The Next Platform - The Next Platform
WEBThe Next Platform covers the latest trends and developments in artificial intelligence, high-performance computing, and data center technologies. Read stories on Dell, Nvidia, VMware, IBM, and more.
First In-Depth Look at Google’s TPU Architecture - The Next Platform
WEBApr 5, 2017 · Learn how Google designed its Tensor Processing Unit (TPU) for efficient and high-performance inference of neural networks. See the details of the TPU's matrix multiplication unit, systolic data flow engine, and memory architecture, and compare it with CPUs and GPUs.
Inside Amazon’s Graviton3 Arm Server Processor - The Next Platform
WEBJan 4, 2022 · Learn how AWS built its own Arm server chip, Graviton3, to improve performance, security, and innovation for its cloud services. Compare Graviton3 with other Arm processors and X86 CPUs and see how it powers AWS's PaaS and SaaS offerings.
This Is What The Most Powerful Server In The World Looks Like
WEBSep 8, 2021 · The Power E1080 is a 16-socket NUMA system based on IBM's Power10 chip, which has 240 cores and 128 threads per socket. It is designed for big data, analytics, and high-performance computing workloads that can benefit from its flat NUMA topology and fast memory access.
Intel Finally Serious About Switching with Barefoot Networks Buy
WEBJun 11, 2019 · In one fell swoop, Intel has finally filled a giant hole in its switching product line by acquiring upstart Barefoot Networks, the creator of the P4 programming language for networking devices and the “Tofino” family of Ethernet switch ASICs that make use of it. Now, perhaps, Broadcom will finally and forcefully get the kind of competition ...
AMD on Why Chiplets—And Why Now - The Next Platform
WEBJun 9, 2021 · Learn how AMD adopted a chiplet approach to overcome the challenges of Moore's Law and deliver high-performance processors across markets. Explore the technical, cost, and interconnect tradeoffs of the chiplet design and its evolution.
Inside Tesla’s Innovative And Homegrown “Dojo” AI Supercomputer
WEBAug 31, 2022 · Tesla's Dojo is a homegrown AI supercomputer that uses a custom processor, interconnect, and software stack for scalability and performance. The D1 core has a RISC-V-like instruction set with ML-specific primitives and a distributed SRAM memory architecture.
Deep Dive On Google’s Exascale TPUv4 AI Systems - The Next …
WEBOct 11, 2022 · Learn about the architecture, performance, and applications of Google's fourth generation Tensor Processing Units, which are custom AI accelerators for training and inference. Compare the TPUv4 with previous generations and with Nvidia GPUs, and find out when TPUv5 and TPUv5i are expected to launch.
Stacking Up Intel Gaudi Against Nvidia GPUs For AI - The Next …
WEBJun 13, 2024 · An Nvidia H100 system configured as we outlined above probably costs somewhere around $375,000, which is $46,875 per petaflops. The Gaudi 3 system with the same configuration would run around $275,000, at a cost of $18,733 per petaflops. That is 2.5X better price/performance than the Nvidia system.
Inside Facebook’s Future Rack And Microserver Iron - The Next …
WEBMay 14, 2020 · Learn how Facebook designs and deploys its own servers, racks, and accelerator modules for various workloads, including inference, video encoding, and training. See the details of the updated Yosemite V2.5 chassis and the Glacier Point V2 carrier card that support M.2 inference sticks.
Skip to content