Part of the Technology photoes in this website are created by rawpixel.com - www.freepik.com

H3 Platform / Blog / Recent

Recent

55 articles

GPU-Accelerated Storage
532

GPU-Accelerated Storage Fundamentals #1 — Why AI Needs a New Kind of Storage-- Extremely High Input-Output Per Second Storage

Modern AI workloads shift storage demands from bandwidth to high IOPS and concurrency. GPUs require continuous small, random data access (512B–8KB), making data delivery—not compute—the main bottleneck. Traditional CPU-centric storage cannot sustain the required parallelism, leading to underutilized GPUs. As a result, storage must evolve from capacity-focused (TB/TCO) to concurrency-driven (IOPS/TCO), becoming an active, high-performance data layer for AI systems.

Read More
AI Storage
5305

AI Storage Fundamentals #1 — Why AI Needs a New Kind of Memory/Storage

AI now bottlenecks on storage, not compute. GPUs demand fast, parallel, low-latency data delivery.

Read More
CXL
3020

Supercharging Your Server Efficiency? CXL Memory Sharing Might Be the Answer

Unlock stranded memory and redefine data center efficiency with H3’s CXL Memory Sharing—delivering scalable, high-performance, and composable infrastructure for next-gen AI and HPC workloads.

Read More
CXL
5965

How CXL Switch Interacts with Host CPUs — and the Configuration Required for Each CPU Type

CXL switch setup guide: BIOS and CLI configs for GNR & Turin, kernel parameter tuning, and DAX setup to enable reliable memory use.

Read More
GPU
264958

Which GPU Is The Best? RTX 5090, RTX PRO 6000 Blackwell, or RTX 4090?

Compare RTX 5090, RTX PRO 6000 Blackwell, RTX 4090, and RTX 6000 Ada performance in deep learning, rendering, and gaming to figure out the best value GPU in different fields.

Read More
CXL
3243

Breaking Through Memory Bottlenecks: 210 Million IOPS with Composable Memory Systems

Composable Memory Systems, powered by CXL, enable dynamic memory pooling to optimize resource utilization, reduce idle rates, and lower TCO in data centers. With up to 210M IOPS and 120GB/s bandwidth, it enhances AI, HPC, and cloud workloads by providing high-bandwidth, low-latency access. Seamlessly integrating with PCIe-based infrastructures, it ensures scalable, cost-efficient memory expansion without extensive hardware modifications.

Read More
CXL
7075

Breakthrough in CXL 2.0 Memory Pooling/Sharing Technology: Four Servers Share 2TB of Memory, Achieving 210M IOPS and 120GB/s Bandwidth

H3 Platform has developed a CXL 2.0 Memory Pooling/Sharing system, enabling four servers to share 2TB of CXL memory. Performance tests showed 66M IOPS and 43GB/s bandwidth per server, scaling to 210M IOPS and 120GB/s across four servers. This breakthrough enhances HPC, big data, and in-memory databases. H3 Platform will continue advancing CXL memory pooling and sharing solutions for data center innovation.

Read More
CXL
3332

Revolutionizing Data-Intensive Applications: Unleashing the Power of CXL Memory Chassis and FAMFS

This blog introduces H3 Platform's CXL 2.0 memory chassis and Micron's FAMFS, showcasing their revolutionary impact on data-intensive applications like RocksDB. It highlights the solution's enhanced memory capacity, performance stability, and compatibility, making it ideal for modern workloads in AI, machine learning, and databases.

Read More
CXL
6326

Exploring the Future of AI with CXL Memory Sharing Architecture

In the rapidly evolving AI landscape, traditional computing architectures struggle to meet the memory and processing demands of modern applications. CXL (Compute Express Link) Memory Sharing Architecture offers a scalable, efficient solution by allowing dynamic resource allocation. This architecture disaggregates traditional servers into nodes dedicated to compute, memory, or GPU resources, which are then logically composed into a unified system as needed. An example is the AI Cloud, which uses CXL memory for AI GPU servers connected via PCIe 5.0 x16 to an H3 CXL Memory Box with E3.S 2T x8 modules. This setup enhances performance through memory pooling and high-bandwidth connections, ensures scalability with flexible resource allocation, reduces costs by optimizing resource utilization, and simplifies management with centralized control and dynamic reconfiguration. As AI advances, CXL Memory Sharing Architecture is crucial for organizations to stay innovative and efficient.

Read More
CXL
1239

Highlights and Progress from Computex 2024: H3's Cutting-Edge Solutions in the Spotlight

H3 showcased innovative solutions like the CXL 2.0 Memory Sharing and PCIe Gen5 GPU at Computex 2024. Engaging a diverse audience, including cloud and AI users, they highlighted seamless integration tests with leading technologies. Looking forward, H3 is committed to advancing technology, fostering partnerships, and pushing the boundaries of computing capabilities.

Read More
CXL
4166

From Message Passing to Distributed Shared Memory: New Possibilities with CXL

With vertical scaling reaching its limits, horizontal scaling has become essential. Traditional message-passing (MP) models were used due to high remote communication costs in distributed shared memory (DSM) models. Compute Express Link (CXL) removes these costs, making DSM feasible. CXL enables high-speed, coherent data transfer, optimized memory pooling, and enhanced memory sharing, revitalizing DSM. While MP simplifies node communication, it involves data copying and complexities. DSM offers unified memory space but faces shared state management challenges. CXL advancements make DSM attractive for modern computing.

Read More
CXL
1501

Forging Ahead: H3 Pioneers CXL Memory Solution Integration within Sunfish™ Framework

Sunfish™ introduces an open-source platform, simplifying the development of composable distributed computing architectures. It streamlines interaction with disaggregated fabrics and computing resources, aiding system administrators, programmers, and architects in managing complex infrastructures. Key features include support for virtual platforms and clusters, RESTful interfaces, diverse system clients, Redfish and Swordfish language support, vendor-specific fabric support, and seamless interaction with high-speed fabrics. Integration with H3's CXL 2.0 Memory Pooling Solution enhances memory resource utilization, with progress including Sunfish agent setup, component information monitoring, and efficient resource allocation. Ongoing developments focus on chassis management functionalities. Stay updated as H3 continues to advance composable distributed computing architectures.

Read More