DirectML vs FidelityFX SR vs Radeon Boost vs DLSS

Rumors recommend that quickly, AMD will launch a brand new driver suite that guarantees two new options and updates: FidelityFX SR and Radeon Boost, which though the latter is already enabled in its drivers, would include enhancements to spice up efficiency. NVIDIA got here first, hit earlier than and now it has to compete, however can it do it with its everlasting rival and towards Microsoft?

One normal or 4 totally different? Where is the online game business going?

SDR vs HDR Gaming

This is absolutely the query we should ask ourselves, the place are the gamers going and the place is the business trying? And it’s that with the panorama that’s offered to us plus some choices which might be free within the business, we’re going to should face DLSS vs Fidelity FX SR vs Radeon Boost vs DirectML.

As now we have talked about, all of them have their peculiarities, however on the similar time all of them do the identical in essence: scale the decision to enhance picture high quality. So why is every firm creating its proprietary expertise and never being provided as an open normal or becoming a member of forces underneath one?

It is a superb query to which now we have no reply, however absolutely since it’s not an API as such, however is a approach of rescaling with AI, every firm intends to implement it within the engines of the primary builders with the excuse of exclusivity for his or her {hardware}.



We have talked at size about DLSS, so we are going to summarize fairly a bit for these causes. It is a expertise that makes use of processors for AI Tensor Cores from NVIDIA to spice up each efficiency and graphics high quality.

All by way of a deep studying neural community, which focuses on 4 major enhancements over its predecessor: superior picture high quality, nice scaling throughout all RTX GPUs, a neural community for all video games, and the flexibility to get choices customizable.

In abstract, NVIDIA is the one which develops and maintains it, the implementation is carried out by the sport engine, it has a specific recreation assist that may allow it, it’s Full Frame and is simply out there for the structure Volta, Turing, Ampere and successive.

AMD Radeon Boost


Although it’s not the direct competitor as such, it’s true that Radeon Boost overshadows the NVIDIA pool. The drawback that AMD has is that the best way of implementation is thru the Adrenalin driver of the corporate, the place for extra issues the video games enabled for this expertise will not be so plentiful.

In addition and on account of its nature, if we analyze it rigorously we are going to see that though it’s Full Frame its approach of working is completely dynamic, though it’s able to supporting architectures from GCN to RDNA and better.

The issues to check it with NVIDIA is that it actually works worse as a result of it’s supported by the motive force and it’s the one which does the work and the dynamic vary that in lots of circumstances makes it lose high quality as an alternative of gaining it.

Microsoft DirectML


It was the final to reach as such and it’s no much less essential for that. Unlike the 2 rivals, it has extra “open” traits, and its implementation requires solely Windows 10, the sport or the motive force having nothing to do with the efficiency or optimization of stated expertise.

In addition, and though it’s a Full Frame expertise, it has the drawback that there is no such thing as a recreation that helps it, which is curious for the reason that builders must make minimal adjustments to assist it.

Also, you might be curious that will probably be appropriate with all GPUs that assist DX12, with out discriminating fashions or architectures, which is a superb benefit and may very well be postulated because the benchmark within the sector.

AMD FidelityFX Super Resolution


It is the final to reach, to the purpose the place it stays a thriller to this present day, however we do know a number of issues. In the primary place, will probably be a expertise that might be applied in every recreation that helps it, which leaves this sort of implementation within the arms of the builders.

It would be the similar as DLSS a Full Frame expertise and though we have no idea what architectures it should assist, it’s greater than probably that will probably be RDNA ahead. But what units it other than Radeon Boost? Well, FidelityFX SR does makes use of Deep Learning algorithms and it’ll require graphics {hardware} to maneuver it, however this level is unclear.

There is a number of controversy right here, since as we have no idea how AMD goes to maneuver the {hardware} sources, it’s speculating that FidelityFX SR may very well be based mostly indirectly on Microsoft DirectML, however all the time inside the software program package deal that receives its title: FidelityFX.

This ends in that at this time we have no idea if AMD is following in NVIDIA’s footsteps, if it shares a path with Microsoft or if it undertakes its personal. Everything signifies that will probably be the final and that the corporate should assign software program engineers in order that the builders know the right way to implement it accurately of their video games, since if it doesn’t occur to NVIDIA when DLSS comes out, which was a somewhat infamous failure.

So which one is healthier?

Well, it’s an unanswered query in case you have been attentive in studying. The solely actuality are DLSS 2.Zero and Radeon Boost, however the latter doesn’t use Deep studying and solely change the decision by driver.

DirectML could be very inexperienced and Microsoft assist tasks appear to be going slowly however absolutely, though in the meanwhile there’s nothing available in the market that’s of the stature. For its half, NVIDIA got here first, struck first with DLSS, and now has a mature expertise that works and to work on. Developers are more and more implementing their expertise in engines, so for now it’s the firm and expertise that’s working essentially the most.

We will see if FidelityFX SR can shade it …