Possible Radeon RX Vega 3DMark Time Spy Benchmark Result
Click here to post a comment for Possible Radeon RX Vega 3DMark Time Spy Benchmark Result on our message forum
rock1m1
That's incredibly disappointing. Hope for AMD's sake it is not true
Bogeyx
What else to say?
Kaotik
Let's put it this way, at reported clocks Fury X would be as fast or faster > It's OC'd Fury (X)
Denial
This could just be the cut down Vega that's supposed to compete with the 1070. Or it could be full Vega clocked lower.
People need to stop forming conclusions on these leaks. They're almost never accurate.
Hilbert Hagedoorn
Administrator
1) Fury is 4GB only
2) HBM memory cannot OC to 700 MHz
ivymike10mt
It's looks like rebranded Fury X, rather than just a new 2017 Vega product.
I don't belive that will look their flagship product in 2017.
GeniusPr0
Hi Hilbert,
Doing a score compare reveals a driver name of 687F:C1
PrMinisterGR
What's interesting is that the reported driver version is the OpenCL one (22.19.384.20), instead of the D3D one.
http://imgur.com/2bpHofJ.jpg
The clock could also just be a part of extrapolating performance, a leak to measure reactions, or a ton of other things. I personally don't believe that any Vega part that goes versus the Ti/XP, will have anything less than 1.5GHz. They were promising passively cooled server parts at 1.5GHz, so...
Hilbert Hagedoorn
Administrator
That's a good spot, and the ID matches the presumed AOTS Vega leak back in December.
Anarion
That would be disappointing if they need one extra year and HBM 2.0 to match GTX 1070. That would mean no competition again when NVIDIA launches Volta.
If it's noticeably cheaper and uses less power... Then it could be nice.
Denial
Calling them passively cooled is a bit of a stretch, but yeah - MI25 probably around 1.5Ghz - the gaming variants should be higher, not 300mhz lower.
In the other thread, GeniusPr0 brought up a good point though - if you take these measurements and pretend the clock was 25% higher you'd get ~5-10 above stock 1080 performance. Which is where we saw Vega perform in BF1/Doom a few months back.
Fox2232
There are "vega" "687F:C1" values from compubench.
And they look like 35~45% above Fury X across the board.
Ryu5uzaku
If this is their topline card it's Fury X at 1200mhz pretty much, with more RAM. It's like everything they should have done has not happened.
Or maybe it is a Fury X done with hbm2 for their 1070 adversary 😀 I don't know.
@ fox
And this looks like it's barely above fury x.
w0nderz
I'd lean more towards an RX 590 candidate.
SirDremor
Uh oh, has someone crashed the hype train?
Endymion
Man, I need something like 1080 Ti level of performance, not this.
rock1m1
There was no RX 490, so there will be no 590.
MorganX
I don't believe it, or it's a low-end Vega that will sell for ~$300. We've already seen demos with higher performance than that and if this was the case for Vega, there's no way Nvidia would have slashed their prices and given up all that free money they've been raking in. Lastly, if this were to be true, AMD is done in GPUs. Period. I call BS.
Srsbsns
Noway no how is this a full Vega because I can get 5600 score on my overclocked Fury X.
Stormyandcold
I refuse to believe it. This isn't big Vega. It's bs.
