Recent content by PhoBoChai

P
Discussion [HWUB] Nvidia has a driver overhead problem. . .

Oh man I just watched that vid from DF, its like they pretend like they don't know what's going on, all perplexing, why are Radeons getting 2x the FPS vs NV GPUs when the CPU is weak?!
- PhoBoChai
- Post #41
- Mar 19, 2021
- Forum: Graphics Cards
P
Discussion [HWUB] Nvidia has a driver overhead problem. . .

Mate you can see this clearly in some reviewers. Primary example, 3060 review by some of them, compared to the 1060, it's a GREAT upgrade path. 6700XT review from same ppl? Not compared to RX 480, or even 5700XT. Its only compared vs 3060Ti and 3070. The narrative is quite different. Then you...
- PhoBoChai
- Post #40
- Mar 19, 2021
- Forum: Graphics Cards
P
Question Speculation: RDNA2 + CDNA Architectures thread

There is nothing to suggest this. AMD's slides refer to the 1.94ghz as the fabric that ties the cache and GPU together, not the actual cache operating freq. AMD claims they used Ryzen L3 SRAM libraries for RDNA2 on 7nm, and we know Ryzen's L3 has no problems running >4.5ghz. There should be no...
- PhoBoChai
- Post #5,395
- Mar 19, 2021
- Forum: Graphics Cards
P
Discussion [HWUB] Nvidia has a driver overhead problem. . .

I think he's confused, because he talks about Async Compute, but this effect is seen in games that don't even use Async C. His results show DX12 is eating more CPU cycles on NV GPUs, and the only thing that runs between hw & CPU is the driver interaction with the API. GameGPU.RU today just did...
- PhoBoChai
- Post #34
- Mar 18, 2021
- Forum: Graphics Cards
P
Question AMD's answer to DLSS, FFSS. Mentioned in [GraphicallyChallenged][YT].

Doesn't make much difference in final perf. Compare DLSS perf in Turing vs Ampere in reviews and see for yourself. Because while it can run concurrently, it still has to share a very limited pool of registers as all the other shaders so you're bottlenecked by something else.
- PhoBoChai
- Post #18
- Feb 11, 2021
- Forum: Graphics Cards
P
Question Speculation: RDNA2 + CDNA Architectures thread

So, quite different, and optimized for DXR 1.1, like many ppl thought.
- PhoBoChai
- Post #4,877
- Nov 24, 2020
- Forum: Graphics Cards
P
Question Speculation: RDNA2 + CDNA Architectures thread

I think folks are not being logical when they say RT makes a huge difference for Minecraft or Quake RTX. It's because those games by default are so simplistic. If one were to enhance them with modern raster techniques for global illumination, dynamic lighting & shadows, SSR, and godrays etc, you...
- PhoBoChai
- Post #4,814
- Nov 22, 2020
- Forum: Graphics Cards
P
Question Speculation: RDNA2 + CDNA Architectures thread

Dirt 5, AMD leads with RT off, and with RT on. RT on vs off, the perf hit is only 20%, less than the perf hit of Ampere & Turing. It's RT shadows, just like RT Shadows in COD Black Ops, Warzone, or Shadow of the Tomb Raider and others. Calling it minimal because its just RT shadows ignore the...
- PhoBoChai
- Post #4,813
- Nov 22, 2020
- Forum: Graphics Cards
P
Question Speculation: RDNA2 + CDNA Architectures thread

That's correct, those numbers are accurate. It clearly shows 6800XT scaling with higher resolution, and at 4K it really shines vs 5700/XT. Its basically scaling as it should. Ampere is going down res it scales worse and worse for all of its tflops. Thats why NV hyped up 8K gaming. As for IC...
- PhoBoChai
- Post #4,793
- Nov 22, 2020
- Forum: Graphics Cards
P
Question Speculation: RDNA2 + CDNA Architectures thread

Correct. Computerbase.de just did a memory OC analysis for the 6800XT. They found almost no gains in perf from memory OC, which shows clearly the GPU isn't bandwidth limited. The IC does it's job really well across all resolutions.
- PhoBoChai
- Post #4,786
- Nov 21, 2020
- Forum: Graphics Cards
P
Question Speculation: RDNA2 + CDNA Architectures thread

More knowledgeable say that? Then they are plain wrong. They compare 4K vs 2x FP32 Ampere designs: But they don't actually say the correct thing, 3080 & 3090 only scales well for it's Tflops at 4K. Its quite inefficient perf/tflops at 1080p and 1440p. 6800XT is efficient at all resolutions...
- PhoBoChai
- Post #4,785
- Nov 21, 2020
- Forum: Graphics Cards
P
Question Speculation: RDNA2 + CDNA Architectures thread

Why do people say IC can't help much at 4K? 128MB hit rate at 4K is 58%. That's not just ~2.1x bandwidth multiplier on vram, it is a perf boost since cache on die access is about 10x faster than memory in terms of latency. If it reduces to 68MB for PS5, that should still have a decent hit...
- PhoBoChai
- Post #4,782
- Nov 21, 2020
- Forum: Graphics Cards
P
Question Speculation: RDNA2 + CDNA Architectures thread

There is no indication at 1440p is CPU limited, at all. Even 1080p is questionable given reviewers testing on 720p show many games still gain nice perf moving down to those resolutions. Facts are the 2x FP32 design on RTX Ampere is less utilized at lower resolution. Accept it and know that it...
- PhoBoChai
- Post #4,758
- Nov 21, 2020
- Forum: Graphics Cards
P
Question Speculation: RDNA2 + CDNA Architectures thread

Both companies do it. When they work with devs, they optimize for their own architectures and when they differ, its going to hurt the competitor. I would never go to the absurd length to pretend like AMD good, NV bad. Both are profit companies. I just point out how absurd some of the ppl are...
- PhoBoChai
- Post #4,757
- Nov 21, 2020
- Forum: Graphics Cards
P
Question Speculation: RDNA2 + CDNA Architectures thread

Are people seriously arguing over a 5% delta at 4K as if it matters? A SINGLE new game that skews heavily to NV or AMD will change that picture entirely. It's not as if the difference is massive to start it, please folks. There's custom 6800XT shown to OC on the air cooler to 2.65Ghz, and with...
- PhoBoChai
- Post #4,752
- Nov 21, 2020
- Forum: Graphics Cards

RESOURCES

Top Bottom