Warning: "continue" targeting switch is equivalent to "break". Did you mean to use "continue 2"? in /homepages/9/d659054407/htdocs/app660029368/wp-content/plugins/magneto-sidebar-generator/html.php on line 83

Warning: session_start(): Cannot start session when headers already sent in /homepages/9/d659054407/htdocs/app660029368/wp-content/plugins/wpmu-dev-facebook/lib/external/facebook.php on line 60

Warning: session_start(): Cannot start session when headers already sent in /homepages/9/d659054407/htdocs/app660029368/wp-content/plugins/wpmu-dev-facebook/lib/external/facebook.php on line 60

Warning: session_start(): Cannot start session when headers already sent in /homepages/9/d659054407/htdocs/app660029368/wp-content/plugins/wpmu-dev-facebook/lib/external/facebook.php on line 60

Warning: session_start(): Cannot start session when headers already sent in /homepages/9/d659054407/htdocs/app660029368/wp-content/plugins/wpmu-dev-facebook/lib/external/facebook.php on line 60
Radeon Chill Put to the Test: How Does it Actually Perform? - Page 2 of 2 - Tech Altar

Rise of the Tomb Raider Radeon Chill Test

Rise of the Tomb Raider is a good test, because it’s dynamic and there’s a lot going on even while standing still. A whole world, at least the visible portions, that’s alive is being rendered. Potentially we could see a drastic decrease in performance as workloads are synced, and become increasingly less frequent.  Our tests are at 1080P so that we can get the best available data possible.

Radeon Chill

Looking at the framerates alone we see that Chill does have an effect on our experience, with slightly lower framerates than without. With more action going on, say when we’re spinning around like confused morons, the data and idle time is less frequent anyway and results in a slightly lower FPS. Not terrible, really. Expected, and the overall subjective effect is less than you might think.

That only tells part of the story, however. If we take a gander at the power consumption we can see that it has a real effect. And thus we also see a cooler GPU as well.

Radeon Chill

Power consumption during some portions tends towards being much lower, and in other situations it can be slightly lower. That translates into a real, positive effect on power consumption. Temperatures for the XFX cooling solution are already very good, but we saw a 5-degree Celsius decrease on average with Radeon Chill activated. That’s not insignificant.

But what of frame times? Theoretically the synchronization of work should make it much more smooth. But in reality that just wasn’t the case. Frame time to the screen actually took longer than without Chill activated.

Radeon Chill

During bouts of heavy movement, we actually found it to be much worse. There was a perceivable delay, though it was only slight. compared to turning Chill off. The idea would be sound if there were enough resources and the architecture were even faster. It could theoretically lead to instantaneous, or sub 2ms, frame times should the card actually be fast enough to deliver. Once it’s rendered, you’re only limited by the amount of time it takes for the monitor to display it. That, of course, would be a future use. Stacking work and allowing idle time gives some semblance of a delay, though not much compared to using Chill right now.

Conclusion

So is this a useful feature or not? Radeon Chill does have some very clear advantages in terms of power consumption and keeping overall usage low. It takes all that work that needs to be done and spreads it out until it’s actually needed, making for a more efficient way of computing. It just won’t satisfy those looking for maximum performance, or even those that just magically want a cooler running GPU. But it is a step into the future, of making things efficient and keeping power consumption low. We just wish that we could use it on every game at our own risk.

Overall a great new feature that seems to work in regards to one aspect they’ve promised, and works well.