Amd sponsored games

If rdna2 cards have performance advantage, biggest gap is at p. If I remember tests right. But it's nothing if it's AMD title which messes with Nvidia cards badly so we can sometimes even see being ahead of From DF's video going from memory here it seemed like when RT was much lighter or much heavier than rasterisation, AMD came closer to nvidia in RT performance as measured by extra ms as added to RT, like you're doing, so the light case isn't simply explained by less RT , while when raster and RT time were more similar, nvidia pulled ahead.

I've got two hypotheses as to why that might be the case. Though note that these are completely unsupported by any sort of actual profiling. First up is nvidia's ray traversal being a dedicated fixed function unit. With light ray tracing, it's mostly idle while shaders are busy. With super heavy ray tracing, it's the bottleneck and shaders aren't fully utilised. With a more balanced mix, both are utilised well.

AMD, using shaders for traversal, wouldn't see this utilisation imbalance. Though the lack of fixed function acceleration does still mean they're behind. Second theory was infinity cache. A workload more balanced between RT and rasterisation will mean more contention over the limited cache size, while if it's nearly all raster or RT hits would be more likely.

I think one of the lighter examples was for GI might even have been RE8. It does make sense from a top down perspective. The ability to do a lot of RT work concurrently is a big advantage for Nvidia, however it's definitely true that if you pass the limit of the amount of RT the card can do then it will choke on that workload and have the shaders waiting.

On the flip side, the fact that AMD can't do it concurrently is hurting them both in general performance as well as stability, frametimes and bugs, its not exactly an easy task to manage. As in my other post, I have both a TI and a XT and the XT sits in it's box, RT on it has definitely got better over time but it's still a mess and not something enjoyable to use on it and also happens to perform around the same as the Ti in many current RT games. So uh, hi old friend, how much would it take to part from your xt, I couldn't care less about RT and enjoy fluid fps plus gaming.

Are you taking offers? I've read enough reviews of the first-gen AMD cards with respect to raytracing and path tracing performance to know if that's what you're after, it's best to stick to Nvidia or wait till AMD's second generation by then Nvidia will be on its third, though. It's pretty sad, because competition is good. You never know. People doubted AMD would even catch Nvidia in raster performance, but it's safe to say that they not only caught up but surpassed Nvidia in some cases, basically anything below 4K is AMD territory.

Keep in mind Nvidia is currently hamstrung by the inferior Samsung "8nm" manufacturing process. Not that AMD didn't come up with some impressive innovations like Infinity Cache, but I would say Nvidia still has the architectural advantage. In quake 2 my overclocked and raised power limit xt maintains at least 45fps at p and works up to 60 in certain areas.

When nvidia released the 2k series my take on RT was that it was just a differentiator for them, a way to lock people into their hardware. But right now it just seems like an expensive bell or whistle. That's not a problem because it's open source. You can always create an AMD-optimized fork yourself. To be honest I do not really see the point or difference from a Gamer's perspective, some reflections become RT seems more like an 3D Animator's Filter setting that they would like and "appreciate" more than a Gamer like me.

As someone that has both perspectives available to him. I have a XT as well in the other pc. Not sure I see the point in testing the differences, should I?

That's the purest RTRT test on a functioning game you can do, there is no complex geometry to process, no big levels to load, it's all about samples and bounces, you won't have to guess, every card will have its hands full on pure RTRT. Quake II RTX was created by Nvidia engineers, they didn't spend time to optimize the code for AMD at all, they just changed the code from their specific implementation to vulkan extensions. I know it was developed by Nvidia unlike Q2VKPT which was made by Christof Schied, who is not an Nvidia employee, he's just done an internship there, he's a researcher as far as I understand , but unlike AAA games it doesn't occupy the card on heavy rasterization, all you get to see is the raw RTRT performance, hybrid games can have multiple bottlenecks that have nothing to do with RT computations, the same reason why I didn't cite Minecraft as, perhaps, loading the chunks can result in bottlenecks, Quake II is small and practically irrelevant in that aspect, which is why I believe it's the best candidate to see RT performance in a functioning game.

Plus, it is open source, you can't blame Nvidia if no one from AMD spent some time on it to optimize it for RDNA2, assuming there are optimizations to be done, the code is there on GitHub for everyone to tinker with and it's been almost a year since RDNA2 came out, I've seen some people proposing changes to fix the performance and bugs on AMD hardware and APanteleev seemed to be quite happy to implement them, so if someone had proposed other optimizations I'm sure they would've been integrated.

Thank you so much for breaking down the numbers from a real world perspective using an actual game. Of course, there are some drawbacks with RSR compared to FSR, but AMD must have weighed the positives and moves by competitors and figured out the change of tack was worth implementing. AMD provided the most obvious reasons why RSR will be inferior to the existing but not so widespread FSR when it discussed the technicalities of its scaling tech. The most important reason that FSR scaling tech needs to be implemented in-game mid-pipeline is so that common game UI and HUD elements don't get softened up by the spatial upscaling algorithm.

Even though this is likely to be a reliable tip, please remember to add a pinch of salt to all this RSR talk until AMD announces it. You shouldn't have to wait long until RSR becomes official, though - it is tipped to be launched next month. Using intelligent image scaling techniques to boost framerates looks like a new flashpoint in the war between AMD, Intel and Nvidia. Rating: Bad Not the best Average Good Eastcoasthandle Senior Member Posts: Joined: Astyanax Senior Member Posts: Joined: I can't take you seriously that you think it's amd driver problem.

Cyberpunk is not sponsored by nvidia. I can't take you seriously that you think it's amd driver problem.. If the game runs perfectly well when you're running The game's, or the driver's? I think people who are wise and experienced enough should be able to answer. If you still haven't figured it out, the problem goes away when you revert to the older I can't take your statements seriously now, can I?

A game having in-game bugs is not tied to having graphical issues due to driver changes. If you are that much into the game you are better off playing it based on it's sponsorship. I never bought that game for that reason. This is probably the most hilarious and at the same time probably the saddest statement I have read on this forum.

It's far worse than the time when some people came to AMD's defense when they said something to the effect of they will not fix issues with DirectX9 games, with saying things like if you're still playing DirectX9 games, you don't deserve support. I don't know what the age of a game has to do with providing proper driver support.

Thankfully, AMD backpedaled from that statement and went on to promise fixes for DX9 games, that they semi delivered. Think of this from a gamer's perspective, as someone who's paying for the hardware, why should I care about whom is sponsoring the game? You're entitled to having working, playable, stable and enjoyable performance for that title, regardless of who your hardware vendor is, be it nvidia or AMD.