I just wanted to start a discussion on Vega, I know the launch is getting close and it seems like it could use some hype in here. So what is everyone's speculation on what Vega is bringing to the table for us?
From what I gathered, i anticipate a HIGH demand for Vega and I would encourage a large production of cards because people are literally scouring the internet everyday for specs of information and basically ready to toss their money into Amds hands. I for one don't care what the price is I will be buying one the second they hit the shelf. I know when I start up games I will be smiling knowing I bought a long anticipated and beautifully crafted video card.
HBM v.2 ?
HBM is always nice - don't see it mentioned in here though: AMD 687F:C1 performance in CompuBench - performance benchmark for various compute APIs (OpenCL, RenderScript)
Either way like a lot of people, Vega is a must have for my new show system
Quote from WCCFTech ---AMD’s high-end Vega 10 GPU will be available to consumers in the first half of 2017. The chip spans a die size of over 500mm2 from early calculations and features two HBM2 stacks, incorporating up to 16 GB of HBM2.
Currently waiting for it to pair with my Ryzen 1800x PC, then my build will be complete.
from some of the rumors ive gathered was 4 gb and 8gb versions higher performance than 1080 but lower than 1080ti at around 399-599
Hope that is true... That is around what I budgeted for a new personal GPU... the last GPU that I spent $400+ was on a GTX 670 = I should've bought a R9 290 because look at these two GPUs now.
I would love to get one when it comes out but I want to see how it performs first, If it is only slightly better than my R9 390 then I don't think I will be getting it. plus I only use a 60 Hz monitor so if the frames are too high i would get a lot of frame tearing. if i was running a higher speed monitor i would be more excited to get one, but right now i am just excited to see Nvidia get taken down a notch by AMD.
What do you consider slightly better? The R9 Fury X was already 25% faster on average than the R9 390, which I think is more than slight. If nothing else there will be substantial gains over the Fury X from the 28nm to 14nm process shrink alone.
I play on triple monitor Eyefinity setup 7680x1440@60Hz with Quadfire R9 290X. On some games I have to lower resolution or graphic quality. I have my fingers crossed that Vega will be near twice a R9 290X in performance.
As far as Vega is concerned, if you look at the base numbers it should be possible to get a reasonable idea of how Vega will perform.
I have listed both the GTX 980 Ti and R9 Fury X in the chart as starting points as the two GPUs were close in performance.
If you look at the GTX 1080, it has had virtually everything cut down from the 980 Ti. Less shaders, less texture mapping units, only a huge boost in clock speed due to the 16nm process and an increase in memory capacity. RX Vega will almost certainly be substantially faster than the GTX 1080 owing to the fact that it has similar gains in clock speed and memory over the R9 Fury X with zero cuts to shaders, TMUs or ROPs.
The GTX 1080 Ti is virtually the same card as the GTX Titan X. to the point where we can consider them the same for the comparison. Compared to the GTX 980 Ti, there is still the substantial increase in core clocks but now the shaders and TMUs are bumped up 27% while the ROPs have been scaled back up as well. Based off that alone, the GTX 1080 Ti should outperform Vega. The two cards have similar increases in memory and core clock, while the 1080 Ti gains shaders and TMUs over its predecessor while Vega does not.
There are however a few other improvements Vega has under the hood that may make up the difference. Most notability in the case of all purpose gaming is the new pixel engine.
Using a tiled approach allows the GPU to only shade the pixels that are actually visible in a scene, making the shaders far more efficient. So while Vega doesn't have more shaders than Fiji, they will be more efficient due to not wasting time shading pixels no one can see. This is actually a similar approach to what NVidia did with Maxwell leading to sizable gains in the 900 series from the 700 series. But will that be enough to overcome a 27% gain in shaders the 1080 Ti enjoys?
We won't know until the card launches but I speculate that NVidia seems to think it will. If you recall, when the R9 Fury X was released, it actually performed closer to the Maxwell based Titan than expected. NVidia countered with the GTX 980 Ti, a card only marginally slower than the aforementioned Titan. Despite the fact that there was limited difference, NVidia priced the 980 Ti at the Fury X's price ($650) and left the Titan at $999. This time, NVidia released a 1080 Ti that was actually slightly faster than the Titan card and then released a new Titan with even more shaders and TMUs. That deviates from previous behavior and leads me to believe that RX Vega will likely end up slightly faster than the 1080 Ti but slower than the Titan Xp. If Vega were slightly slower than the 1080 Ti, they would have had no reason to create a new +$1000 card in the product stack.
Retrieving data ...