Eurogamer Article: http://www.eurogamer.net/articles/digitalfoundry-2017-microsofts-xbox-one-x-benchmarks-revealed
YouTube Video: https://www.youtube.com/watch?v=R_7iD9wjdCE [10:26]
Remember when we revealed the hardware specs for what was then called Project Scorpio? Well, developers had a similar briefing - but with a fat pile of GPU benchmarks. We've managed to source them, and here they are.
YouTube Video: https://www.youtube.com/watch?v=R_7iD9wjdCE [10:26]
OP can you at least post the pictures?
1 of 3: Xbox One at native resolution compared to performance at 2160p from a barebones Xbox One X port. Higher is better. The data is expressed here in terms of percentage. 1080p scales well, 900p holds up better than you might expect. 720p? Predictably not.
2 of 3: Here are the frame-rates, converted from frame-times (seen in the next slide). Higher is better. 1080p titles gain frame-rate moving to 4K, with two out of three 900p games within five to six per cent of matching performance at 4K.
3 of 3: This is the original Microsoft data based on frame-time measured in milliseconds from PIX GPU captures. Lower is better. Title D is a 900p game with scaling issues, but everything else looks promising bearing in mind the very basic nature of the ports.
1 of 3: A percentage increase comparison showing base Xbox One performance (blue at 100 per cent), followed by back-compat (red) and native Xbox One X code compiled under the latest XDK (green). Older titles can be recompiled to get a big performance boost.
2 of 3: Here's the same data, this time in basic frame-rate terms. Only Title A doesn't see a huge improvement in performance with back-compat, while the boost with native Xbox X code (a barebones port it should be noted) is massive in all cases.
3 of 3: Just like the 4K captures, the original data is based on PIX captures measured in frame-time (milliseconds). Here's that data - lower is better here. You might say we're looking at 1.31TF vs 3TF vs 6TF performance here at the same resolution.