Vega Rumors

The 1600MHz came about from people calculating the required clocks to hit 12.5 TFLOPs based upon what was known at the time and the 4096 cores, compared to Fiji.

However this was applied to MI25, and the MI25 is not watercooled but passive, meaning cooling from the server-node-blade chassis structure.
It is a lot to ask especially when Nvidia is more conservative with their Boost clocks and the limits they are at with HPC Pascal on the larger dies (even GV102 that is smaller than Vega).

If AMD manage 1600MHz and raw performance actually scales to the theoretical peak limit and importantly sustained, then that is one great engineering feat from AMD IMO.
Personally I am expecting performance around 1080 level.
CHeers
 
DuronBurgerMan said:
Sounds about right. Although I suspect we'll see a 1600 MHz part. Been hearing that number often enough to suspect there's some truth to it.

So maybe something like this for a lineup?

Vega light - 1200 MHz-ish. 1070 competitor
Vega regular - 1500 MHz-ish. 1080 competitor
Vega heavy - 1600 MHz-ish. Not really a 1080 Ti competitor... but a way for AMD to offer a bit extra for less money than a Ti. Cherry-picked silicon.

Which goes back to Kyle's guess some time ago that the top Vega will probably be between a 1080 and a 1080 Ti.

Don't think we're seeing 16GB on a consumer card, though. Maybe for the professional cards (or a quasi-professional card offered for the 'extreme' segment, like Titan Xp). I'd guess 8GB is our top end for consumer parts, at least for now.
Click to expand...


Yeah I'm kinda in the same boat you are, between 1080 and 1080ti performance sounds about right closer to the 1080 though from what AMD has shown so far.
 
Magic Hate Ball said:
This is awkward to say, but is the moderator trying to derail a Vega thread with VR?
Click to expand...

Maybe, though I do feel a lot of us here are interested to see how Vega perform in VR since AMD have made improvements in VR from RX 400 series to RX 500 series. So it will be interesting to see how Vega fair with GTX 1080 and above.
 
Magic Hate Ball said:
This is awkward to say, but is the moderator trying to derail a Vega thread with VR?
Click to expand...

Someone earlier asked how could 16 GB of VRAM be usable? I was just pointing out that newer VR headsets are entering the 4K realm. I think that 4K resolution per eye would require a lot of VRAM. I don't think that is derailing the topic.

Do you?
 
DuronBurgerMan said:
Sounds about right. Although I suspect we'll see a 1600 MHz part. Been hearing that number often enough to suspect there's some truth to it.

So maybe something like this for a lineup?

Vega light - 1200 MHz-ish. 1070 competitor
Vega regular - 1500 MHz-ish. 1080 competitor
Vega heavy - 1600 MHz-ish. Not really a 1080 Ti competitor... but a way for AMD to offer a bit extra for less money than a Ti. Cherry-picked silicon.

Which goes back to Kyle's guess some time ago that the top Vega will probably be between a 1080 and a 1080 Ti.

Don't think we're seeing 16GB on a consumer card, though. Maybe for the professional cards (or a quasi-professional card offered for the 'extreme' segment, like Titan Xp). I'd guess 8GB is our top end for consumer parts, at least for now.
Click to expand...

Agree, going back to rumors, MI25 is suppose to sport 16GB HBM2. Still not sure if the 16GB HMB2 is true but very easy to imagine 16GB in a professional card though.
 
cageymaru said:
Someone earlier asked how could 16 GB of VRAM be usable? I was just pointing out that newer VR headsets are entering the 4K realm. I think that 4K resolution per eye would require a lot of VRAM. I don't think that is derailing the topic.

Do you?
Click to expand...

Gotcha, it was just odd with the links to videos of headsets and such without any mention of why it was being mentioned.

Yeah, I can see what you mean with the double 4k projection regarding memory needs.
 
Well it looks like nothing was released today after all.
 
cageymaru said:
Someone earlier asked how could 16 GB of VRAM be usable? I was just pointing out that newer VR headsets are entering the 4K realm. I think that 4K resolution per eye would require a lot of VRAM. I don't think that is derailing the topic.
Click to expand...

No. That's not how VRAM works. You don't need a separate framebuffer per eye. And proper VR is not yet entering 4K. Chinese knockoff headsets trying to capitalize on VR hype with glorified personal theater devices is not proper VR.

When HTC announces a Vive 2, or Oculus announces a Rift 2, or another SteamVR partner or maybe even Samsung announces a 4K HMD, then we'll be entering 4K VR.
 
Last edited:
Darakian said:
It shouldn't be. This is what the seamicro purchase was all about.
Click to expand...

More like what was left after SeaMicro collapsed and still had any value. The dream and hopes by buying SeaMicro was quite different.
 
financial presentation, Vega Cores in their Zen CPU's, 40% boost over what they had in BD or last gen and total Ryzen mobile CPU +GPU 50% reduction in power, sounds mostly from the CPU side.
 
razor1 said:
financial presentation, Vega Cores in their Zen CPU's, 40% boost over what they had in BD or last gen and total Ryzen mobile CPU +GPU 50% reduction in power, sounds mostly from the CPU side.
Click to expand...

Still be a decent upgrade especially in the mobile field, probably something I will look forward to.
 
Big announce was threadripper

DUD announcement on VEGA

Edit: Maybe not...Raja Up
 
Interesting stuff
 

Attachments

  • vega frontier.JPG
    vega frontier.JPG
    34 KB · Views: 48
  • vega vs p100.JPG
    vega vs p100.JPG
    23.1 KB · Views: 53
  • vega vs p100-2.JPG
    vega vs p100-2.JPG
    35.9 KB · Views: 48
Well I thought it be 16 GB GDDR variant instead of HBM, I have no problem saying I am wrong in my assumption.
 
Pieter3dnow said:
Have you seen the price already , I would doubt that it is "avg" consumer friendly
Click to expand...
He is just getting at the self-proclaimed-geniuses that obvious aren't at all that proclaimed there is no way there will be 16gb HBM2 on an AMD card.
 
You must log in or register to reply here.
Back
Top