Sidebar Sidebar
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.

T-Minus (3) Days Until VEGA!!

koyoot said:
You lose basically 20% of performance under macOS.
Click to expand...

When running OpenGL software, sure, due to Windows having drivers better optimised for gaming. But it has always been like that?

Ma2k5 said:
I mean I was hoping it competed better with the GTX 1050/TI Max-Q personally
Click to expand...

Are we looking at different benchmarks? What I've seen so far puts the Vega Pro 20 definitely above the GTX 1050 Ti. Not that I've seen much, there was only one fairly uninformative benchmark here...
 
leman said:
Are we looking at different benchmarks? What I've seen so far puts the Vega Pro 20 definitely above the GTX 1050 Ti.
Click to expand...

I've been seeing all sorts of different Uniengine Heaven benchmarks for the 1050 TI so maybe we are :D.
 
Ma2k5 said:
I've been seeing all sorts of different Uniengine Heaven benchmarks for the 1050 TI so maybe we are :D.
Click to expand...

Yeah, the problem with Heaven is that a) its not really clear what settings he is using so you can't really directly compare it (I can't discern it on the video, picture is too small) b) Heaven is a really old benchmark to begin with (at least use Valley if it has to be Unigine) c) Windows scores are on average 20-25% better due to better drivers.

Still, he is seeing 80% increase of performance, which should put the Vega Pro 20 somewhere midway the 1050 Ti max-Q and 1060 GTX Max-Q. Which is fairly impressive, given how small the chip is.

What we need is for people to run 3dmark and friends. Once mine arrives, I'll certainly do it, but that will still take two weeks...

P.S. The GFXBench results from the other thread also put the GPU around the (desktop?) 1050 Ti. And that is Metal on Mac vs. DX12/Vulcan.
 
Last edited:
  • Like
Reactions: Ma2k5
koyoot said:
Unigine is heavily CPU dependent.
Click to expand...

Yep, though the XPS 9570 has the same CPU so hopefully the main contributing factor was elsewhere (be it OS, dGPU or TDP limits).
[doublepost=1542810336][/doublepost]
leman said:
When running OpenGL software, sure, due to Windows having drivers better optimised for gaming. But it has always been like that?



Are we looking at different benchmarks? What I've seen so far puts the Vega Pro 20 definitely above the GTX 1050 Ti. Not that I've seen much, there was only one fairly uninformative benchmark here...
Click to expand...

As much as I worry about the KB, it is still niggling at me to buy one to try. I do find I end up going on a bender where after I upgrade the dGPU, I am tempted to also upgrade the SSD to 1TB, at which point you wonder should you upgrade it to 32GB RAM although unnecessary and suddenly you are up +£1000 :).
 
0-0 said:

Impressive performance. Beating a razer 1070.
Click to expand...

His test results released on Twitter.
PNG image.png
 
mehdijobs said:
Really? Why?
Click to expand...

Well, it is an expensive upgrade going from the RX560 to the Vega dGPU's - so if you are going for it, I would spend the extra to make it the Vega 20, it is a small % of the overall cost of the machine anyway. So Either stick with the base RX 555/560 or go all out Vega 20 is my personal opinion!
 
  • Like
Reactions: 0-0
koyoot said:
Im a bit shocked people are impressed with those scores.
Click to expand...
So we swap places I guess, I'm blown away, never expected such results to happen and I'm happy to admit I was wrong in my predictions. This never happened before, AMD having higher power efficiency than Nvidia in 3D, and by a large margin. This thing benches higher in 3D than my RX580 running as eGPU - so I guess this time Blackmagic owners are going to revolt. Higher than a 190W GPU. 560X to Vega Pro 20 is literally the same proportion in performance across compute and 3D as RX580 to Vega 56, but this time maintaining the same (or even lower as it looks) power consumption as the predecessor. I didn't see that coming.
 
  • Like
Reactions: 0-0
Ma2k5 said:
Well, it is an expensive upgrade going from the RX560 to the Vega dGPU's - so if you are going for it, I would spend the extra to make it the Vega 20, it is a small % of the overall cost of the machine anyway. So Either stick with the base RX 555/560 or go all out Vega 20 is my personal opinion!
Click to expand...

I didn’t wanna end up being trapped in Apple’s game of always being 100 dollars behind the better option.
 
mehdijobs said:
I didn’t wanna end up being trapped in Apple’s game of always being 100 dollars behind the better option.
Click to expand...

One could argue that you already fallen into a trap. $250 is a fairly steep price for only a moderate performance increase. If you really want a performance jump, Vega Pro 20 is pretty much the only real option. After all, Vega Pro 16 is a byproduct (defective Pro 20 are made into Pro 16). The only reason why it’s an option is to increase the revenue.
 
mehdijobs said:
I didn’t wanna end up being trapped in Apple’s game of always being 100 dollars behind the better option.
Click to expand...

What is $100 over 3+ years of use for a much better dGPU on a machine that costs ~$3000+? If you can cancel and re-order, I would. I would rather save $250 and go back to RX 560 than save $100 to keep the Vega 16 over the Vega 20.
 
Why do apple use dhl shipping in the uk it was spose to come today but dhl delayed it due to tech issue, so annoying I had a day of work to wait for parcel
 
leman said:
One could argue that you already fallen into a trap. $250 is a fairly steep price for only a moderate performance increase. If you really want a performance jump, Vega Pro 20 is pretty much the only real option. After all, Vega Pro 16 is a byproduct (defective Pro 20 are made into Pro 16). The only reason why it’s an option is to increase the revenue.
Click to expand...

The way you put it, it sounds like the Vega 16 is worse than the RX 560..
 
Under MacOS Vega 20 performs just like GTX 1050 Ti under Windows. Which means that the software disadvantage macOS has over Windows is around 20-25%.

Rubbish.
 
You must log in or register to reply here.
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.
Top Bottom
Back