[Rumor] RX480 AIB Card Leaked and Tested!

[Recon-UK]

if you're in Saigon I will be happy to let you take the cards for the test dude. here is the driver I upload https://www.mediafire.com/?h13f6szrakcp11i

I don't have AMD GPU's, though that could change if RX 480 is good.

And i don't understand your image, it's just Batman....

 

[gasolina]

i did say i will upload batman max settings on FHD res no nvidia gameworks enable and the pic is the benchmark result.

 

[Deleted member 50521]

I can confirm now the new RX480 beta driver cannot be applied to Fury X.

I used the driver gasolina sent me. It can only install AMD Install manager and HDMI audio driver. It cannot even detect I have radeon Fury X installed. Weird.

 

[gasolina]

I can confirm now the new RX480 beta driver cannot be applied to Fury X.

I used the driver gasolina sent me. It can only install AMD Install manager and HDMI audio driver. It cannot even detect I have radeon Fury X installed. Weird.

before the beta driver i tried other drivers same thing happened

 

[Shou Miko]

Mby the beta driver is for the RX4xx series only?

 

[xorbe]

nVidia does the same thing with the first driver.

 

[gasolina]

base on the batman test it's the same fps as the 980

 

[Caring1]

base on the batman test it's the same fps as the 980

That's what I've been saying

 

[Solid State Brain]

Gasolina, what's the average (roughly speaking) and maximum power draw from GPU-z during these benchmarks?
Could you try with default core/memory clocks with the default power limit and with the maximum power limit (from the Crimson drivers)?

If the power draw increases just by increasing the power limit, you might also end up having better performance, even with default clocks.

 

[gasolina]

Gasolina, what's the average (roughly speaking) and maximum power draw from GPU-z during these benchmarks?
Could you try with default core/memory clocks with the default power limit and with the maximum power limit (from the Crimson drivers)?

If the power draw increases just by increasing the power limit, you might also end up having better performance, even with default clocks.

150 i believe, ok i will try

 

[gasolina]

Gasolina, what's the average (roughly speaking) and maximum power draw from GPU-z during these benchmarks?
Could you try with default core/memory clocks with the default power limit and with the maximum power limit (from the Crimson drivers)?

If the power draw increases just by increasing the power limit, you might also end up having better performance, even with default clocks.

i have increase the power limit and the result is the same , here is the benchmark with 2560x1440

same fps as the 980 with 2560x1440

 

[Solid State Brain]

I see, so performance is not power limited in this benchmark. Thanks!
It looks like it might not scale as well with resolution as the R9 390X. This would have some implications for VR games.

 

[gasolina]

i will try 4k benchmark with batman hopefully with official driver it could make to 390x level and can overclock

 

[matteo]

thanks for the benchmarks Gasoline, if you have time to test also unigine @ 1080p with ultra/extreme preset would be great as well.
If the aib rumor that can handle a good overclock (1500+) is real this 480 would be a great card.

Also i would recommend to wait for the official drivers to come out, the card seems already on par with 390x/980.

 

[medi01]

So it is a waste of money to buy this RX480 if you already on GTX 970, 290X, 390 or any faster card?

It actually depends on a number of things, for how much do you can sell it and how much do you care for power consumption and "future proofness".

If one can sell 970/290x/390 for more than 480 costs, I'd say he/she must do it, if reviews the day after tomorrow deliver what we were expecting, namely, slightly below 390x-ish performance.

RX480 will be a great upgrade for 70-75% of the gamers though.

More like 80-85%. (only less than 14% were on 970/290 or better)

 

[gasolina]

thanks for the benchmarks Gasoline, if you have time to test also unigine @ 1080p with ultra/extreme preset would be great as well.
If the aib rumor that can handle a good overclock (1500+) is real this 480 would be a great card.

Also i would recommend to wait for the official drivers to come out, the card seems already on par with 390x/980.

I have the browser_x86.exe application error '' the application was unable to start correctly (0xc000007b). Click Ok to close the application'' =-=

 

[silentbogo]

I have the browser_x86.exe application error '' the application was unable to start correctly (0xc000007b). Click Ok to close the application'' =-=

Is it caused by Unigine benchmark? If yes, then install DX redistributable and MSVC11 redistributable to fix this problem.

 

[gupsterg]

I can confirm now the new RX480 beta driver cannot be applied to Fury X.

I used the driver gasolina sent me. It can only install AMD Install manager and HDMI audio driver. It cannot even detect I have radeon Fury X installed. Weird.

There is a way to add support that may work possibly, see this post on Guru3D. I have used this method in the past for other experiments.

 

[Shou Miko]

it's not uncommon for new products with "BETA" drivers to be locked to only that model or series bcs not everything has to be official to work out of the box for reviewers and testers

 

[Solid State Brain]

I have the browser_x86.exe application error '' the application was unable to start correctly (0xc000007b). Click Ok to close the application'' =-=

Unrelated with this, but speaking of benchmarks, are you able to download from Steam the SteamVR Performance test and check out the final detailed score?
There were indications earlier on that the score might have actually been lower than that of the R9 390. However, it might have run at lower clocks for this slide. It's also possible that this test could benefit from a higher power limit, so if you have time I would suggest trying it with both the default and maximum setting from the Crimson drivers.

http://store.steampowered.com/app/323910/

 

[medi01]

SteamVR benchmark results vary so wildly from card to a card with a tiny clock change, that I don't even know why they bother mentioning it.

Probably just to "demonstrate" that it is passing VR test...

 

[Solid State Brain]

The score does vary, but not as wildly as many claim, if no other programs are running in background. With an R9 390 I could hit according to the test 6.8-7.0 on stock clocks and 7.6-7.8 with a mild (100 MHz) core overclock. The latter score would put it at about stock R9 390X performance.

 

[gasolina]

i'm downloading now

 

[gasolina]

the score is 6,8

 

[Frick]

the score is 6,8

ABout the same as many GTX970's made, iirc.