P
Why cant AI write code for other hardware beside Nvidia GPU?
Chinese semiconductor thread II
- Thread starter vincent
- Start date
I think his point is that if AI can be trained to program in say SASS, then it could be trained to program in the SASS equivalent on other GPUs. CUDA is a moat because most human AI researchers know it by far the best, and it works really well for humans. If the programming language used does not need to work well for humans and does not need humans to be fluent with it, then there is no moat.
With that said, I think that AI being able to program in a language that humans don't really understand may have issues of its own.
ai can add enough comments to code for human consumption
Sure, but should AI learn to lie?
Anyhow, programming in machine language could be another frontier in AI coding. There's a lot of inefficiency in AI learning in higher level language which is ultimately translated into 0's and 1's that AI operates in, and then producing codes that again need to be compiled into 0's and 1's computers operate in. Make the whole process into 0's and 1's would probably improve efficiency by orders of magnitude.
in a sign of times, China is now clearly winning the 3rd generation power semiconductor battle. Wolfspeed continues to get less profitable and losing boatload of money. I think they are going down the drain
YMTC has been blacklisted since 2022. and they achieved this with domestic tools from AMEC+NAURA+Piotech and ASML's DUV..
According to these papers DSA could cut the immersion DUV mask use for contact interconnects by 75%
Influence of sidewall affinity on the directed self-assembly for contact hole multiplication
Abstract
Directed self-assembly of block copolymer (BCP) has been extensively explored in application of contact hole multiplication at the sub-7 nm technology node. The aim of this study is to investigate the effect of surface affinity on the directed self-assembly of BCP, polystyrene-b-poly (methyl methacrylate) (PS-b-PMMA) for contact hole multiplication. The sidewall affinity of the guiding templates for PS and PMMA blocks is manipulated by controlling the grafting process or the PS mole fraction (fPS) of the brushes, which is favorable for tuning the Center-to-Center Distance (CCD) of the twin-hole pattern. The result demonstrates that the CCD increases with increasing the sidewall affinity for PS block of BCP, which is promising to fabricate high-density hole patterns at the sub-7 nm technology node. |
Quadruple-hole multiplication by directed self-assembly of block copolymer
Abstract
Directed Self-Assembly (DSA) of cylindrical block copolymer with graphoepitaxy strategy offering significant potential for contact hole multiplication in semiconductor manufacturing at technology nodes below 7 nm (sub-7 nm). This technique allows precise control over the number of DSA-generated holes and their critical dimensions (CD) by manipulating the guiding template geometry and size. The results indicate that DSA quadruple-hole multiplication patterns aligned top-bottom and left-right, with a long-axis spacing of 55-65 nm and a short-axis spacing of 30-40 nm could be achieved through precisely designing template holes with long-axis dimension and short-axis dimension. The formed DSA pattern was successfully transferred to the underlying hard mask layer, creating a large-area quadruple-hole array with a CD of approximately 17 nm. A comprehensive investigation of guiding template size and morphology on multiplication hole patterning enhances the understanding of the self-assembly behavior in confined elliptical spaces. In conclusion, this work highlights the importance of optimizing guiding template size for contact hole multiplication in integrated circuit fabrication. |
