Ran two quick tests on RTS / Extreme with D3D. First test ended up at average 10.25 FPS, second test 27.66 FPS. For the second test I disabled motion blur in the extreme settings. Can't wait to do the same tests with Mantle drivers when they are released.

I'll have to try it out when I get home. Exciting

From the Steam Community Hub it seems like people are getting decent performance even without mantle. Seems like this demo is a lot more optimized than what we were seeing in the previous videos? The demo never saw the game go higher than 30fps

Of course now I have to wonder if I need to get a wretched AMD card.... ewww...

OK, i gave it a spin last night and its very nice indeed, although HW demanding. My system is Core i7 980x @ 3,78 GHz, 24 GBs of RAM, geforce gtx590, win 7. Installed on SSD. Since Nvidia GPU, no Mantle, just D3D.

Few random thoughts>

i tried the RTS only with high difficulty (not extreme) and the FPS plummeted on occasion (zooming out, rotating camera) to 0-10 range.... i take this is down to GPU not being fast enough (played it with just one GPU, not both, equal in performance to cca GTX570), that CPU would be fine to run it without stutter otherwise....or not?

turning the temporal AA off helped the performance quite a lot. It ran almost fluidly, without any major hiccups, i dont think the FPS dropped under 20, atleast i did not notice. 

the graphics were not so great . I mean, i expect bit more. Obviously the smallish fighters, when you zoom in onto them, are completely on another level polygon- and detail-wise, when you compare them to fighter/bomber models from Rebellion. But the specular highlights were bit over the top metallic, it seemed like the models lacked some kind of bump or normal maps...it kinda looked bit cartoonish. Was not blown away by the lighting solution in general - CryEngine it is not. Maybe this was only down to art direction, rather than technical solution though.

Do not want to sound pessimistic or what, it was indeed cool otherwise and i cant wait for another Sins (or any other potentially similar game in general) to be released on this engine.

it would have been nice if the custom settings has some ui bits to go with it other than telling you to go edit a .txt

.... not that my machine even meet the min req (core 2 duo + 7770)

got a grey screen out of it and a hugely lagged mouse cursor. a look at (mid) settings suggest it was using 1920*x1080 or even 2560*? when my desktop resolution is set at 1280*800... did a hard reset and after that, the launcher doesn't even show the menu.. just a greyed out app window... trying to use command line on a custom lower setting gave me a grey screen and less lagged mouse cursor.. deleted and gave up. probably too optimistic. might give it another go when there's a non-beta mantle driver. (nevermind the absent beta driver )

(speaking of min req... saints row 4 has a higher req than saints row 3... but strangely 4 ran better than 3 on my machine which doesn't meet the min req of 4. then again, i suppose it seems to me more like an expansion than a new game.. refined engine and re-used assets)

There has been an FAQ posted which answers some questions people have been asking.

http://www.oxidegames.com/2014/01/31/star-swarm-faq/

Hi Found this as well.

http://arstechnica.com/gaming/2014/01/amd-almost-rolls-out-mantle-its-high-performance-alternative-to-direct3d-and-opengl/

Glad that's up.  I've read a lot of posts from people not understanding that the default config for Star Swarm is a stress test.  If it didn't bring a typical machine today to its knees then it's kind of failed as a stress test.  

Well, the Mantle vs DirectX review site comparisons are beginning to surface. This german site shows a 2x to 3x increase using Mantle...

http://translate.google.com/translate?hl=en&sl=de&tl=en&u=http%3A%2F%2Fwww.golem.de%2Fnews%2Famds-mantle-api-im-test-der-prozessor-katalysator-1402-104261-4.html&sandbox=1

It would appear that us regular peeps will see the Mantle/Catalyst 14.2 drivers sometime today, probably sometime this afternoon... note that you will need an AMD GCN GPU to take advantage of the Mantle drivers - the rest of us are still stuck with DirectX!

Well, that's impressive.

So you did not run those stress test on right settings.... it happened to me to. I OC´ed my CPU to 3,9Ghz and ran Intel Burn Test on certain settings and it completed just fine... but then i played CoD and on occasion, after one or 2 hours of gameplay time, booom - BSOD. So i inquired a little on Anandtech forums and i was told i shall be running the stress test on different settings, since the ones i used are not as "stressful". So i did and the test gave me error right away this time around....that moment i knew i have to either underclock the CPU somewhat or bump the voltage.

I'll just remove the overclock on the CPU for now, it's not doing me any good anyway. Star Swarm runs fine without it on mantle and DirectX. It was more that I was curious what the difference would be between a 4670k on 3.4 ghz and a 4670k on 4.2ghz (with a radeon 7990 behind it).

The reason I was interested in that is that Star Swarm seems very uneven in how it uses my GPU. It'll go up to 100% sometimes, and sometimes it'll stay around 60%. So I imagine the CPU was bottlenecking at the 60% point.

I'd imagine that if the game designers push things too far with their game releases, and GPUs start catching fire left and right, that there could be a class action suit involved...

That being said, I don't see people running a stress benchmark 24/7/365, so I think the Oxide guys are safe for now...