RejZoRBut in a way we also need to look at the situation how NVIDIA has money to throw around and they felt despite performance penalty, it won't be an issue. AMD on the other hand doesn't have resources to throw around (at least not graphics division) and they aren't sure if it's worth investing resources into officially supporting a feature that will by default be inferior to implementation on RDNA4. They don't want to outright say no because they might one day implement it, but they don't want to promise anything long term either.

yes I do see that... I also meant the whole sentence i quoted (not only the part I highlighted) -> I understand them not introducing the new feature to old cards. BUT I would really appreciate it if they did.

RejZoRPersonally, if they are unsure they should just say it, but leave it open. Instead they seem to flip flop around and not say anything. Like, say it's not planned or in the works right now, but they won't say no either. And that would be it. If they one day drop it, people will be happy. And if some moved on to RDNA4 specifically because of this in the meanwhile, then so be it.

Main issue with AMD is really this flip flopping around like they themselves don't know what they are doing or they just really suck at communicating. In which case, they should maybe hire new PR team. AMD released a lot of really good stuff, but they are just so bad at telling the world what they have and what they are bringing.

^ Also this, it annoys me how they communicate with us/the consumer

it just fosters these very unproductive conversations everytime a new "FSR4 on old hardware" Headline drops...

wolfis the attempts to invalidate and gatekeep having an opinion and indeed real empathy on this subject because of irrelevant and arbitrary barriers, so really, it's them telling us how we should feel about it.

Considering most comments here are boring, you at least managed to make me laugh with this one. But go ahead, keep listening to the voices in your head, because all you managed to do was turn this thread into another “them vs us” debacle.

I've put people on ignore for less than that, so off you go. It’ll eventually become so quiet that I might even hear my bowel movements.

ThomasKall you managed to do was turn this thread into another "them vs us" debacle.

Projecting it seems, you literally said it first ["feel like they have to come up here and tell people how they should feel about it"] and I provided another perspective that you apparently didn't want to hear. A position seemingly so fragile that you immediately go to 'you're on the ignore list'. You sure showed me :rolleyes:

Looks like AMD finally figured out that the less they say about something the better.

Keeping things ambiguous is better that just saying what everyone knows and putting their foot in their mouth.

Also let's fan theories flourish. I'm now hearing that they're not saying anything because contractual obligations to Sony and their PSSR upscalers which would essentially be FSR4 Int8 given PS hardware, so they have to keep mum until that's launched and patched up.

Meh don’t use it anyway.

RexterNow there is another reason "not" to buy AMD, yes you pay premium for NVIDIA, but you get 7+ years of feature updates. Lets hope they change course if there is enough pressure.

Nvidia locks newer tech to newer generations as well. Not always of course, same as AMD.

I've no idea why AMD is holding back the INT8 version from older cards though. We've seen performance is good enough on those cards. The bad will they've gained is not going to be worth it for them and it's just another baffling AMD choice that nets them nothing in return.

cinemawareGod forbid some top end - and actually not that important - features could be exclusive to the latest HW.

Meanwhile, the substantial IQ improvements from 4.5 are available across the board.

The issue is that you are arbitrarily deciding what is and isn't top end. Such logic would be a get out of jail free card for Nvidia or AMD to exclude any software features they want. It's arguable that all these features are "top-end". DLSS, FG, Reflex, VRR. All could be considered "top-end" depending on the perspective of the speaker. After all discrete GPUs are inherently much much more powerful than what the majority of people have. Both Nvidia and AMD have even used this logic in the past to market some of their cards.

I don't see this as a valid argument, it's just a crux that fanboys will use to justify their big corpa crush screwing people over.

In addition, the post you are replying to was in turn replying to a comment stating Nvidia gives it's cards 7 years of feature support. I've seen you and so many people tout Nvidia's software support. Well, part of that being a big selling point of your cards is ensuring all the latest features get ported to cards at least 7 years old (if not more if the hardware is there). It doesn't mean "oh well only the features I care about" or "only non "high end" features". I expect significant effort to make these features available on older cards if it's even remotely workable. Every PC gamer should be demanding this because the extended software support is a big part of modern GPU pricing, particularly for Nvidia where you are paying 15-25% more just for the extra support over what AMD provides. You are paying a lot for it.

If Nvidia want to have tiered software support where "top-end" features are locked to higher SKUs, it needs to stipulate that when selling the video cards. Same goes for AMD. You can't run RT well on lower end cards but it will run.

RejZoRNot only that, people don't mention that while RTX 2000 and 3000 series support latest models, no one mentions both have severe performance penalties when using newer models.

This is true but at least they have access to using the newer models. If it's a question of performance, they should push game devs to add a warning message in game about the performance hit on prior generations. I can understand that most people are not wired into the tech loop and may assume performance is bad for the tech.

nguyenparticipate in online arguments

Say: AMD- facilitate online GPU arguments since 7970 (GHz edition..)

ThomasKSelective criticism, that's what it is.

But what amuses me the most is the fake empathy coming from people who do not own the hardware in question, but feel like they have to come up here and tell people how they should feel about it, just to stir the pot. If you care enough to pay attention, is the same people, over and over.

Exactly, it's selective criticism, people high on the mind share will say how great Radeon cards are, until someone really asks for buying recommendations, then it's all about " mah dee ell ess ess" and frame gen so magical it can retrieve the next frame with a time machine so you'll never feel the extra input lag. Or that a GeForce card is always worth a $300 premium because everyone cares about workstation uses.

It's always the same people over and over again riding in on a high horse, first to tell everyone how shit they think AMD is while passive aggressively mocking and attacking anyone with and AMD GPU. This same bullshit evangelism every time for a brand while being condescending to others only wants me to buy that brand less, and heck if AMD stops making dGPU's I'd rather find a new hobby, Nvidia users would love the full monopoly anyway.

Hecate91Exactly, it's selective criticism, people high on the mind share will say how great Radeon cards are, until someone really asks for buying recommendations, then it's all about " mah dee ell ess ess" and frame gen so magical it can retrieve the next frame with a time machine so you'll never feel the extra input lag. Or that a GeForce card is always worth a $300 premium because everyone cares about workstation uses.

It's always the same people over and over again riding in on a high horse, first to tell everyone how shit they think AMD while passive aggressively mocking and attacking anyone with and AMD GPU. This same bullshit evangelism every time for a brand while being condescending to others only wants me to buy that brand less, and heck if AMD stops making dGPU's I'd rather find a new hobby, Nvidia users would love the full monopoly anyway.

It's always you and Thomas defending AMD no matter what too. The difference is that you're not unreasonable like him. It cuts both ways. We always go over the pros and cons.

Folks - drop the personal jabs.

Nobody cares who ignores who - don't post about it.

RejZoRAnd that's perfectly valid feeling. But in a way we also need to look at the situation how NVIDIA has money to throw around and they felt despite performance penalty, it won't be an issue. AMD on the other hand doesn't have resources to throw around (at least not graphics division) and they aren't sure if it's worth investing resources into officially supporting a feature that will by default be inferior to implementation on RDNA4. They don't want to outright say no because they might one day implement it, but they don't want to promise anything long term either.

Personally, if they are unsure they should just say it, but leave it open. Instead they seem to flip flop around and not say anything. Like, say it's not planned or in the works right now, but they won't say no either. And that would be it. If they one day drop it, people will be happy. And if some moved on to RDNA4 specifically because of this in the meanwhile, then so be it.

Main issue with AMD is really this flip flopping around like they themselves don't know what they are doing or they just really suck at communicating. In which case, they should maybe hire new PR team. AMD released a lot of really good stuff, but they are just so bad at telling the world what they have and what they are bringing.

That may have been true 10 years ago. Today, AMD is clearing $35 billion in revenue. They could afford to let a dev or two just work on backporting features.

evernessinceNvidia locks newer tech to newer generations as well. Not always of course, same as AMD.

I've no idea why AMD is holding back the INT8 version from older cards though. We've seen performance is good enough on those cards. The bad will they've gained is not going to be worth it for them and it's just another baffling AMD choice that nets them nothing in return.

Its simple: AMD wants you to buy new cards, not keep old ones. So they gatekeep features.

evernessinceThe issue is that you are arbitrarily deciding what is and isn't top end. Such logic would be a get out of jail free card for Nvidia or AMD to exclude any software features they want. It's arguable that all these features are "top-end". DLSS, FG, Reflex, VRR. All could be considered "top-end" depending on the perspective of the speaker. After all discrete GPUs are inherently much much more powerful than what the majority of people have. Both Nvidia and AMD have even used this logic in the past to market some of their cards.

I don't see this as a valid argument, it's just a crux that fanboys will use to justify their big corpa crush screwing people over.

In addition, the post you are replying to was in turn replying to a comment stating Nvidia gives it's cards 7 years of feature support. I've seen you and so many people tout Nvidia's software support. Well, part of that being a big selling point of your cards is ensuring all the latest features get ported to cards at least 7 years old (if not more if the hardware is there). It doesn't mean "oh well only the features I care about" or "only non "high end" features". I expect significant effort to make these features available on older cards if it's even remotely workable. Every PC gamer should be demanding this because the extended software support is a big part of modern GPU pricing, particularly for Nvidia where you are paying 15-25% more just for the extra support over what AMD provides. You are paying a lot for it.

I have no idea what you are on about with this "top end". The only feature nvidia has "gated" is MFG over 2x, which doesnt function on ada and earlier because they dont have the hardware to support it. I feel like everyone wilfully ignores that part. Everything they CAN support, DLSS 4, transformer, 2x FG, all of that has been backported to the 2000 series.

AMD, meanwhile, knows the INT8 version of FSR4 runs on rDNA3, yet refuses to backport any of it. Apples and oranges responses from the two. Nobody is complaining that rDNA3 doesnt have the new FG from redstone (because they dont have the hardware to support it), they are complaining that the INT8 version of FSR4 is being artificially blocked.

evernessinceIf Nvidia want to have tiered software support where "top-end" features are locked to higher SKUs, it needs to stipulate that when selling the video cards. Same goes for AMD. You can't run RT well on lower end cards but it will run.

What "top end feature" has nvidia artificially locked to a higher SKU? What can a 5090 do that a 5060 cannot?

The only time nVidia has ever done this was the GTX 1600 series, which were stripped of RT cores to make budget hardware. They were sold with full acknowledgement that they were stripped cores that couldnt do RT and they were lauded for giving the consumer choice.

Hecate91Exactly, it's selective criticism, people high on the mind share will say how great Radeon cards are, until someone really asks for buying recommendations, then it's all about " mah dee ell ess ess" and frame gen so magical it can retrieve the next frame with a time machine so you'll never feel the extra input lag. Or that a GeForce card is always worth a $300 premium because everyone cares about workstation uses.

Selective criticism would imply that people ignore nvidia doing what AMD is also doing. This is untrue. That would require nVidia to not port DLSS 4 backwards to the RTX 2000/3000 series.

Whether these features work well on older cards is a completely different argument, the fact is nvidia allows users to do so, AMD does not.

Hecate91It's always the same people over and over again riding in on a high horse, first to tell everyone how shit they think AMD is while passive aggressively mocking and attacking anyone with and AMD GPU. This same bullshit evangelism every time for a brand while being condescending to others only wants me to buy that brand less, and heck if AMD stops making dGPU's I'd rather find a new hobby, Nvidia users would love the full monopoly anyway.

AMD is being criticised for their artificial gatekeeping, and rightfully so. If they want to be a premium brand with premium prices, they need to compete with the big boy. When the evil, greedy nGREEDia is out here giving users 8 years of feature support (so far) and 10 years of driver updates, while you put out 4-5 years of updates for some cards and only provide feature updates for the newest stuff, youre gonna get heat for it.

Theres an easy solution, one that AMD refuses to do: port FSR4 to rDNA 2/3. Others have already done the work. But AMD refuses, because at the end of the day AMD is just as greedy and wants you to upgrade every generation to get the new toys.

RejZoRAnd that's perfectly valid feeling. But in a way we also need to look at the situation how NVIDIA has money to throw around and they felt despite performance penalty, it won't be an issue. AMD on the other hand doesn't have resources to throw around (at least not graphics division) and they aren't sure if it's worth investing resources into officially supporting a feature that will by default be inferior to implementation on RDNA4. They don't want to outright say no because they might one day implement it, but they don't want to promise anything long term either.

Personally, if they are unsure they should just say it, but leave it open. Instead they seem to flip flop around and not say anything. Like, say it's not planned or in the works right now, but they won't say no either. And that would be it. If they one day drop it, people will be happy. And if some moved on to RDNA4 specifically because of this in the meanwhile, then so be it.

Main issue with AMD is really this flip flopping around like they themselves don't know what they are doing or they just really suck at communicating. In which case, they should maybe hire new PR team. AMD released a lot of really good stuff, but they are just so bad at telling the world what they have and what they are bringing.

AMD has about 18% margin on client and gaming according to their revenue earnings news, I wonder if that includes gpus but their Radeon division might be seriously underfunded as AMD has prioritized the CPU market,servers and compute hardware before gpus. Compared to Nvidia having 60-70% margins with billions to throw at game developers and an influence on nearly the entire market. I expect someone will think I'm blindly defending AMD for not offering wider FSR support, but the fact is development costs money, playing catch up on features can't be easy when Nvidia can just pull some feature out of their ass and convince the gaming masses they can't live without it.

I think AMD should have just went ahead with an FSR4 lite or FSR4 beta to make everyone happy and to avoid the negative press, because tech sites will nitpick everything AMD does.
AMD has communicated things badly before, so them taking missteps here isn't a surprise, it's like Radeon engineers aren't talking to the marketing people, or the marketing people have more say on what goes on than the engineers do. It does sadden me that AMD missed the chance they could've taken to fully launch FSR4 in some form to cards at least back to RDNA2, now people would be less accepting of an FSR4 beta test or an open source driver. I have no use for upscaling but it would be interesting to try without the need for workarounds.

Hecate91AMD has about 18% margin on client and gaming according to their revenue earnings news, I wonder if that includes gpus but their Radeon division might be seriously underfunded as AMD has prioritized the CPU market,servers and compute hardware before gpus. Compared to Nvidia having 60-70% margins with billions to throw at game developers and an influence on nearly the entire market. I expect someone will think I'm blindly defending AMD for not offering wider FSR support, but the fact is development costs money, playing catch up on features can't be easy when Nvidia can just pull some feature out of their ass and convince the gaming masses they can't live without it.

I think AMD should have just went ahead with an FSR4 lite or FSR4 beta to make everyone happy and to avoid the negative press, because tech sites will nitpick everything AMD does.
AMD has communicated things badly before, so them taking missteps here isn't a surprise, it's like Radeon engineers aren't talking to the marketing people, or the marketing people have more say on what goes on than the engineers do. It does sadden me that AMD missed the chance they could've taken to fully launch FSR4 in some form to cards at least back to RDNA2, now people would be less accepting of an FSR4 beta test or an open source driver. I have no use for upscaling but it would be interesting to try without the need for workarounds.

Youre right, development DOES cost money.

FSR4 was ported to rdna3 by one guy. ONE guy. So yeah, the multi billion dollar AMD is going to get nitpicked for not doing what both intel and nVidia can do, and what one hobbyist can do.

AMDs 18% margin gave them $4.335 billion in net profits. Think they can support hiring ONE guy to move this stuff?

Its not a money problem. AMD isnt some poor underdog. They have the staff to do it already, and redstone having INT8 code shows the had already developed it. Valve was doing it FOR FREE on linux and got shut down by AMD. It was a business decision, done to segregate the market. It wasnt done to save money or because they are suffering.

Theyre gonna get a lot of heat for using greed to drive their decision to abandon older users. If nVidia did this, they would be criticized too. If Intel did this, they would be criticized too (and they WERE, there was a lot of heat when intel straight up abandoned users with 1 year old hardware using Xe cores in their iGPUs).

Costs about $0 to open source their FSR code. Intel and Nvidia are already ahead of them in the ML game so I don't know who they're hiding their code from

outlw6669FSR 4 is available through a Driver Override of FSR 3.1 when using AMD GPU's.

It's one switch in the Adrenalin menu. If clicking that is too much stress for you, then I'm really out of things to say.

Dr. DroJust read this.



This is modern AMD in a nutshell. Suits run the show, and the developers are just basically code monkeys with no real agency over their work. This has been an endless source of frustration for me, and if I may be honest, more than half the reason I dropped them after the VII.

Like that isn't true of every single western company. I'm sure developers can customise Nvidia code to their liking, right? ;)

Harun8I think the biggest crime AMD has committed here is not only not open sourcing their code

Please, show me open source driver for nVidia cards released by nVidia.

GrimlingIf this limitation of FSR4 would only effect OLD products, i could understand.
But the fact that brand new Products from AMD, like Strix Halo, that are based on RDNA3 variants, and therefore limited to FSR3.1 (which is absolute crap)

Leave that crap for yourself.

Radeon 8060S is not brand new.
Jan 6th, 2025 - RDNA3.5 based iGPU in Ryzen AI Max+
Mar 6th, 2025 - RDNA4 based GPU 9070XT

Am I happy taht RDNA 3.5 has not got ML upscaling? No.
Do I care about ML upscaling on pre-RDNA4 dGPUs? No.
I just do not have any game slowed down by RT effects, so even 760M (iGPU from Ryzezn 8600G) is able to run them without upscaling.

AusWolfIt's one switch in the Adrenalin menu. If clicking that is too much stress for you, then I'm really out of things to say.

Yes, one switch in the control panel, IF you are playing a game that supports FSR 3.1, and then only IF you have a RDNA 4 GPU.

In comparison, FSR 1 works with every GPU that is able to start the game, and FSR 2 through FSR 3.1 work with Polaris/Pascal and newer.
XeSS also works with RDNA 1/Pascal and newer GPUs.

The issue is, compared to the user base of non-RDNA 4 GPUs, there are only a miniscule amount of gamers using RDNA 4 GPUs.
To this, while FSR 3.1 is still often better than TAA, it is still worse than other options, such as DLSS and XeSS (which supports most GPUs).
As such, there is minimal incentive for devs to implement support for FSR 3.1 or FSR 4 in games.

The entire point of this thread is that INT8 support for FSR 4 has already been developed.
That it is not being officially implemented, in my opinion, is doing a disservice to both gamers, as well as AMD, as it would allow most GPUs to use it, and would make it an enticing feature for devs to add to their games.

wolfI've been using the new L and M presets on my 3080 rig, as has my best mate too. It's genuinely shocking how good preset L looks in Ultra Performance mode at 4k, or perhaps my new favourite scaling factor, a custom forced 42% render scale fed into preset L (UP is 33% and P is 50%). The performance penalties are absolutely not so severe that they make the new models undesirable, on the whole, for all users of all supported hardware.

Now do that on 2060/3050 with 6GB of VRAM and tell about the not severe impact on performance.

melkor_unlimitedPlease, show me open source driver for nVidia cards released by nVidia.

AMD was the only GPU company with open source upscaling. Now they're not special and closed source like everyone else

TheinsanegamerNSelective criticism would imply that people ignore nvidia doing what AMD is also doing. This is untrue. That would require nVidia to not port DLSS 4 backwards to the RTX 2000/3000 series.

nVidia did not backported whole DLSS4 to rtx 20xx/30xx cards.
It just happened that requirements of some parts of DLSS pack are low enough they are still able to run on them. Somehow.

Harun8AMD was the only GPU company with open source upscaling.

Did that help to get more marketshare?
Did that help to get higher revenue?
Did that pleased shareholders?

AusWolfIt's one switch in the Adrenalin menu. If clicking that is too much stress for you, then I'm really out of things to say.

I assume they're referring to how Optiscaler works?
Then again I've seen people whine the Adrenalin menu is too hard to use. 🤣

AusWolfLike that isn't true of every single western company. I'm sure developers can customise Nvidia code to their liking, right? ;)

This. And I'm sure Nvidia drivers have been completely perfect too.

TheinsanegamerNSelective criticism would imply that people ignore nvidia doing what AMD is also doing. This is untrue. That would require nVidia to not port DLSS 4 backwards to the RTX 2000/3000 series.

Whether these features work well on older cards is a completely different argument, the fact is nvidia allows users to do so, AMD does not.

Ignoring that MFG 6X requires a RTX 5000 series card is selective criticism as well.
And so is saying how well features work, if the features cause a significant performance impact those features are worthless,2000&3000 don't work well at all with DLSS4.

TheinsanegamerNAMD is being criticised for their artificial gatekeeping, and rightfully so. If they want to be a premium brand with premium prices, they need to compete with the big boy. When the evil, greedy nGREEDia is out here giving users 8 years of feature support (so far) and 10 years of driver updates, while you put out 4-5 years of updates for some cards and only provide feature updates for the newest stuff, youre gonna get heat for it.

Theres an easy solution, one that AMD refuses to do: port FSR4 to rDNA 2/3. Others have already done the work. But AMD refuses, because at the end of the day AMD is just as greedy and wants you to upgrade every generation to get the new toys.

AMD is hardly a premium brand with premium prices, remember what Nvidia tried with the 4080 12GB for $900? Or how a $800 x70ti tier is "midrange" now. That's thanks to Nvidia putting a premium on top of their premium, the more you buy the more you save.

But only AMD can be greedy for artificial gatekeeping, Nvidia is the knight in a shining leather jacket even though they gatekept DLSS for too long with some optical flow bullshit, older cards don't get MFG 6X even though it would most benefit the 8GB cards they keep selling.
And yes, FSR4 would benefit 8GB cards too, however official development with getting game companies to use their features costs money, not just a third party tool.

TheinsanegamerNYoure right, development DOES cost money.

FSR4 was ported to rdna3 by one guy. ONE guy. So yeah, the multi billion dollar AMD is going to get nitpicked for not doing what both intel and nVidia can do, and what one hobbyist can do.

AMDs 18% margin gave them $4.335 billion in net profits. Think they can support hiring ONE guy to move this stuff?

Its not a money problem. AMD isnt some poor underdog. They have the staff to do it already, and redstone having INT8 code shows the had already developed it. Valve was doing it FOR FREE on linux and got shut down by AMD. It was a business decision, done to segregate the market. It wasnt done to save money or because they are suffering.

Theyre gonna get a lot of heat for using greed to drive their decision to abandon older users. If nVidia did this, they would be criticized too. If Intel did this, they would be criticized too (and they WERE, there was a lot of heat when intel straight up abandoned users with 1 year old hardware using Xe cores in their iGPUs).

One guy who made a workaround that requires a bunch of work to get it working, and then it always doesn't work or look nearly as good as FSR4 on RDNA4 hardware. And you have to run Optiscaler on Linux to get any real benefit from it, also while running the risk of getting banned in any multiplayer game for using unsupported code.
AMD has $4.4335 billion for their entire client and gaming side of the business, go compare that to how much Nvidia or Intel has in net profits.
It is a money problem to be keeping up with Nvidia and Intel, they are the tiny underdog fighting against two much larger corporations which are also collaborating together. It isn't as simple as taking the tool one guy made relabling it as FSR4 Lite and stuffing it into the Adrenalin panel.

They would've gotten heat regardless of launching it for all GPU's or not since Nvidia pulled MFG 6X out of their butts, people would've moved the goalposts and critcized AMD for not having the same level of AI fake frames. As I mentioned, Nvidia doesn't get nearly the same amount of criticism when they pull dumb stuff, neither does Intel, they got excused for Xe support getting dropped for 1 year old hardware because it only affected iGPU's.

melkor_unlimitedDid that help to get more marketshare?
Did that help to get higher revenue?
Did that pleased shareholders?

Idk I don't have business numbers. I just know I as a consumer have one less reason to buy Radeon now. Did close sourcing it increase marketshare & revenue?

melkor_unlimitednVidia did not backported whole DLSS4 to rtx 20xx/30xx cards.
It just happened that requirements of some parts of DLSS pack are low enough they are still able to run on them. Somehow.

nVidia backported all of DLSS4 except for one thing, MFG, which requires the new blackwell hardware to function.

Again, by definition "selective criticism" would mean people pointing out AMD doing something were ignoring nVidia doing the same thing. That is categorically untrue. nVidia is not doing what AMD is doing, and people will compete at the olympic level in mental gymnastics to justify AMDs actions.

melkor_unlimitedDid that help to get more marketshare?
Did that help to get higher revenue?
Did that pleased shareholders?

Did closing the software get them more marketshare? More revenue? Did it please shareholders?

Hecate91I assume they're referring to how Optiscaler works?
Then again I've seen people whine the Adrenalin menu is too hard to use. 🤣

This. And I'm sure Nvidia drivers have been completely perfect too.

Ignoring that MFG 6X requires a RTX 5000 series card is selective criticism as well.
And so is saying how well features work, if the features cause a significant performance impact those features are worthless,2000&3000 don't work well at all with DLSS4.

AMD is hardly a premium brand with premium prices, remember what Nvidia tried with the 4080 12GB for $900? Or how a $800 x70ti tier is "midrange" now. That's thanks to Nvidia putting a premium on top of their premium, the more you buy the more you save.

But only AMD can be greedy for artificial gatekeeping, Nvidia is the knight in a shining leather jacket even though they gatekept DLSS for too long with some optical flow bullshit, older cards don't get MFG 6X even though it would most benefit the 8GB cards they keep selling.
And yes, FSR4 would benefit 8GB cards too, however official development with getting game companies to use their features costs money, not just a third party tool.

One guy who made a workaround that requires a bunch of work to get it working, and then it always doesn't work or look nearly as good as FSR4 on RDNA4 hardware. And you have to run Optiscaler on Linux to get any real benefit from it, also while running the risk of getting banned in any multiplayer game for using unsupported code.
AMD has $4.4335 billion for their entire client and gaming side of the business, go compare that to how much Nvidia or Intel has in net profits.
It is a money problem to be keeping up with Nvidia and Intel, they are the tiny underdog fighting against two much larger corporations which are also collaborating together. It isn't as simple as taking the tool one guy made relabling it as FSR4 Lite and stuffing it into the Adrenalin panel.

They would've gotten heat regardless of launching it for all GPU's or not since Nvidia pulled MFG 6X out of their butts, people would've moved the goalposts and critcized AMD for not having the same level of AI fake frames. As I mentioned, Nvidia doesn't get nearly the same amount of criticism when they pull dumb stuff, neither does Intel, they got excused for Xe support getting dropped for 1 year old hardware because it only affected iGPU's.

Im not going to address the "muh MFG" cope anymore. Asked and answered why you are wrong already.

AMD wants to be a premium brand, but isnt willing to put in the effort that the ACTUAL premium brand does in supporting their products. This is a self made problem that is well within AMDs capability to fix. They make enough money to hire someone to finish the INT8 code already made. The fact a hobbyist makes it work at all is a strong indication that an official employee with official code could do a much better job. AMD can sacrifice 0.001% of their annual profits to make that happen, if they choose.

If nVidia was doing this, they would be torched. But they are not. AMD is doing it, and people keep making excuses for it. Enjoy 6% marketshare, I guess? Id rather AMD not act like this, and provide support that mirrors the price of their hardware, and actually COMPETE with nvidia instead of lapping up scraps.

TheinsanegamerNIts simple: AMD wants you to buy new cards, not keep old ones. So they gatekeep features.

I mean they can try but they don't have the market leverage to do that. Their marketshare is the smallest it's ever been at 4% and GPU and memory price increases ensure sales are going to heavily dip. It's akin to a small time fruit stall vendor jacking the price of their product up to try bilk customers for more, only to find out they aren't hot stuff and they lost their long term customers as they all moved back to shopping at the supermarket.

I have no idea what you are on about with this "top end"

That's more or less one of the points I was making. The guy I was responding to coined that term. As I stated in my comment, It's as capricious as it sounds.

melkor_unlimitedNow do that on 2060/3050 with 6GB of VRAM and tell about the not severe impact on performance.

So follow me here, that's why I said:

wolfThe performance penalties are absolutely not so severe that they make the new models undesirable, on the whole, for all users of all supported hardware

You want to pick the absolute lowest possible SKU to hang your hat on, I haven't personally tested that one. Might try with my A2000 which performs around that level. It's been demonstrated to work well with a 2070, so a 2060S, 2070(S), 2080(S), 2080Ti, 3060, 3060Ti, 3070(Ti), 3080(ti), 3090(Ti) at a minimum would all be able to reap rewards.

I don't argue the impact is considerably higher, I'm saying it's still worth it in certain situations and it's absolutely awesome that Nvidia implemented it officially that far back.

Hecate91then it's all about " mah dee ell ess ess" and frame gen so magical it can retrieve the next frame with a time machine so you'll never feel the extra input lag. Or that a GeForce card is always worth a $300 premium because everyone cares about workstation uses.

I like that your points are so strong you need to wilfully misrepresent the points you don't like, essentially another strawman by phrasing them the way you do, which nobody who talks of the merits of those features does. Classic.

Hecate91if the features cause a significant performance impact those features are worthless,2000&3000 don't work well at all with DLSS4.

And a straight up lie. It's not worthless at all on RTX2000 and 3000, ask me how I know. Please, stop spreading misinformation, it's not a good look.

TheinsanegamerNAMD can sacrifice 0.001% of their annual profits to make that happen, if they choose.

See it wouldn't even be a sacrifice though. A tiny, utterly miniscule portion of their revenue/profits would need to be allocated to it, but the benefits it brings would in turn bring in more revenue and profits. It's not doing it that's costing them money, and I'd wager it's going to cost them a boatload more through their mishandling of it than it would releasing it.