graphics processing unit
PulseAugur coverage of graphics processing unit — every cluster mentioning graphics processing unit across labs, papers, and developer communities, ranked by signal.
- competes with central processing unit 70%
- used by H.1000 Gnome 70%
- competes with application-specific integrated circuit 70%
- used by Dohuk Polytechnic University 70%
- used by Vulkan 70%
- uses data processing unit 70%
- competes with Cerebras Systems 70%
- competes with Tensor Processing Unit 60%
- uses central processing unit 50%
- used by central processing unit 50%
- used by Cerebras Systems 50%
- competes with data processing unit 50%
18 天有情绪数据
-
Hierarchical Transformer Preconditioner speeds up physics simulations
Researchers have developed a Hierarchical Transformer Preconditioner designed to improve the efficiency of real-time physics simulations. This new method utilizes a multiscale structural prior derived from an H-matrix p…
-
Godot Engine Tech Lead Discusses AI's Practical Role and GPU Needs
An interview with Clay John, technical lead at the Godot Engine, focused on the open-source philosophy and rapid development cycles of the game engine. John highlighted Godot's emphasis on practical rendering features a…
-
Nscale secures $790M for Norway AI data center amid energy scramble
Nscale, an AI infrastructure developer, has secured $790 million in financing for its data center campus in Narvik, Norway. The deal, backed by several Nordic and European banks, signals a shift towards treating AI infr…
-
LoKA framework enables low-precision FP8 for large recommendation models
Researchers have developed LoKA, a framework designed to make low-precision arithmetic, specifically FP8, practical for large recommendation models (LRMs). Unlike previous attempts that often degraded model quality, LoK…
-
CPUs are sufficient for most AI tasks, offering cost savings
Most AI applications do not require a GPU and can perform optimally using CPU infrastructure. This approach can be more cost-effective for businesses. The article provides guidance on how to integrate AI into applicatio…
-
Cerebras Systems boosts IPO on AI compute demand
Cerebras Systems is significantly increasing its IPO price and share count due to high demand driven by the AI industry's need for compute power. While GPUs, particularly from Nvidia, have dominated AI workloads like tr…
-
Sakana AI, NVIDIA unveil TwELL for faster LLM training and inference
Researchers from Sakana AI and NVIDIA have developed TwELL, a novel method that significantly speeds up large language model (LLM) operations. By targeting the feedforward layers, which are computationally intensive, Tw…
-
AI models increasingly run on-device, reducing service reliance
The shift towards running AI models locally on devices is a positive development, moving away from a reliance on "LLM as a Service" models. While the necessary hardware, such as GPUs, remains costly, there is an expecta…
-
China's CPI Rises 1.2% in April; Stock Market Opens Higher
China's National Bureau of Statistics reported that the Consumer Price Index (CPI) rose by 1.2% year-on-year in April, with a 0.3% increase month-on-month. Concurrently, the Producer Price Index (PPI) saw a 2.8% year-on…
-
A-share Market Surges, Shanghai Index Breaks 4200 Amidst Tech Gains
The A-share market saw a broad increase, with the Shanghai Composite Index surpassing the 4200-point mark. Key sectors leading the gains included engineering machinery and semiconductors, while precious metals and spiri…
-
Companies Lay Off Staff to Fund AI Investments Amid Rising Costs
Companies are continuing to lay off employees to fund their investments in AI, according to a report. The primary driver for these job cuts is not automation replacing workers, but rather the escalating costs associated…
-
Commentator calls AI boom a 'giant con' reliant on hyperscalers
Tech commentator Ed Zitron argues that the current AI boom, particularly for companies like OpenAI and Anthropic, is an unsustainable "con" propped up by hyperscalers. He believes this reliance on massive infrastructure…
-
GPU Memory Bandwidth Crucial for Local LLM Speed, Outpacing VRAM
For running large language models locally, GPU memory bandwidth is a more critical factor than VRAM capacity. Higher bandwidth allows the GPU to process data more quickly, preventing it from being bottlenecked while wai…
-
New GPU solver cuRegOT accelerates optimal transport for machine learning
Researchers have developed cuRegOT, a new GPU-accelerated solver designed to overcome the computational challenges of optimal transport (OT) in large-scale machine learning applications. The solver addresses the limitat…
-
Mac mini outperforms expensive workstations running large AI models
A $1,999 Mac mini equipped with Apple Silicon can run a 70-billion parameter AI model, outperforming a $4,000 Windows workstation. This is attributed to Apple's unified memory architecture, which eliminates VRAM and PCI…
-
Kunluncore files for dual IPO, touts China's first 32K GPU AI cluster
Kunluncore, an AI chip spinoff from Baidu, has officially filed for an IPO on Shanghai's STAR Market, alongside a concurrent filing for a Hong Kong listing on January 1st. The company announced its P800 GPU cluster, fea…
-
HCInfer system enables LLMs on resource-constrained devices with error compensation
Researchers have developed HCInfer, a novel inference system designed to enable large language models (LLMs) to run efficiently on devices with limited memory. This system offloads parts of the model's compensation mech…
-
Rongxin Zhiyuan raises hundreds of millions for GPU-centric AI architecture
Rongxin Zhiyuan, an AI infrastructure company founded by Tsinghua University alumni, has secured hundreds of millions of yuan in an angel funding round. The company is developing its novel AGC architecture, which positi…
-
Galaxy Securities: Token consumption to surge, benefiting AIDC, telcos, fiber optics, and optical modules
Galaxy Securities predicts a significant increase in Token consumption, driven by the growing demand for AI inference and rapid iteration of large language models. This surge is expected to accelerate growth across four…
-
AWS offers EC2 Capacity Blocks for short-term GPU needs
Amazon Web Services (AWS) is introducing EC2 Capacity Blocks for Machine Learning (ML) and SageMaker training plans to address the scarcity of GPU capacity. These new options allow customers to secure short-term GPU res…