• 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.
LoadsaMoney said:
I know it doesn't actually say that anywhere, but saying yes, grab the drivers from here, and switch to 'Gaming Mode', to 'I Just Game', is saying, you can just use it purely as a 'Gaming Card'.
Click to expand...

AMD are obviously catering for people who want to also game using this card, who in their right mind would buy a Vega FE for purely gaming purposes when there are better options available, and with Vega RX just around the corner which is rumoured to be faster and will no doubt be considerably cheaper..
 
N19h7m4r3 said:
Well that sucks, as without a similar CPU or system it's not too accurate. It would be like testing a FE on Ryzen vs Phenom 2.
Click to expand...

If it's a purely GPU limited test then within limits the CPU it's tested on won't make much difference, should be within a few % margin of error. Eitherway, assume worst case scenario and the CPU is causing a 16% reduction in performance, that still only puts it on par with Fury...
 
HeX said:
If it's a purely GPU limited test then within limits the CPU it's tested on won't make much difference, should be within a few % margin of error. Eitherway, assume worst case scenario and the CPU is causing a 16% reduction in performance, that still only puts it on par with Fury...
Click to expand...

Aye, just pointing it out. If someone was testing NVIDIA vs AMD they'd point out the massive discrepancy in the CPU platforms as well.
 
akarypid said:
Let's not forget that AMD had a very clear picture of where the competition was at. They've know what the 1080 performance is for over a year. Same thing for the Titan X that was launched last August.

What I have a hard time believing is that they knew all this and said "let's make a card that performs like the 1080, and release it a year later". I am pretty sure they would have aimed higher than that.

Either something has gone horribly wrong (if this is all the card can achieve) on the hardware side, OR there is a lot of untapped potential due to the drivers not using any of the new tech.

This is the only thing that makes me think we will get a nice bump (as in 20%) for RX Vega, with perhaps more added later as they keep working on these drivers.
Click to expand...
My guess is a combination of manufacturing issues and HBM2 supply combined with having a complete mare writing the drivers. Given the architecture changes the driver work must be substantial and it seems like the're having trouble getting them working, otherwise why would core features like the tiling be disabled? Maybe they bit off more than they could chew having the Ryzen launch so close.

I can't imagine the FE is the launch AMD wanted and it seems unlikely that they can do so much driver work in just a month. I wouldn't be surprised if there are substantial gains over the next few months though, that's pretty standard for them. They really need to get their launch drivers sorted.
 
N19h7m4r3 said:
Aye, just pointing it out. If someone was testing NVIDIA vs AMD they'd point out the massive discrepancy in the CPU platforms as well.
Click to expand...

Indeed, the ideal would be identical platforms without question. I think, however, that's why they've gone with this particular test as it negates the CPU advantage of the Ryzen platform. Hopefully someone with a Ryzen and an FE will do the same test to compare.
 
N19h7m4r3 said:
Well that sucks, as without a similar CPU or system it's not too accurate. It would be like testing a FE on Ryzen vs Phenom 2.
Click to expand...

wait a second I've just looked into it a bit more, the benches are on this page (post 41) http://www.planet3dnow.de/vbulletin...tion-Bilder-Benchmarks-Energiemessungen/page2

and he is using a 1700 ryzen, and he got 2718 on DX and 2039 at 1050/stock (or maybe 800) the dude with the xeon don't know where he came in but it's irrelevant, as these 2 results on this page that was linked above ( https://www.forum-3dcenter.org/vbulletin/showpost.php?p=11419198 ) shows only Ryzen cpus

On a side note the guy with the Xeon got better results ( https://www.forum-3dcenter.org/vbulletin/showpost.php?p=11419302 ) than the one with ryzen, I don't know if his ram was OC'd to 1000 or if he meant 500X2

EDIT : and his Xeon is at 4.1 by the looks of it
 
Last edited:
Rezident said:
Great, there is nothing quite like the lead-up to a new high-end gpu architecture. I can almost taste the reviews now, can’t wait.


And Vega has had some buildup!
Click to expand...

That's an understatement if ever there was one :D

I expect we'll seem some leaks (some of them real :D) shortly. They'll give us something to talk about :p
 
HeX said:
Apples to Apples: Fury X vs. Vega FE at same clock rates

Radeon R9 Fury X - clock rates 1050/500 MHz (Stock) - Superposition 1080P Extreme 3322

Radeon Vega FE - clocked at 1050/1000 MHz - Superposition 1080P Extreme 2794

https://www.forum-3dcenter.org/vbulletin/showpost.php?p=11419198

https://www.forum-3dcenter.org/vbulletin/showpost.php?p=11419302
Click to expand...


First clear indication that this could be a driver issue OR AMD have regressed with extra 5 billion transistors, die shrink, faster memory and 30% IPC increase.
 
Steampunk said:
Someone at AMD has decided there will be no gaming benchmarks until the RX gaming card is released. FE is a prosumer card, so they don't care about gaming for the initial release. Notice how well FE does in professional apps benchmarks, and how badly it does in gaming? I really think there's loads of stuff that's not yet enabled on the FE because they are waiting on the RX release.

It makes no sense to release a card on a new process and loads more transistors if it can't even match their own card from a couple of years back. So either FE is awful and it's representative of RX, or FE is not representative of RX because AMD are not showing their hand until the release of the RX.
Click to expand...

Yep that is it. We will do this, oh it looks terrible dont care do it anyway. Stick to the plan. They are just resolute in their vision of this what we will do. I mean if they are right on some things I hope they keep doing what they think is best.
Just playing to the crowd just means they dont go past dx11 performance or anything really.

https://youtu.be/yXC7pv39y8o?t=633
The scale of this generation for them is quintuple from 580 capability ? so results dont match that potential

+10% for drivers in a month is fair guess unless they as above AMD done this deliberately and already knew how to improve performance but just have not wanted to place Frontier in that way (yet) Then who knows, maybe +40% I just wonder how much is mistake and how much is misunderstood product placement etc

or wait few months for volta.
Click to expand...
Volta is next March my guessimatometer
 
chaosophy said:
From what I remember FS2 just does some extra processing for HDR before it hits the monitor.
The card+FS2 does some extra work for the HDR image (something to do with colour mapping) that means that the monitor is doing less to the image than say how a normal HDR TV does it.
So basically reduces the time it takes the monitor to display the image, there by reducing overall input lag.
Click to expand...

I haven't seen anything on that but it sounds worth a read, Do you know where I can find the article is please
 
iB CoMpL3XiTy said:
AMD are obviously catering for people who want to also game using this card, who in their right mind would buy a Vega FE for purely gaming purposes when there are better options available, and with Vega RX just around the corner which is rumoured to be faster and will no doubt be considerably cheaper..
Click to expand...

Yes, but with the FE not even topping the 1070 at gaming, no one can see how the RX would be any/much different, as its the same chip, its basically just an FE, with 8GB, instead of 16.

Higher clocks, and getting better optimized drivers for gaming, might just take it inbetween the 70 and 80, which lets be honest, would still be crap.
 
Last edited:
LoadsaMoney said:
Yes, but with the FE not even topping the 1070 at gaming, no one can see how the RX would be any/much different, as its the same chip, its basically just an FE, with 8GB, instead of 16.
Click to expand...

Well putting aside for a moment that there seems to be some issue(s) with FE performance, Do we actually know for sure that RX is exactly the same as FE, minus some RAM, have AMD said as much?
 
LoadsaMoney said:
Yes, but with the FE not even topping the 1070 at gaming, no one can see how the RX would be any/much different, as its the same chip, its basically just an FE, with 8GB, instead of 16.

Higher clocks, and getting better optimized drivers for gaming, might just take it inbetween the 70 and 80, which lets be honest, would still be crap.
Click to expand...

This is what I fear, it needs to be well ahead of a 1070 in everything. if it matches a 1080 or beats it in some areas and is around 1070 price (wishful thinking) then id be happy. I am really starting to doubt though.
 
This is where they've cocked up.

  • I Just Want to Game
    If you want to game on a Windows® platform, get your driver here and switch to “Gaming Mode.”
Click to expand...

If you 'Just Want To Game', then you should wait for the RX to be launched at Siggraph, as Raja said back at that event (which ive forgotten was called).
 
Status
Not open for further replies.
Back
Top Bottom