Overclock.net banner
21 - 40 of 49 Posts
Quote:
Originally Posted by PontiacGTX View Post

if its at Fury x level it wouldnt be working, in both they are above 580. and above Fury X
Not quite. Vega FE's triangle throughput is higher because it's running at higher clocks. Notice how the Vega FE clocked at 1050 MHz has the same triangle throughput as a Fury X?

What you need to look at is the difference between partially culled and fully culled tests with the Vega cards. There is no difference. The primitive discard accelerator is not functional.
 
Quote:
Originally Posted by AmericanLoco View Post

Not quite. Vega FE's triangle throughput is higher because it's running at higher clocks. Notice how the Vega FE clocked at 1050 MHz has the same triangle throughput as a Fury X?

What you need to look at is the difference between partially culled and fully culled tests with the Vega cards. There is no difference. The primitive discard accelerator is not functional.
at 100% culling i see the RX Vega ahead at same clock speed
 
Don't worry, miners will sweep it off the shelves anyway.
 
Quote:
Originally Posted by The Robot View Post

Don't worry, miners will sweep it off the shelves anyway.
Even with ETH crashing to $120?
 
Quote:
Originally Posted by xTesla1856 View Post

One thing I still haven't seen a reasonable justification for is the horrible power draw for the performance given. 350 watts and it's "competitive" with a 1080? I'm looking forward to the X299 + CrossFire Vega work logs on this forum
biggrin.gif


Buy stock in PSU makers that have 1200+ watt models
thumb.gif
I plan on going cfx. 1300G2, mobo with plx and an extra 6 pin for the gpus, a dedicated 20a line straight to the breaker and look at this:

" It's more muted than the Radeon Pro Duo or Fury X, that's for certain." -from op source
This thing is made for obscene power consumption.
I may need to pull that old window unit ac out of the garage
tongue.gif
 
Quote:
Originally Posted by AmericanLoco View Post

Ignore the comparison between the other cards. Look at the results themselves. They do not change between 50% culling and 100% culling. The primitive discard accelerator is not working properly in Vega.
Yep. Fury X did particularly poorly at 50% for whatever reason, but there should be a very large difference for Vega between 50% and 100%. Even Polaris has a solid gain.
 
Quote:
Originally Posted by AmericanLoco View Post

Ignore the comparison between the other cards. Look at the results themselves. They do not change between 50% culling and 100% culling. The primitive discard accelerator is not working properly in Vega.
they did for the 100%
as for strip ends being higher than list on polaris aswell
Quote:
Originally Posted by Tojara View Post

Yep. Fury X did particularly poorly at 50% for whatever reason, but there should be a very large difference for Vega between 50% and 100%. Even Polaris has a solid gain.
Because has no culling.
 
Quote:
Originally Posted by PontiacGTX View Post

they did for the 100%
as for strip ends being higher than list on polaris aswell
Because has no culling.
I need to do more search into what exactly these numbers mean, because looking through Beyond3D, it appears I may not understand as well as I thought. I'll report back when I find out.
 
Idk what I can say.How come AMD approached Vega's Performance that I can't Explain ?
 
Quote:
Originally Posted by Xuper View Post

Idk what I can say.How come AMD approached Vega's Performance that I can't Explain ?
More questionable is who the manager was giving green light to this project with such numbers. Since when did AMD know this? Would be interesting to hear. Something is definately off here.

Hopefully RX Vega can at least stay under 300W.
 
There is a whole lot of awesome going on with this card, except the performance, and the mystery of the not enabled, yet specifically stated arch improvements.
Could be the best built card to deliver lackluster performance ever.
Still hoping it goes Maxwell on that Fiji.
 
I'm an Nvidia guy for sure but I really want AMD to pull through with Vega but isn't looking good. How can a card use so much power and perform at or below a 1080? Everyone makes fun of GTX480 but that card was the fastest card at its time while consuming the most power, a trade off I accepted. Water cooled Vega gets destroyed by 1080ti while using less power... I doubt magic drivers like the HD5800 series will save this card.
 
That power is being used on something but what? Cant be that power inefficient right?
 
Quote:
Originally Posted by reqq View Post

Cant be that power inefficient right?
I think it can...
 
Quote:
Originally Posted by Wimpzilla View Post

Unfortunately not, 500$ a water cooler. A good one but still!

Hope for amd that this really bad lunch and amd evictions have a sense!

Otherwise yes it is a nice joke!
For 500$ you can get a better AIO solution once block are out, and even save some money.
I really don't understand AMD's pricing.
 
21 - 40 of 49 Posts