AMD Crossfire Freesync Update: Delayed
by Ryan Smith on April 30, 2015 6:45 PM EST
When AMD launched Freesync back in March, one of the limitations of the initial launch version was that only single-GPU configurations were supported. Multi-GPU Crossfire systems could not be used with Freesync, requiring users to trade-off between Crossfire and Freesync. At the time AMD claimed that Crossfire Freesync support would be coming in April, however as April comes to a close it has become clear that such a release isn’t going to happen.
To that end, AMD has posted a short update on the status of Crossfire Freesync over on their forums. In the update, AMD states that after QA testing they believe that Crossfire Freesync is “is not quite ready for release” and that they will be holding it back as a result. Unfortunately AMD is not committing to a new release date for the feature, but given the fact that it’s more important to get this right than to get it out quickly, this is likely for the best.
Source: AMD
Facebook
Twitter
RSS
92 Comments
View All Comments
Socius - Friday, May 1, 2015 - link
After having tried gsync, and loving it, I realized that both gsync and freesync are baby steps towards what needs to come next: ULMB. Current iterations of ULMB in gsync monitors like the Acer Predator XB270HU are limited to 100Hz, do not work at the same time as gsync, and cut down on screen brightness.In the testing I did, I came to the realization that ULMB a is absolutely godly and 144Hz ULMB with gsync/freesync would be an even better step than going to something like 240Hz.
So put the gsync/freesync crap aside. ULMB has the potential to be significantly more important than this technology.
edzieba - Friday, May 1, 2015 - link
Unfortunately, the two are mutually exclusive until the brightness variance problem can be solved. When you don't know how long until the next display update will be, you don't know how bright to pulse the backlight for the current update. Underpredict, and the display will dim. Overpredict and it will be overbright.chizow - Friday, May 1, 2015 - link
No, the dimming byproduct is strictly due to the fact the light is pulsing for a lower amount of time, so less light is hitting your retina. Even if they fix ULMB to work with G-Sync you will get reduced brightness, this is well-known trade off of ULMB that will never get fixed. They already strobe the LED to higher brightness, but if you are only getting light for a fraction of the time due to the pulse, the macro effect is overall dimmer light. It is also not unlike the effect using 3D Active shutter glasses, the dimness is due to the shutters cutting at least 50% of the light hitting your retina.chizow - Friday, May 1, 2015 - link
I do think ULMB + G-Sync is the next evolution of G-Sync, possibly G-Sync 2.0. Tom Petersen did leave this as a possibility in his interview with PCPer with a "no comment but stay tuned" response, which I have found generally means they are already working on it and there is a high likelihood it makes it to market if they find a way to implement it.Personally I like ULMB at high FPS, but at lower FPS, I find G-Sync is definitely the more noteworthy tech, but a solution that incorporates both without compromise would be great. Petersen explained how PWM was used to control the strobe backlighting with ULMB and said that was hard to get right with a variable refresh rate, but since that interview, PCPer was able to conclude (and Nvidia more or less confirmed) Nvidia's G-Sync module takes over the OverDrive process from the monitor. If they could implement similar controls over whatever IC is responsible for the PWM rate, that would open the door for G-Sync + ULMB together.
The Von Matrices - Friday, May 1, 2015 - link
I wish AMD's driver team would stop promising release dates it can't realistically meet; it does more harm to their brand than the benefit of the sales gained by promising the feature. AMD did the same thing with their Crossfire frame pacing driver, which was also missing deadline after deadline. If you are not sure when you will have a driver ready, then you shouldn't give a specific date of its release. Also, AMD, despite what you may argue otherwise, releasing a buggy beta driver on the deadline does not qualify as a full release.chizow - Friday, May 1, 2015 - link
I agree, and I fully expected more AMD fans and users to have this same sentiment, because honestly, you are the ones who are impacted the most by this kind of behavior, and this is the main reason it would be difficult for me to buy and support AMD products. I just don't agree with their development and release philosophy as it just doesn't meet my needs. Plus, I know there's an alternative out there that does it better.Oxford Guy - Sunday, May 3, 2015 - link
I also wish Nvidia wouldn't do things like release $300+ cards with half the VRAM speed of a 2007 midrange card, and not bother to tell anyone about it.chizow - Sunday, May 3, 2015 - link
Cool yeah I wish AMD didn't sell everyone extra graphics cards as CrossFire and claim it actually improved their gaming experience when it actually made it worst.Gunbuster - Friday, May 1, 2015 - link
Coming Soon™, it will be great just keep waiting and chugging down the FUD they spread about the competitors...chizow - Friday, May 1, 2015 - link
Exactly, it would have been much better if AMD just took their time developing this feature and released it when it was ready, but they spent much of the last 18 months disparaging G-Sync, buying time, trying to slow adoption with all of their FUD. And now, they're stuck with an inferior solution that they won't be able to backtrack from (DP AdaptiveSync) because of all the disparaging FUD they said about why Nvidia needed to use their own, expensive, "license fee" ASIC.