• 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 VEGA confirmed for 2017 H1

Status
Not open for further replies.
Again you are inventing your own specifics into things I've posted that I don't even mention or allude to - seriously re-read your own posts and then try and find anything like for instance where I've made any mention of software flags or whatever directly or indirectly. My posts are both a lot more general and a lot less anti-AMD than you are perceiving them to be.

I've no interest in seeing Vega fail for instance as if nothing else nVidia need a kick up the rear to stop them trickle feeding consumers I'm merely being somewhat realistic.

EDIT: PS I've been very quiet in the Volta thread for a reason - it'll become more clear but remember what I said about people not understanding what Volta is all about :S

You are the one inventing your own specifics by making snarky comments like this:

Just because they might be saying what you want to hear doesn't mean they are right :P

Which is YOU making up stuff about what I am saying - you started it with pointless comments like,and now you are backtracking when flagged on it.


So by extension I will call you out now:

Just because they might be saying what you don't want to hear doesn't mean they are wrong:P

If you argue with that statement you are inventing your own specifics.

However much you want to bury,there are greater changes to Vega than with Polaris - Beyond3d forums is discussing where those changes might be.
 
Last edited:
Plus how many times have I actually been more realistic on how parts perform than many here??

Look in the Polaris thread - Polaris 10 will be beat a GTX980TI and won't launch in June.

Yet I said R9 390 to Fury level(depended on what clockspeeds the part would ship at) performance launch late June or early July and the whole point of Polaris 10 was to cut costs. I even had people having a go at me for suggesting it in that thread.

Then when we had posts from China confirming it had 2304 shaders after I said that,and another once saying 5% to R9 390X before launch,still the same overblown expectations and some of it was done on purpose IMHO.

Then at launch we had that:

https://tpucdn.com/reviews/AMD/RX_480/images/perfrel_1920_1080.png

Don't ever try to say I am being optimistic about purported AMD performance figures,or raise what expectations people might have.

Even my performance expectations for the next Vega are much lower than for many here.

The same goes for the RX490 which people think is a bigger Polaris chip and I expect it will be a dual Polaris 10 card,and I have said many times multi-GPU is fail currently.

In fact I was more disappointed by the RX460. If it had been a full 16CU part it would have been close to what I was predicting(nearer to a GTX960) yet the cut to 14CUs made it utterly meh.
 
Last edited:
Lets look at what we know about Vega so far:
1.)Has different software flags to Tonga/Fiji/Polaris(which are apparently the same) meaning it has greater changes
2.)Engineer leaked that a new 4096 core is being developed
3.)Both Polaris 10 and 11 hit higher clockspeeds than the previous cards - about 1.2GHZ to 1.4GHZ or thereabouts
4.)Pictures of the XBox Scorpio indicates the use of GDDR5 or GDDR5X not HBM or HBM2
5.)Rumours the XBox Scorpio has 12GB of VRAM. This indicates the GPU either has a 384 bit memory controller with GDDR5 or a GDDR5X implementation
6.)AMD Polaris 10 and 11 apparently have optimal performance/watt at 900MHZ or around that,so that hints at what we will see the XBox Scorpio GPU at
7.)The PS4 PRO uses a 2304 shader GPU at 911MHZ and is the same configuration as the Polaris 10 core,hence there is strong evidence to indicate we might see one of the Vega cores being based on the one in the XBox Scorpio
8.)Move to GDDR5 or GDDR5X will mean probably less bandwidth tha HBM2 and AMD bandwidth compression is not as good as Nvidia,so we might find some performance drops there.
9.)I will be surprised if the smaller Vega uses HBM2 since it makes me wonder whether AMD can fight Nvidia on price that easily and it wouldn't surprise me if the delay is down to getting clockspeeds up,or even waiting for GDDR5X prices to drop

So based on that,we can get some rough guess where the next core AMD launches will land. Now OFC,that assumes that 4096 core is still being developed.

AMD says Vega will have a slight performance/watt advantage over Polaris from their own slides.

This is why I think the smallest of the next GPU launches will be more GTX1080 level performance. A Fiji core with uarch improvements,better tessellation,etc with a 30% clockspeed bump. If GF 14NM improves by then,maybe it will be close to 1.4GHZ but I doubt any higher.

Such a core should be between 300MM2 to 400MM2,so AMD can attempt to be price competitive in some way. I also expect the larger Vega will be a bigger core than the GP102 as AMD does not have separate gaming and professional lines,so have to make a jack of all trades core just like with Hawaii.

So I will put that as a disclaimer as I don't want to be accused of pumping expectations for AMD,and I will also put the disclaimer I could be wrong just for future reference.
 
Last edited:
the part about disappointing performance per watt couldn't be further from what everyone else is seeing. If anything, AMD have hugely improved on their performance per watt ratios with Polaris, leaving negligible difference between each camp.

This is actually very wrong. Polaris does have disappointing perf/w, and there's a massive gap between Nvidia and AMD right now.

Source, newest techpowerup review: https://www.techpowerup.com/reviews/MSI/GTX_1070_Quick_Silver_OC/30.html

I like AMD and REALLY REALLY hope Vega is good, but the only thing Polaris is good for is its DX12 performance and it's performance per £. In every other technical metric it was disappointing.
 
This is actually very wrong. Polaris does have disappointing perf/w, and there's a massive gap between Nvidia and AMD right now.

Source, newest techpowerup review: https://www.techpowerup.com/reviews/MSI/GTX_1070_Quick_Silver_OC/30.html

I like AMD and REALLY REALLY hope Vega is good, but the only thing Polaris is good for is its DX12 performance and it's performance per £. In every other technical metric it was disappointing.

AMD managed to go from a R9 390 being 60% worse than a GTX970 to a RX480 against a GTX1060 being 34% worse off! :p

They are getting better!! :p

Plus AMD was using a 440MM2 salvaged part with a 512 bit memory controller and double the RAM chips when compared to a 398MM2 salvaged part with a 224 bit memory controller. AMD also had to use higher density chips too.

Now they are using a 232MM2 part against a 200MM2 part,with a 256 bit memory controller against a 192 bit memory controller,whilst dropping TDP and power consumption significantly.

Then you have the RX470 4GB and RX480 4GB cards,also again using a 232MM2 GPU against the 365MM2 Tonga GPU and generally dropping power consumption against the R9 380 and R9 380X(which should drop costs too).

Not ideal at all still,but still far better than the situation of the R9 390 8GB and GTX970 4GB in terms of manufacturing costs.

AMD is moving in the right direction,but still has a lot of work to do IMHO OFC.
 
Last edited:
The 1080TI will be a $900 GPU, only about 1 in 5,000 give a #### about it.

Its not relevant.

It entirely depends on AMD's lineup, so here's to hoping AMD will actually deliver something great! :cool: Definitely leaning toward AMD's Zen CPUs if they offer haswell-type performance with more cores at a much cheaper price (8 cores or more with 64Gb RAM sounds like a good combo :D).
 
all I would like is 1080ti like performance at a lower cost, in a similar style to the FuryX

Ta

The GTX 980 Ti was the better buy by far, and the Fury/Fury X were seriously overpriced for what they offered at the time; 4Gb ram, no overclocking headroom, and much higher TDP, and let's not forget the pump-issues that many people had to endure. If history is going to repeat itself, then let's hope AMD has learned from their mistakes. :(
 
I still think AMD will be releasing a GTX1080 competitor first - unless they can beat a GTX1080TI/Titan X by a reasonable amount and be a bit cheaper,people will still buy a GTX1080TI. The problem is if they are using a 500MM2+ GPU(which is quite possible) and loads of expensive HBM2,then it puts them in a tight space IMHO OFC(could be wrong).
 
Although the half-missing segment does make one wonder, the part about disappointing performance per watt couldn't be further from what everyone else is seeing. If anything, AMD have hugely improved on their performance per watt ratios with Polaris, leaving negligible difference between each camp.

That in it self is a bit of an achievement given the high power requirements for the previously released GCN lineup and even going further back.

Absolute rubbish! A 1070 is much more efficient than a 480. AMD improved their efficiency but no where near the claims / rumours.
 
Absolute rubbish! A 1070 is much more efficient than a 480. AMD improved their efficiency but no where near the claims / rumours.

+1, AMD now have a 980 equivalent with the rx480, nothing more, nothing less.

It's still a massive improvement for them, which I'm sure will be further improved upon with Vega.
 
I still think AMD will be releasing a GTX1080 competitor first - unless they can beat a GTX1080TI/Titan X by a reasonable amount and be a bit cheaper,people will still buy a GTX1080TI. The problem is if they are using a 500MM2+ GPU(which is quite possible) and loads of expensive HBM2,then it puts them in a tight space IMHO OFC(could be wrong).

The problem is though that the GTX 1080 doesn't offer next-gen performance; it's basically a slight improvement over what we've had for a long time already - we need something close to the Titan X, but cheaper (so people can put them in CF/SLI), in order for 4K gaming to become a reality. The Titan X is the only decent card this gen imho, and we know nvidia haven't even unleashed the full potential of Pascal yet due to AMD lagging so far behind.
 
The problem is though that the GTX 1080 doesn't offer next-gen performance; it's basically a slight improvement over what we've had for a long time already - we need something close to the Titan X, but cheaper (so people can put them in CF/SLI), in order for 4K gaming to become a reality. The Titan X is the only decent card this gen imho, and we know nvidia haven't even unleashed the full potential of Pascal yet due to AMD lagging so far behind.

A 1080 is more than a slight improvement over a 980, or 980ti, or Fury X IMO.
 
The problem is though that the GTX 1080 doesn't offer next-gen performance; it's basically a slight improvement over what we've had for a long time already

Ok so in your mind what about the AMD RX480 is that a reasonable step over the outgoing AMD380 ?


Just a hint both the 480 (replaces the 380) and 1080 (replaces the 980) are about 35-40% faster than the card they are replacing. ;)
 
Status
Not open for further replies.
Back
Top Bottom