NVIDIA
NVIDIA is an American semiconductor and technology company primarily designing and selling graphics processing units (GPUs), system-on-chip processors, and related software platforms. Its products are widely used in gaming graphics rendering, data center computing, artificial intelligence training and inference, autonomous driving, and professional visualization. NVIDIA also developed the CUDA parallel computing platform, enabling GPUs for general-purpose scientific computation, and is a principal supplier of modern AI infrastructure.
Core Facts
TensorRT is a high-performance inference optimization engine provided by NVIDIA that compiles PyTorch/ONNX models into GPU-optimized execution engines, typically improving inference speed by 2–5×.
The NVIDIA A100's HBM2e capacity is 80GB, the H100 has 80GB HBM3, and the H200 has 141GB.
The HBM3 memory used by NVIDIA's H100 has a bandwidth of up to 3.35 TB/s.
Timeline (last 90 days)
WGMMA（Warpgroup Matrix Multiply-Accumulate）是 NVIDIA Hopper 架构引入的异步矩阵乘加指令，由整个 warpgroup（通常 128 个线程）协同执行，可直接从共享内存读取操作数并写入张量核心累加器。
Although the PTX specification is public, the correspondence between its semantics and the hardware microarchitecture still depends heavily on NVIDIA's implementation details, which is the fundamental reason why translation layers struggle to reproduce performance perfectly.
NVIDIA's moat is not just the software API, but also the deep co-optimization between hardware and compiler; reverse engineering can make CUDA programs "run" on non-NVIDIA hardware, but achieving "fast and stable" operation is exponentially harder.
The same PTX can run on NVIDIA GPUs of different architectures, providing a compatibility layer across GPU generations.
NVIDIA announced that it will fully open-source its Run:ai platform, as part of the NVIDIA DSX AI Factory platform strategy.
Currently the officially disclosed information is concentrated at the level of open-sourcing the complete platform; the specific code repository structure, license type, and community governance model have not yet been elaborated in detail.
Run:ai was initially known for helping enterprises efficiently schedule GPU resources on Kubernetes, and after being acquired by NVIDIA its positioning shifted toward being the hub of the AI factory.
When NVIDIA opens up the scheduling layer for free, the value proposition of third-party GPU schedulers needs to be rethought.
The deep coupling of Run:ai with NVIDIA hardware and the DSX platform means that the freedom brought by the open-source code will noticeably diminish when stepping outside the NVIDIA ecosystem.
The goal of the NVIDIA DSX AI Factory platform is to provide an end-to-end solution from hardware to scheduling to model deployment.
40 more timeline events
All Facts (20)
TensorRT is a high-performance inference optimization engine provided by NVIDIA that compiles PyTorch/ONNX models into GPU-optimized execution engines, typically improving inference speed by 2–5×.
90%VerifiedThe NVIDIA A100's HBM2e capacity is 80GB, the H100 has 80GB HBM3, and the H200 has 141GB.
90%VerifiedThe HBM3 memory used by NVIDIA's H100 has a bandwidth of up to 3.35 TB/s.
90%VerifiedNVFP4 is a 4-bit floating-point (FP4) quantization format introduced by NVIDIA, designed specifically for the hardware acceleration capabilities of Blackwell architecture GPUs
90%VerifiedA single NVIDIA A100 card has 80GB HBM2e memory and 312 TFLOPS of FP16 computing power
90%VerifiedNVIDIA CUDA was launched in 2006 and was originally a programming framework for enabling developers to use GPUs for general-purpose computing.
90%VerifiedThe NVIDIA H100 single-card FP64 double-precision floating-point performance reaches approximately 34 TFLOPS.
90%VerifiedxAI built a supercomputing cluster called Colossus, reportedly equipped with more than 100,000 NVIDIA H100 GPUs.
90%VerifiedNVIDIA NIM (Inference Microservices) is an inference microservice platform launched by NVIDIA that packages optimized AI models as standardized container images, supporting one-click deployment on enterprise on-premises GPUs.
90%Verified训练一次GPT-4级别的模型需要数万块A100/H100 GPU协同工作数月
90%VerifiedThe retail price of a single NVIDIA H100/H200 GPU is as high as $30,000–$40,000.
90%VerifiedCUDA (Compute Unified Device Architecture) is a parallel computing platform launched by NVIDIA.
90%VerifiedNVIDIA's H100/H200 series GPUs hold more than 80% of the AI training and inference market.
90%VerifiedROCm is AMD's open-source software platform for high-performance computing and AI workloads, including the HIP programming interface, which allows CUDA code to be ported to AMD GPUs at relatively low cost.
90%VerifiedNVIDIA's CUDA ecosystem is the de facto standard in the AI inference industry.
85%VerifiedNVIDIA's TensorRT-LLM deeply optimizes inference speed for its hardware architecture.
85%VerifiedThe HBM memory bandwidth of the NVIDIA H100 is approximately 3.35 TB/s, and its utilization rate in autoregressive generation tasks is often below 10%.
85%VerifiedThe H100 GPU adopts NVIDIA's Hopper architecture, with a single-card peak compute of 3,958 TFLOPS (FP16 precision).
85%VerifiedThe H100 SXM5's HBM3 bandwidth is approximately 3.35 TB/s, and its FP16 peak compute is approximately 989 TFLOPS.
85%Verified美国自2022年起对中国实施先进半导体出口管制，限制高端AI训练芯片如NVIDIA A100/H100及其后续产品的出口
85%