I'm interested in semiconductors, for example CPU, GPU, AI chips, NVIDIA...
Minisforum has launched the AtomMan G7 PT Mini PC, featuring an AMD Ryzen 9 7945HX CPU and an RX 7600M XT discrete GPU. The device is available in two versions: a complete system with 32GB of RAM and a 1TB NVMe SSD for $1,199, and a barebones version without components or an OS for $999. The early bird prices are $300 lower than the MSRP.
TechnologyMinisforum | Rating: 90 | 2024-07-10 07:26:29 PM |
Nvidia's Maxwell graphics cards, specifically the GeForce GTX 900 series and above, now support OBS's Twitch Enhanced Broadcasting feature. This feature was initially released for GeForce RTX graphics cards and has been updated to work with older Maxwell-based cards. To use this feature, streamers must be accepted into Twitch's BETA program and have the proper hardware.
TechnologyNvidia | Rating: 88 | 2024-07-10 05:26:39 PM |
Intel's upcoming CPU models, including Arrow Lake-S desktop processors and Lunar Lake mobile CPUs, have been confirmed through patch notes for CPU-Z version 2.10. The models include Core Ultra 9, Ultra 7, and Ultra 5 variants with unlocked 'K' SKUs and lower-end locked models. The Arrow Lake-S model names feature an unlocked 'K' SKU variant for each lineup, accompanied by a vanilla model with no locking scheme.
TechnologyIntel | Rating: 89 | 2024-07-10 03:36:33 PM |
Samsung Electronics has partnered with Preferred Networks to supply 2nm-class processors with HBM memory for datacenter AI accelerators. The processors will be developed by Gaonchips, a South Korea-based contract designer of chips, and packaged using Samsung's Interposer-Cube S 2.5D technology. The partnership aims to provide advanced AI accelerators for Preferred Networks' upcoming projects.
TechnologySamsung | Rating: 89 | 2024-07-10 12:36:27 PM |
Researchers at the U.S Department of Energy's Princeton Plasma Physics Laboratory (PPPL) are developing the next generation of chips using transition-metal dichalcogenide (TMD) to replace silicon in processors. Led by Shoaib Khalid, the team aims to create smaller, thinner, and more efficient chips. According to Moore's Law, the count of transistors in semiconductors will double every two years, but this rate is expected to slow down to three years as process nodes become smaller.
TechnologyEnergy | Rating: 86 | 2024-07-10 12:16:51 PM |
TSMC, the world's largest contract maker of chips, is planning to increase wafer pricing by up to 10% in 2025 due to high demand for advanced processors. The price hike will affect all types of customers, including AI and HPC customers like Nvidia. The 4nm and 5nm nodes, used by companies like AMD and Nvidia, are expected to see a price increase from around $18,000 to $20,000 per wafer.
TechnologyTSMC | Rating: 89 | 2024-07-10 10:46:54 AM |
TSMC, the world's No.1 contract maker of semiconductors, is on track to reach a market capitalization of $1 trillion. The company produces processors for leading chip designers, including Apple, AMD, Broadcom, Intel, Nvidia, and Qualcomm. TSMC's processors are in high demand due to their transistor density, performance, and power efficiency, which are unmatched by competitors Intel Foundry and Samsung Foundry.
TechnologyTSMC | Duplicated with: 1 | Rating: 90 | 2024-07-09 06:56:55 PM |
Nvidia has announced a bundle with Black Myth- Wukong, a game that will be released on August 19, with RTX 40-series GPUs ranging from the GeForce RTX 4090 to GeForce RTX 4070. The game will leverage path tracing and is a natural fit for mid-range to high-end RTX 40-series GPUs.
TechnologyNvidia | Rating: 88 | 2024-07-09 06:06:35 PM |
Intel's Arrow Lake-S desktop architecture will feature a neural processing unit (NPU) with 13 TOPS of AI performance, which is slightly more powerful than Meteor Lake's NPU. This is under the 45 TOPS of NPU-only performance of Lunar Lake and not enough to qualify for Copilot+ certification, which requires at least 40 TOPS of AI performance.
TechnologyIntel | Rating: 81 | 2024-07-09 03:46:48 PM |
Nvidia, Georgia Institute of Technology, and others have organized a contest to create an open-source public dataset to help large language models (LLMs) design hardware. The contest aims to address the shortage of high-quality data specific to hardware design, which is a bottleneck for LLM-Assisted Hardware Design. Nvidia's director of design automation research, Haoxing (Mark) Ren, announced the collaboration on X.
TechnologyNvidia | Rating: 89 | 2024-07-09 03:46:35 PM |