- Jun 4, 2004
- 17,408
- 15,253
- 146
I can’t say I totally agree with his performance conclusions. He compared an aftermarket MSI Gaming 1070Ti against a reference RX Vega 56.
With the various turbo algorithms used by each manufacturer the improved cooling capabilities of an aftermarket cards are very important to get the best performance.
I have a Powercolor Red Devil RX Vega 56 that I’ve been testing.

(I’m hoping to finish up a review of the Red Devil like i did for my Threadripper 1900X - https://forums.anandtech.com/threads/my-threadripper-1900x-semi-formal-review-part-1.2544112/ )
In 3DMark Timespy Guru3D had the stock RX56 at 6601/6264 (total score/GPU score) and the MSI 1070Ti at 7123/6840. That’s 9% slower than the 1070Ti.
My Red Devil on the TR1900X Out of the box scores 7120/6885 @223W of GPU power. It’s effectively equivalent to the MSI 1070Ti.
OCing the MEM to 900mhz and undervolting (neither of which I’ve maxed out) gets me 7234/7025 for 208W, 15W less than stock. (Noise is basically the same as stock as I haven’t played with fan speeds yet).
Playing around with core clocks I’ve hit ~7300/7120 which puts me only 3% behind a stock 1080.
So my basic point is - compared apples to apples the aftermarket RX Vega 56s and 1070Tis are basically equivalent. And both are almost on top of the 1080.