Filter

×
Active Filters Clear All
Keyword: 异构计算 ×
8 Total Reports
Intel Other High Signal 2026-04-08

Intel and SambaNova Announce Heterogeneous Inference Architecture for Agentic AI

Intel and SambaNova have announced a collaborative blueprint for Agentic AI production workloads. The heterogeneous design combines GPUs, SambaNova RDUs, and Intel Xeon 6 processors to address performance, efficiency, and software compatibility issues, with availability expected in H2 2026.

AMD Other High Signal 2026-03-19

AMD Defines Agent Computer Vision for Edge AI Architecture

AMD releases 2026 AI PC roadmap, proposing Agent Computer concept with expanded Ryzen AI stack featuring NPU-GPU-CPU heterogeneous architecture. Enables local multimodal AI agents, shifting PC from productivity tool to proactive AI partner.

AMD Other Medium Signal 2026-03-19

AMD Highlights CPU's Critical Role in Agentic AI Orchestration and Inference

AMD states Agentic AI workloads require serial decision-making and context management, better suited for CPUs. The company emphasizes high-core-count, high-memory-bandwidth server CPUs will lead in agent orchestration and lightweight inference, complementing GPUs in training. This signals a strategic repositioning of CPUs in AI data center architecture.

AMD Other Medium Signal 2026-03-05

AMD Releases Complete ROCm Technical Documentation to Strengthen AI Development Ecosystem

AMD released comprehensive ROCm technical documentation covering installation, system optimization, and performance tuning guides, with specialized optimization for MI300X GPUs. The documentation supports multiple programming models including HIP and OpenCL, improving GPU utilization efficiency for AI/HPC workloads.

AMD Other Medium Signal 2026-03-04

AMD Releases Updated Device Models for FPGA and Adaptive SoC Verification Acceleration

AMD released updated device models for core product lines including Versal ACAP and UltraScale+, enhancing pre-silicon simulation accuracy and toolchain compatibility. The update aims to accelerate hardware design verification and reduce development risks for complex systems like AI and data centers.

AMD Other Medium Signal 2026-03-04

AMD Launches Vitis Unified Platform for AI and HPC Development

AMD launches Vitis unified software platform, simplifying FPGA and adaptive SoC development through high-level programming models. The platform integrates optimized libraries for AI inference and data analytics, supports mainstream AI frameworks, and provides performance analysis tools. This lowers the barrier for heterogeneous computing development.

AMD Other Medium Signal 2026-03-02

AMD Launches Vitis AI Developer Tools to Strengthen AI Inference Ecosystem

AMD releases Vitis AI developer tool suite, providing a unified AI development environment for its adaptive computing platforms. The tools support mainstream deep learning frameworks and offer model optimization, quantization, and compilation capabilities to lower deployment barriers for AI models on AMD hardware.

NVIDIA Other 1970-01-01

NVIDIA Acquires Groq LPU: Inference Architecture Shift from HBM to On-Chip SRAM

NVIDIA signs ~$20B licensing deal with Groq for LPU tech, featuring 230MB on-chip SRAM at 80TB/s bandwidth. This targets Transformer inference decode, replacing HBM bottlenecks with ultra-low latency on-chip storage, potentially reshaping the AI inference chip landscape.