, AMD is set to once again revolutionize the industry, from next-level gaming to VR and beyond.
Welcome to 3 years ago.
I think it is fairly safe to say that unfortunately HBM hasn't revolutionised the industry. That is not to say there is anything wrong with it particularly, but GDDR with 5x and now 6 hasn't been left behind as expected.
Welcome to 3 years ago.
I think it is fairly safe to say that unfortunately HBM hasn't revolutionised the industry. That is not to say there is anything wrong with it particularly, but GDDR with 5x and now 6 hasn't been left behind as expected.
Well, I believe it will in the next couple of years.
After all, with AMD, we are still stuck at Polaris refreshes...
Very true there is still plenty of time for it to come good yet.