The World Opinion

Tag: NVIDIA

  • Samsung Electronics Signals Progress In HBM Chip Supply To Nvidia |

    Seoul: Samsung Electronics on Thursday hinted at the possibility of supplying its advanced high bandwidth memory (HBM) chips to the US AI giant Nvidia in the near future.

    The South Korean tech giant has been struggling to have its HBM3E chips pass Nvidia’s quality tests, while its local chip making rival SK hynix recently began mass-producing industry-leading 12-layer HBM3E chips, Yonhap news agency reported.

    “We are currently mass-producing both eight-layer and 12-layer HBM3E products,” said Kim Jae-june, vice president for memory business at Samsung Electronics, during a conference call following the release of its third-quarter earnings report.

    Kim noted that the company has achieved “meaningful progress” in meeting quality testing requirements for a “major customer,” expected to be Nvidia, whose GPUs are essential for AI computing.

    “We expect to expand sales in the fourth quarter,” he added, addressing previous concerns that Samsung Electronics had fallen behind in supplying HBM products to Nvidia.

    Samsung Electronics reported that its third-quarter HBM sales rose more than 70 per cent from the previous quarter, with the fifth-generation HBM3E chips expected to contribute 50 per cent of total HBM sales in the fourth quarter.

    “We are expanding sales of our eight- and 12-layer HBM3E chips to multiple customers,” the company said. “We are working on improving our HBM3E in line with the next-generation GPU plans of a major customer.”

    Samsung Electronics said it is developing sixth-generation HBM4 products and aims to begin mass production in the second half of next year.

    In its third-quarter earnings report, Samsung Electronics’ semiconductor division posted a 3.86 trillion won ($2.8 billion) operating profit, missing the market expectation of 4.2 trillion won.

    Shares of Samsung Electronics rose 0.17 per cent to close at 59,200 won on the Seoul main bourse, outperforming the KOSPI’s 1.45 per cent drop.

  • ChatGPT-maker OpenAI Appoints Political Veteran Chris Lehane As Head of Global Policy: Reports |

    New Delhi: ChatGPT-maker OpenAI has appointed political veteran Chris Lehane, who is a member of the executive team at OpenAI, as its vice president of global policy, according to New York Times reports. Chris Lehane, a former policy chief at Airbnb and a member of the Clinton White House, could not be reached for comment by OpenAI when contacted by Reuters.

    This appointment comes after the Cupertino-based tech giant Apple and chip giant Nvidia are reportedly in talks to invest in OpenAI as part of a new fundraising round that could value the Microsoft-backed startup above $100 billion. Earlier in the day, the Financial Times reported that OpenAI is considering restructuring its corporate framework to be more appealing to investors.

    OpenAI names political veteran Lehane as head of global policy, NYT reports https://t.co/BfChnf7ybw pic.twitter.com/XWGX45QbHG — Reuters (@Reuters) August 31, 2024

    Lehane has also served in the Clinton White House as a lawyer and spokesperson, specializing in opposition research. He earned a reputation as “the master of disaster” during his time working for President Bill Clinton.

    A spokeswoman for OpenAI, Liz Bourgeois, said, “Just as the company is making changes in other areas of the business to scale the impact of various teams as we enter this next chapter, we recently made changes to our global affairs organization.”

  • Nvidia Faces Lawsuit Over AI Copyright Infringement By Authors |

    New Delhi: Nvidia, a leading provider of chips for artificial intelligence, finds itself embroiled in legal action as three authors Brian Keene, Abdi Nazemian, and Stewart O'Nan have filed a lawsuit against the company. The authors alleged that Nvidia used their copyrighted books without permission in the development of its NeMo platform, sparking a contentious legal battle over intellectual property rights.

    The authors argue that their books were part of a dataset of 196,640 books used to train Nvidia's NeMo AI platform. They claim that the platform aimed to replicate everyday written language but was shut down in October following allegations of copyright infringement. The legal document suggests that Nvidia's action to remove the NeMo dataset implies acknowledgment of copyright violation. (Also Read: Google Wallet Will Auto-Add Movie Tickets And Boarding Passes From Gmail)

    The authors are seeking unspecified compensation on behalf of people in the United States whose copyrighted works were used to train NeMo's large language models during the past three years, in the proposed class action. (Also Read: Xiaomi 14 Ultra vs Samsung Galaxy S24 Ultra: Which Smartphone Should You Buy?)

    The lawsuit covers various types of literary works, such as Keene's novel “Ghost Walk” from 2008, Nazemian's novel “Like a Love Story” from 2019, and O'Nan's novella “Last Night at the Lobster” from 2007.

    Nvidia chose not to provide a comment on Sunday, and lawyers representing the authors did not respond immediately to requests for further comment on Sunday. This lawsuit pulls Nvidia into a larger pool of legal disputes initiated by writers, as well as The New York Times, concerning generative AI. Generative AI generates new content using inputs like text, images, and sounds.

    Other tech giants, including OpenAI, the developer of the AI ​​platform ChatGPT, and its collaborator Microsoft, are also facing legal issues linked to generative AI.

    Despite facing legal hurdles, Nvidia continues to attract investors' favor, benefiting from the ongoing growth of AI technology. Since the close of 2022, the company's stock price has surged by nearly 600%, propelling Nvidia's market capitalization to nearly $2.2 trillion.

    The lawsuit is identified as Nazemian et al v Nvidia Corp in the US District Court for the Northern District of California, under case number 24-01454.

  • Why NVIDIA’s previous RTX GPUs may not be getting the brand new DLSS 3.0 function

    GPUs are crucial a part of any PC, and choosing the proper one could be a difficult determination. In case you are taking a look at NVIDIA GPUs, you might be almost definitely questioning whether or not DLSS 3.0 is an to be had function. DLSS 3.0 is the most recent model of the generation, and it is just to be had on GPUs from the RTX 4000 sequence. So, if in case you have an older NVIDIA GPU, good fortune is not truly for your facet.

    DLSS is a generation that makes use of the Tensor Cores present in NVIDIA GPUs to support sport efficiency whilst keeping up symbol high quality. It’s an AI-based generation that upscales in-game pictures, permitting video games to run at upper resolutions and framerates.

    NVIDIA’s Ada GPUs will be offering important enhancements over the corporate’s earlier era of GPUs, together with new structure and strengthen for NVIDIA’s Deep Studying Tremendous Sampling 3.0 generation. Alternatively, NVIDIA has no plans to provide strengthen for its older era of RTX GPUs.

    What’s DLSS 3?

    A ground-breaking development in AI-powered graphics, DLSS 3 dramatically will increase sport fluidity whilst keeping very good visible high quality and responsiveness. Development on Tremendous Answer, DLSS accommodates NVIDIA Reflex low latency generation for the most productive responsiveness and Optical Multi Body Technology to provide utterly new frames. The fourth-generation NVIDIA Ada Lovelace structure, which additionally powers the GeForce RTX 40 Collection graphics playing cards, is a supply of energy for tremendous sampling.

    The Body Technology Convolutional Autoencoder makes use of two consecutive in-game frames and Ada’s Optical Glide Accelerator. The optical drift box data the path and fee of motion of pixels between frames 1 and a pair of. Debris, reflections, shadows, lights, and different pixel-level main points that don’t seem to be taken into consideration via movement vector computations in sport engines could also be noticed via the optical drift Accelerator. Optical drift fields are calculated the use of two consecutive frames. The present and former sport frames and engine information corresponding to movement vectors and intensity are the 4 inputs.

    The DLSS Body Technology AI community selects the best way to use information from sequential sport frames, optical drift fields, and sport engine information like movement vectors to build intermediate frames for each and every pixel. DLSS 3 moreover uses a sport engine movement vector to correctly practice the motion of geometry within the scene, by contrast to the Optical Glide Accelerator, which data pixel-level results corresponding to reflections. Body era can exactly rebuild geometry and results via using each engine movement vectors and optical drift to trace movement.

    Why may not DLSS 3 be supported within the earlier era RTX GPUs?

    Since DLSS Body Technology runs as a post-process at the GPU, DLSS3 does no longer serve as on prior era playing cards. The upgraded optical drift analyzers at the RTX 4000 sequence playing cards are gentle years forward of the prior era’s fashions. Even if older era playing cards strengthen OFA, they are no longer as fast or efficient as Ada.

    Even if last-generation RTX playing cards may strengthen DLSS 3, NVIDIA has selected to discontinue strengthen for them as a result of they imagine the DLSS they are able to reach will likely be laggy and feature a decrease total high quality than what may well be accomplished on RTX4000. This determination can have contributed to a basic loss of hobby in DLSS.