Nvidia is also using a dedicated hardware solution, whereas the VESA and AMD solution is using software AFAIK. There would be some overhead and definitely less performance
According to the previous AMD FreeSync demo, any display that currently supports Variable VBLANK (like those eDP Toshiba laptops) would be compatible when using AMD GPUs and APUs. I'm guessing there are likely some displays in the wild right now that would support A-Sync in spirit, but not in name.
I suspect we'll see users on notebookreview trying to make this work...
I must say, I like you Ryan. You mention that there are technical differences between A-Sync and G-Sync, yet you don't mention what these differences are because you are busy or if you are under an NDA.
By saying, "what these differences are because you are busy" I mean that you don't mention that you are busy (if you are). Or if you are just being lazy.
Telling us that it would be difficult to cover without hardware in hand in the article would be useful, especially since the hardware is so far out.
People are way, way overstating the impact that the VESA update will have. All we have are vague promises that someone, somewhere, will come up with the necessary hardware to make Adaptive Sync displays work. If it were as easy as AMD claims, why did Nvidia choose a different path?
There's plenty of reason to make an article about it, if nothing else than to clear up the significant amount of misinformation and misrepresentation about FreeSync.
The slow motion video from CES looks identical to 50fps 50hz v-sync, was the demonstration able to show the display adapting the refresh rate to changing frame rates?
Cause g-sync syncs to the GPU for gaming, which this new display port standard will also require via something in the monitor itself. G-Sync is available now, Adaptive sync is not.
The biggest difference is availability, and both approaches require new hardware inside the monitor. It's expensive so far because it's low volume and implemented via FPGA. It will stay expensive until there are ASIC implementations produced in volume.
From memory, the difference is that the nVidia cards did not support the sync implementation of eDP so they had to do a hardware based work-around. Whereas some (newer) AMD cards supported the specific standard in eDP and they could exploit it by using a slightly modified laptop display that was connected via eDP.
And this is wrong... Both Nvidia and AMD supports the synching features in eDP, however, nvidia tried to extend this feature to external monitors and did that through a proprietary standard, G-Sync. AMD tried to counter it with a demo of the unrelated eDP capability (that nvidia supports too, but it NOT the point of G-Sync, confusing the public by conflating the issue). VESA is now announcing that an equivalent solution to G-Sync will be an optional part of the DP 1.2a standard, thus extending the eDP capability to external monitor through DP connection. So far, G-Sync has been demoed and has impressed most of the press. It is available today and work well (although is quite expensive as it is proprietary and recoup the RnD investment). Free-Sync using eDP was demoed, but is NOT available yet. A-Sync standard is just announced and will required 6-12 months for implementation (maybe even more if single tile 4k monitor delays are an indication of the speed of implementation of new display controllers).
Glad this made it into a display spec officially. Would also love it for TVs. Standards based innovation is time consuming and I understand the temptation to go-your-own like NVIDIA did but they really should have gone with a display standard to begin with. Glad to see this corrective action.
For those that will say this wouldn't have happened if NVIDIA didn't make g sync, I disagree. NVIDIA easily could have pushed this forward instead of making g sync... Years ago
I'm not sure what I want to see happen here. One the one hand, if this causes GSync to crater a lot of vendors will probably conclude that the whole idea isn't worth it and ASync will never take off at all. On the other hand if GSync gets a lot of momentum we'd have to worry about nVidia deciding to try and use it as a lockin lever by refusing to support ASync and forbidding their GSync board from being installed in a monitor that also supports ASync.
Open standard is more exciting for the tv and mobile market. If tv support it then it can be picked up by consoles and streaming devices and push units.
Instead of just dimming the screen when the computer thinks you're idle it could dim the screen and lower the refresh rate and perhaps give you a few extra minutes of battery life.
No idea how much power you'd save but I'm pretty sure that the laptop vendors could make new marketing material from it.
AMD and VESA are doing what's right, NVIDIA just wants to lock you in to NVIDIA.
I see G-Sync cratering badly, now there's an open standard that is actually better, due to going lower than 30Hz. Also, an open standard will be more likely to be adopted by other bits of hardware, that aren't or cannot be NVIDIA based. Good work, VESA.
The exact same thing could be said about mantle vs DirectX12 : "Nvidia and Microsoft are doing what's right, AMD just wants to lock you (and developpers) to AMD".
At the end of they day everything should be standardized.
Now Adaptive Sync just needs to fold in the other things G-Sync does like non-resistance refreshes (backlight pulsing). It's probably marginally lower latency (assuming you have a controller that can start working during display readout, which the G-Sync's board FPGa and big RAM pile can), which while somewhat pointless for desktop monitor may be more important for HMDs.
Now it's mostly down to AMD persuading monitor controller manufacturers to develop and add this feature to their upcoming DP1.2a controllers, rather than going with Nvidia's already developed solution (and probably a hefty cash incentive too, going by past Nvidia practices).
I don't think this is except to hardcore gamers. Monitors can support multiple frame rates including 24p without adaptive sync. The fact that they do not do this is a problem with the monitors not the connection.
I'd say most gamers really unless you enjoy tearing? Open standard here is definitely better if we want to see universal adoption down the road based on past experiences.
CSMR phrased it poorly; but only niche consumers, probably HC Gamers/Movie fans, will be willing to pay a significant premium for it. If it eventually becomes part of a future monitor baseline by being added to commodity panel control ASICs a much wider segment of the market will benefit from it.
What about non-hardcore gamers, for whom it will provide a smoother experience? What about people who want to watch movies with their very-difficult-to-vsync 23.976 Hz refresh rate?
Now, what are the chances we'll see cheap (ish) 4K Korean displays with Adaptive Sync in two years? Or ever? Do they go thru any sorta VESA cert? I'm still happy with my 3x 24" 1920x1200 IPS displays but the future can't come soon enough! :p
If this is implemented in eDP 1.4 then how can we test it now? There are notebooks but how to test that it is actualy working? Is there some software to tell you what is the actual "momentary" refresh rate of the notebook and if it is the same as the rendered (or video) frame rate? https://intel.activeevents.com/sf13/connect/fileDo...
We’ve updated our terms. By continuing to use the site and/or by logging into your account, you agree to the Site’s updated Terms of Use and Privacy Policy.
Facebook
Twitter
RSS
38 Comments
Back to Article
Conduit - Monday, May 12, 2014 - link
Goodbye G-Sync, it was nice knowing ya.Morawka - Tuesday, May 13, 2014 - link
IIRC Gsync performed better than AMD's FreeSync when it was first announced and tested by Anandtech.Morawka - Tuesday, May 13, 2014 - link
Nvidia is also using a dedicated hardware solution, whereas the VESA and AMD solution is using software AFAIK. There would be some overhead and definitely less performanceSoulwager - Tuesday, May 13, 2014 - link
Nope, both solutions require changes to the monitor, in order to allow the monitor to wait until the frame has finished rendering.nathanddrews - Tuesday, May 13, 2014 - link
According to the previous AMD FreeSync demo, any display that currently supports Variable VBLANK (like those eDP Toshiba laptops) would be compatible when using AMD GPUs and APUs. I'm guessing there are likely some displays in the wild right now that would support A-Sync in spirit, but not in name.I suspect we'll see users on notebookreview trying to make this work...
Soulwager - Tuesday, May 13, 2014 - link
I've seen very little in the way of FreeSync testing, and the slow motion video from CES only proves the panel in that laptop can run at 50hz.Ryan Smith - Tuesday, May 13, 2014 - link
We haven't done any comparative testing at this time. Never mind the fact that the CES demo is just a proof of concept...blanarahul - Tuesday, May 13, 2014 - link
I must say, I like you Ryan. You mention that there are technical differences between A-Sync and G-Sync, yet you don't mention what these differences are because you are busy or if you are under an NDA.blanarahul - Tuesday, May 13, 2014 - link
By saying, "what these differences are because you are busy" I mean that you don't mention that you are busy (if you are). Or if you are just being lazy.Ryan Smith - Tuesday, May 13, 2014 - link
It's beyond the scope of this article, and honestly would be difficult to cover without hardware in hand.Mand - Tuesday, May 13, 2014 - link
Telling us that it would be difficult to cover without hardware in hand in the article would be useful, especially since the hardware is so far out.People are way, way overstating the impact that the VESA update will have. All we have are vague promises that someone, somewhere, will come up with the necessary hardware to make Adaptive Sync displays work. If it were as easy as AMD claims, why did Nvidia choose a different path?
There's plenty of reason to make an article about it, if nothing else than to clear up the significant amount of misinformation and misrepresentation about FreeSync.
Soulwager - Tuesday, May 13, 2014 - link
The slow motion video from CES looks identical to 50fps 50hz v-sync, was the demonstration able to show the display adapting the refresh rate to changing frame rates?saneblane - Monday, May 12, 2014 - link
Wow, Nvidia can't catch a break these days. When asking them such tough questions at least offer them a box juice to drink.Wreckage - Monday, May 12, 2014 - link
An optional spec, that no manufacturer has pledged support for, that may not even work well...Can't wait!
SunLord - Tuesday, May 13, 2014 - link
One would assume AMD will support it and as for the display manufactures will G-sync is just as optionaliwod - Monday, May 12, 2014 - link
What is the difference between G-Sync, and A-Sync? Why does G needs an expensive extra piece of hardware to get it done?cmdrdredd - Monday, May 12, 2014 - link
Cause g-sync syncs to the GPU for gaming, which this new display port standard will also require via something in the monitor itself. G-Sync is available now, Adaptive sync is not.Soulwager - Tuesday, May 13, 2014 - link
The biggest difference is availability, and both approaches require new hardware inside the monitor. It's expensive so far because it's low volume and implemented via FPGA. It will stay expensive until there are ASIC implementations produced in volume.Death666Angel - Tuesday, May 13, 2014 - link
From memory, the difference is that the nVidia cards did not support the sync implementation of eDP so they had to do a hardware based work-around. Whereas some (newer) AMD cards supported the specific standard in eDP and they could exploit it by using a slightly modified laptop display that was connected via eDP.frenchy_2001 - Wednesday, May 14, 2014 - link
And this is wrong...Both Nvidia and AMD supports the synching features in eDP, however, nvidia tried to extend this feature to external monitors and did that through a proprietary standard, G-Sync.
AMD tried to counter it with a demo of the unrelated eDP capability (that nvidia supports too, but it NOT the point of G-Sync, confusing the public by conflating the issue).
VESA is now announcing that an equivalent solution to G-Sync will be an optional part of the DP 1.2a standard, thus extending the eDP capability to external monitor through DP connection.
So far, G-Sync has been demoed and has impressed most of the press. It is available today and work well (although is quite expensive as it is proprietary and recoup the RnD investment). Free-Sync using eDP was demoed, but is NOT available yet. A-Sync standard is just announced and will required 6-12 months for implementation (maybe even more if single tile 4k monitor delays are an indication of the speed of implementation of new display controllers).
skiboysteve - Monday, May 12, 2014 - link
Glad this made it into a display spec officially. Would also love it for TVs. Standards based innovation is time consuming and I understand the temptation to go-your-own like NVIDIA did but they really should have gone with a display standard to begin with. Glad to see this corrective action.For those that will say this wouldn't have happened if NVIDIA didn't make g sync, I disagree. NVIDIA easily could have pushed this forward instead of making g sync... Years ago
DanNeely - Tuesday, May 13, 2014 - link
I'm not sure what I want to see happen here. One the one hand, if this causes GSync to crater a lot of vendors will probably conclude that the whole idea isn't worth it and ASync will never take off at all. On the other hand if GSync gets a lot of momentum we'd have to worry about nVidia deciding to try and use it as a lockin lever by refusing to support ASync and forbidding their GSync board from being installed in a monitor that also supports ASync.chocosmitj - Tuesday, May 13, 2014 - link
Open standard is more exciting for the tv and mobile market.If tv support it then it can be picked up by consoles and streaming devices and push units.
junky77 - Tuesday, May 13, 2014 - link
Any word about whether laptops will be updated too?Raniz - Tuesday, May 13, 2014 - link
Instead of just dimming the screen when the computer thinks you're idle it could dim the screen and lower the refresh rate and perhaps give you a few extra minutes of battery life.No idea how much power you'd save but I'm pretty sure that the laptop vendors could make new marketing material from it.
shlagevuk - Tuesday, May 13, 2014 - link
Does Intel plan to implement A-Sync in their driver too?piroroadkill - Tuesday, May 13, 2014 - link
AMD and VESA are doing what's right, NVIDIA just wants to lock you in to NVIDIA.I see G-Sync cratering badly, now there's an open standard that is actually better, due to going lower than 30Hz. Also, an open standard will be more likely to be adopted by other bits of hardware, that aren't or cannot be NVIDIA based. Good work, VESA.
euler007 - Tuesday, May 13, 2014 - link
The exact same thing could be said about mantle vs DirectX12 : "Nvidia and Microsoft are doing what's right, AMD just wants to lock you (and developpers) to AMD".At the end of they day everything should be standardized.
edzieba - Tuesday, May 13, 2014 - link
Now Adaptive Sync just needs to fold in the other things G-Sync does like non-resistance refreshes (backlight pulsing). It's probably marginally lower latency (assuming you have a controller that can start working during display readout, which the G-Sync's board FPGa and big RAM pile can), which while somewhat pointless for desktop monitor may be more important for HMDs.Now it's mostly down to AMD persuading monitor controller manufacturers to develop and add this feature to their upcoming DP1.2a controllers, rather than going with Nvidia's already developed solution (and probably a hefty cash incentive too, going by past Nvidia practices).
HaryHr - Tuesday, May 13, 2014 - link
What is Intel going to do? Intel holds largest percent of market, true not gaming oriented but marketshare non-the-less.Do you think Intel will support proprietary Nvidia solution or Vesa standard?
edzieba - Tuesday, May 13, 2014 - link
Given that the extension to Adaptive Refresh is only an optional part of the DP1.2a spec, probably neither.CSMR - Tuesday, May 13, 2014 - link
I don't think this is except to hardcore gamers.Monitors can support multiple frame rates including 24p without adaptive sync.
The fact that they do not do this is a problem with the monitors not the connection.
CSMR - Tuesday, May 13, 2014 - link
I don't think this is _useful_ except to hardcore gamers.Icehawk - Tuesday, May 13, 2014 - link
I'd say most gamers really unless you enjoy tearing? Open standard here is definitely better if we want to see universal adoption down the road based on past experiences.DanNeely - Tuesday, May 13, 2014 - link
CSMR phrased it poorly; but only niche consumers, probably HC Gamers/Movie fans, will be willing to pay a significant premium for it. If it eventually becomes part of a future monitor baseline by being added to commodity panel control ASICs a much wider segment of the market will benefit from it.Guspaz - Tuesday, May 13, 2014 - link
What about non-hardcore gamers, for whom it will provide a smoother experience? What about people who want to watch movies with their very-difficult-to-vsync 23.976 Hz refresh rate?Impulses - Tuesday, May 13, 2014 - link
Now, what are the chances we'll see cheap (ish) 4K Korean displays with Adaptive Sync in two years? Or ever? Do they go thru any sorta VESA cert? I'm still happy with my 3x 24" 1920x1200 IPS displays but the future can't come soon enough! :pvlado08 - Wednesday, May 14, 2014 - link
If this is implemented in eDP 1.4 then how can we test it now? There are notebooks but how to test that it is actualy working? Is there some software to tell you what is the actual "momentary" refresh rate of the notebook and if it is the same as the rendered (or video) frame rate? https://intel.activeevents.com/sf13/connect/fileDo...