Your math escapes me, care to elaborate more so that I can understand the reasoning?
If Vega can't beat Nvidia in perf/w and/or outright performance, then we already know how it will do against Volta. Assuming Vega is 50% more efficient than RX 480, it'll have to consume 270-280 watts on average to match the 1080 TI. That won't bode well at all against Volta.
If Vega can't beat Nvidia in perf/w and/or outright performance, then we already know how it will do against Volta. Assuming Vega is 50% more efficient than RX 480, it'll have to consume 270-280 watts on average to match the 1080 TI. That won't bode well at all against Volta.
Your math escapes me, care to elaborate more so that I can understand the reasoning?
It uses HBM2 and the current 14nm process used in the RX580 is better than the RX480. Old slides state 225w TDP for Vega and 300w for Vega x2 (and don't forget that 14nm+ should be out before Volta).
If the current process for the RX 580 is better, I'd love to see tangible proof. RX 580 and 570 took a huge dump on efficiency and had massive regressions vs. RX 480 and 470.
I'm asking about the actual performance of the chip.
If Big Vega can't beat Big Pascal, how in god's name will Vega 11 be competitive against Volta midrange in 2018?
We don't know what Vega performs like yet, that's why we're in this thread.... A thread for rumors and speculation. If you don't want to discuss rumors or speculation on Vega, stop posting and wait til the chip is released.We don't even know what Vega 10 performs like yet, so there isn't much value in speculating on speculation. I assume AMD has the relevant information to make good decisions about how to design the product and where it should fit into the market. If nothing else, we just see a repeat of what always tends to happen, AMD lowers their prices and NVidia gets a better margin.
It appears I was off. Vega would only be about 10% less efficient than 1080 TI IF it were to be 50% more efficient than RX 480, putting it's power consumption in the 250-260w average range if it's the same speed as 1080 TI.
We don't know what Vega performs like yet, that's why we're in this thread.... A thread for rumors and speculation. If you don't want to discuss rumors or speculation on Vega, stop posting and wait til the chip is released.
That's why we rolled out the Crystal Ballshit.Yes, its a form for rumors and speculation, but not all rumors and speculation are equally good.
If the current process for the RX 580 is better, I'd love to see tangible proof. RX 580 and 570 took a huge dump on efficiency and had massive regressions vs. RX 480 and 470.
That's because the clockspeed is higher which uses way more power.
If the current process for the RX 580 is better, I'd love to see tangible proof. RX 580 and 570 took a huge dump on efficiency and had massive regressions vs. RX 480 and 470.
To be fair, he is asking about the process, not the architecture. Vega will presumably be built on the same process as the 580, and whether there are any benefits in that regard over the 480/470 is a good question. I'd say the answer is no. Of course, as you point out, that's not the whole story.you simply can't compare 580 or 480 to Vega. They aren't the same cards. This isn't one giant launch of various forms of Pascal, where you can accurately extrapolate performance across nVidia's entire release this last year, from card to card.
They are different architectures.
I don't think that Vega is late. I think that due to limited R&D, they skipped over another major revision in order to add features to compete with Volta before it came out. AMD is focused on the professional market and they need to make sure they get chips in before it's too late. As you've seen, both Vega and Volta have a heavy enthuses on packed math.
I think this is a very valid point, especially when you take into consideration how different GPU architectures can be when compared to CPU architectures. The much bigger inter-generational changes in GPU architectures make driver optimization almost as important as hardware optimization, which is never the case with CPUs. Although Fiji hasn't seen the runaway improvements of Hawaii (and to a certain degree Polaris), it's still pretty clear that AMD's Achilles heel in the GPU space has been drivers and software. Fiji was a very unbalanced architecture - probably a bet on some requirements being higher than others at higher resolutions, or some such - which mostly explains why it hasn't blown past the 980Ti (yet?).Why I don't take much stock in Ryzen vs Vega marketing is because GPU's are different from CPU's.
Drivers can make a far bigger impact to even a mature architecture than BIOS or microcode can to a CPU. Essentially, Ryzen was near peak performance for months while Vega might not be ready until launch.
If I can't preorder Vega, and can't pick it up on day 1, I doubt I'll get a Vega high end chip by the end of 2017.IllogicalGlory, your comments about not boding well make sense and then I tried to factor in the recent announcement of Vega in the upcoming Apple I Mac. I wonder if AMD's RTG division is spread too thin and is really focusing on Apple instead ?
I've decided to stay with a single RX480 for my "all AMD" build with Ryzen 1800x and not jump ship back to Nvidia. I have Nvidia cards in my other 2 rigs and they both work great.
The RX 480 is "right sized" for the monitor I'm using, LG 29UM67 (2560 x 1080 UW) but starts running out of gas at higher resolutions. Conversely both of my Nvidia cards GTX1080 and even the GTX 980 TI SC still rock at 1440p. I had 2 RX480s in CF to "compensate" but just sold one in anticipation of the Vega launch. Missed the tidal wave surge of the RX480/580 bit mining by about a week-such is luck!
I love custom water cooling but held off on adding blocks to both 480s due to cost and the mistaken belief that Vega would appear sooner.
I now get a vibe that big Vega for consumer/gamers (not the Frontier FE) will launch in the end of July (that's been confirmed) but be in extremely short supply for awhile. I prefer to add my own water block and not use AMD's AIO solution such as Fury X so I will probably have to wait significantly longer before a big Vega that uses an EK designed block becomes available.
I just snagged a "scratch and dent" EK RX 480 acetal/copper full water block for a great price from PC Performance and it should be delivered in a few days. I've decided to complete the water cooling cpu/gpu loop for the R7 1800x/RX 480 and ride it out until the Vega I want is really available with the right water block. I estimate 6 months.
This also gives time for RTG to polish the Vega drivers.
He may have been talking about Vega 11.For this to be true, Vega would have to make 1080Ti it's chew toy.
Would be great if true, but I think this is "optimistic" based on what we've seen so far.
This kind of thing would turn the GPU market upside down, at least if priced in the $500-$600 range. (NVIDIA has proved pretty conclusively you can add sell cards that beat up all competition pretty badly for over $1000 but not sell a lot of cards in terms of market share)
Add in the FE cards, likely a chunk going to Apple and currency miners and it could be potentially a difficult time obtaining one at release. I honestly doing know what I'm going to do if we end up with a trickle for 2017. I'm definitely not buying Pascal and imo Volta consumer won't be available until 2018.If I can't preorder Vega, and can't pick it up on day 1, I doubt I'll get a Vega high end chip by the end of 2017.
If I can't get it on day 1, I'll potentially be looking at Volta.
Between all of the potential people that want Vega, I think lots of normal gamers will get crowded out.
Add in the FE cards, likely a chunk going to Apple and currency miners and it could be potentially a difficult time obtaining one at release. I honestly doing know what I'm going to do if we end up with a trickle for 2017. I'm definitely not buying Pascal and imo Volta consumer won't be available until 2018.