• 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: ** The AMD VEGA Thread **

On or off the hype train?

  • (off) Train has derailed

    Votes: 207 39.2%
  • (on) Overcrowding, standing room only

    Votes: 100 18.9%
  • (never ever got on) Chinese escalator

    Votes: 221 41.9%

  • Total voters
    528
Status
Not open for further replies.
Why do you expect the RX Vega drivers to be significiantly better. They have 1 month to optimize, when they have been developing the drivers for 9-12 months minimum, likely more.

Well we don't know if the architecture is a complete overhaul because the dudes at AMD won't say anything about the GPU... So if NCU is to GCN what Zen is to bulldozer, it could make a difference. It is not logic that this gpu underperforms compared to Fiji AND Polaris, at this stage even Polaris has a better or on par IPC.
 
I think the only saving grace is looking back to RX vega demo in january that was faster by a big margin in doom at 4k going by recent results. Drivers should have came on since then. Really don't understand why this card ain't performing but there is still hope that when the gaming card arrives it will be way better especially if that rumour about a flaw in these chips is true. Very unusual to say the least.
 
Well we don't know if the architecture is a complete overhaul because the dudes at AMD won't say anything about the GPU... So if NCU is to GCN what Zen is to bulldozer, it could make a difference.
Its not a complete overhaul, AMD have said as much, vega is an evolution of GCN, just a much bigger step than the previous updates.
And even if it was a whole new architecture, that still doesn't explain why 1 extra month of driver development against what is most like 18-24 months of prior driver development is going to magically change things.
 
I think the only saving grace is looking back to RX vega demo in january that was faster by a big margin in doom at 4k going by recent results. Drivers should have came on since then. Really don't understand why this card ain't performing but there is still hope that when the gaming card arrives it will be way better especially if that rumour about a flaw in these chips is true. Very unusual to say the least.

Yes its strange ....I really want AMD to do well for all i sakes ..our wallets ...

But something is not right with the numbers i am hearing maybe there is some truth in a design flaw and the gaming RX vega will be a revised chip....Who knows ...But at this moment its not looking good
 
Isn't it down to his experience?

What happened to the "wait for the pro reviews"?

Last i checked he has the Vega FE, he hasn't got the RX Vega. Pulling percentages out of the air for a card he doesn't have is pointless. As i said its best to wait for the reviews to hit.
 
I think the only saving grace is looking back to RX vega demo in january that was faster by a big margin in doom at 4k going by recent results. Drivers should have came on since then. Really don't understand why this card ain't performing but there is still hope that when the gaming card arrives it will be way better especially if that rumour about a flaw in these chips is true. Very unusual to say the least.

Well in DIRT Rally its more like a GTX1080(he said around 100FPS which is around what they saw with the GTX1080),and in Fallout 4 they saw this:

https://youtu.be/bhGAS_oGN3c

This is what they see for the Titan Xp:

https://www.pcper.com/image/view/79272?return=node/67277

The sharp dip for the Titan Xp is just above 100FPS,so it looks like its averaging between 100FPS to 150FPS,which is slightly better than a GTX1080.

This is significantly faster than a Fury X in Fallout 4.

Edit!!

eKFzDu4.jpg


UTfHk7N.png
 
Its not a complete overhaul, AMD have said as much, vega is an evolution of GCN, just a much bigger step than the previous updates.
And even if it was a whole new architecture, that still doesn't explain why 1 extra month of driver development against what is most like 18-24 months of prior driver development is going to magically change things.

Yep seems to be a refinement, damn that's not looking good...
 
Well in DIRT Rally its more like a GTX1080(he said around 100FPS which is around what they saw with the GTX1080),and in Fallout 4 they saw this:

https://youtu.be/bhGAS_oGN3c

This is what they see for the Titan Xp:

https://www.pcper.com/image/view/79272?return=node/67277

The sharp dip for the Titan Xp is just above 100FPS,so it looks like its averaging between 100FPS to 150FPS,which is slightly better than a GTX1080.

This is significantly faster than a Fury X in Fallout 4.

This just further shows that there is a lot more work to be done on the driver front.... You can't have a GPU matching a 1080 one minute and then have it barely matching a 1070 the next minute.....

But nope, lets all just act like the drivers are as good as they are going to get :D :p :o
 
A lot of optimisation may still need to be done tbf, esp due to the cutting edge features of the architecture.

I added a screen capture of the FO4 results and one of their reviews. So it looks like its probably GTX1080 ballpark performance again I suspect.

This just further shows that there is a lot more work to be done on the driver front.... You can't have a GPU matching a 1080 one minute and then have it barely matching a 1070 the next minute.....

But nope, lets all just act like the drivers are as good as they are going to get :D :p

Well if he says another 10% extra performance for the gaming drivers in his estimates,and if the RX Vega gaming card is water cooled,which means maybe another 10% extra clockspeed,that should probably put it between a GTX1080 and a GTX1080TI.
 
This just further shows that there is a lot more work to be done on the driver front.... You can't have a GPU matching a 1080 one minute and then have it barely matching a 1070 the next minute.....

If its architecturally different to Pascal (which it is) it could vary in performance against it quite a bit depending on the type of load it was dealing with if it has specifically different strengths and weaknesses at dealing with different types of processing i.e. just as an example it could have ridiculously fast shader performance but not be great for texture fill rate, etc.

A lot of optimisation may still need to be done tbf, esp due to the cutting edge features of the architecture.

Problem is it has increasingly become apparent that the headline tech features they've talked about, which on paper would increase performance a lot, either only support a limited number of titles or need developers to either produce their game in a particular way or program support for specifically.
 
Status
Not open for further replies.
Back
Top Bottom