Recent content by DavidGraham

They are not going to change anything, companies like TSMC and Samsung are not agile as you think, they rely on fabs and complicated machinery from other companies (such as ASML) to do their business, these facilities and equipment take years to build and set up, and right now they are set up to...

This chart is showing the effect of crypto bubbles, increased demand during the bubble causing a spike, then a slump after the bubble pops, then a return to normalcy. There is also the effect of COVID. Here are the latest results from Q4 2023, showing return to normal demand. Total AIB...

Ok I might have had the wrong impression from your post. You said many things. But still, pricing people out will mean decrease in production and the gradual withdrawal from the market. Which isn't financially valid. Why would they do that when they can sell all chips? Expensive GPUs to AI...

They can sell both at the same time, and make even more money. That's how businesses work. That's not entirely true, only when you compare against Crypto bubbles, but exclude those bubbles, and the number is relatively stable each year. That's not true, even in the hypothetical situation of...

First of all, AI GPUs are NOT higher yields, they are chips sized at the reticle limit, which means they have much lower yeilds. Secondly NVIDIA can barely produce 2 to 3 million AI GPUs per year, that's across H100, H200, B100 and B200. That's because these chips need advanced packaging (for...

Disagree, native path tracing lacks reflections clarity, lacks dozens of shadows entirely, and have flickering unstable and laggy indirect lighting, It's simply no match.

Well, they have that AI money, why would they care if a 4080 costs 1200$ instead of 1000$? It's the argument people are making to incorrectly expect a 1500$ 5080.

However, 4x ML upscaling gives you 2x the performance of native for maybe ~10% reduction in certain image aspects, so ... It's a no brainer bargain really. If you factor in ML denoising (Ray Reconstruction) into the equation, then 4x upscaling exceeds native TAA -quality wise- in path traced...

If someone thinks that way about NVIDIA, then he/she probably doesn't understand NVIDIA well enough, they are an opportunistic and conquering entity, meaning they love to conquer new markets, for years they kept their Automotive segment on life support even though it brings them very little...

Ark Survival Evolved seems to use hardware Lumen, at native resolution (2K and 4K), the 3080 is as fast as or faster than 7900XTX. https://www.pcgameshardware.de/ARK-Survival-Evolved-Spiel-55571/Tests/Ascended-Tech-Check-Unreal-Engine-5-1445393/4/

What the hell is wrong with Take Two? I mean GTA V sold way too big, Red Dead Redemption 2 also sold big, GTA Online is a golden goose with billions in revenue from microtransactions. In 2016 (only 3 years after the release of GTA V), GTA Online alone made 500 million $ from microtransactions...

By the end of 2024, Microsoft is going double it's arsenal of AI GPUs to 1.8 million GPUs, mostly from NVIDIA. Which means buying around 900K of GPUs. Also by 2027, they plan to spend around a 100 billion dollars on GPUs and data centers...

While true, their critics and haters have successfully weaponized this against them, as the 5700XT lacks ray tracing, mesh shaders, DLSS, DLAA, DLDSR, ray reconstruction, Reflex and a myriad of other new features like RTX HDR, RTX Remix, RTX Video, RTX Chat .. etc. Pretty much anytime NVIDIA...

F1 2024.