POSTS

tld image.jpg

FUGAKU

2020 world's fastest supercomputer, made by Fujitsu and based on 2.2 GHz A64FX 48C ARM cores. With 7,299,072 cores, 4.8 PB memory, offers 416 PFLOPS and consumes around 28 MW of electric power. It runs Red Hat Enterprise Linux, and involved heavily COVID-19 related medical research

tld image.jpg

SOFTWARE-DEFINED VEHICLES

The automakers giant Mercedes-Benz and the semiconductor giant Nvidia inked a deal to power all the vehicles starting 2024 with standard Nvidia Drive AGX Orin technology, sensors and software. They will jointly develop AI-powered SAE level 2 and 3 automated vehicle applications, such as address-to-address automated driving, as well as automated parking functions up to level 4. Vehicles will receive updates on the air

tld image.jpg

RASPBERRY 4 SBC

Broadcom BCM2711, Quad core Cortex-A72 (ARM v8) 64-bit SoC @ 1.5GHz, 8GB LPDDR4-3200 SDRAM, Besides the ports listed: 2-lane MIPI DSI display port, 2-lane MIPI CSI camera port, 4-pole stereo audio and composite video port and Micro-SD card slot for loading operating system and data storage. The board is only 88mm x 58mm x 19.5mm, 46g, and 15W max

tld image.jpg

NVIDIA JETSON XAVIER

Smaller than a credit card (70x45 mm), the energy-efficient Jetson Xavier NX module delivers server-class performance of 21 TOPS @ 15 W. It can run multiple modern neural networks in parallel and process data from multiple high-resolution sensors which is a requirement for full AI systems

tld image.jpg

24/7 CARBON NEUTRAL

Google's 24/7 carbon-free energy Data Center. The carbon-intelligent computing platform moved the timing of non-urgent tasks like YouTube video processing, adding new words to Google translate and creating new Google Photo filter features to when wind and solar energy are most plentiful

tld image.jpg

SINGULARITY

Industrial revolution started with the steam engine, then the electricity, then the computer, and AI is expected to be the next big thing. The paradigm shift is called the singularity; referring to the point when the technological progress of computers and AI, with big data as key element, surpasses human intelligence; predicted to occur in 2045 with exponential technological progress has tremendous impacts on society, leading to irreversible changes in daily living

tld image.jpg

80-CORE ARM

Ampere's Altra ARM is a 7nm, 80-core, single threaded, 8 memory channels (4 TB/Socket), 128 PCIe lanes processor. With less than 3W/Core, it meets the Data Centers demand on low power high core count

tld image.jpg

CHIP SCALE ATOMIC CLOCK

A chip scale, low power (115 mW) and low cost ($1,500) atomic clock, to be used in portable devices that require precise timing and have no access  via GPS, like deep-sea diving, mining and seismic research. This clock has a good accuracy in the range of 1 second per 4,000 years. compared to heavy power consuming space taking atomic clocks with accuracy in the range of 1 second per 100 million years. Let alone the nuclear clock with accuracy in the range of 1 second per 10 billions years or the quantum logic with accuracy in the range of 1 second per 200 billions years. Quartz or Crystal clocks have accuracy of 1 second per 2 days

tld image.jpg

HELIUM HDD

Helium, as opposed to Oxygen & Nitrogen of the air, causes less flutter, hence platters can run faster and packed closer. Helium filled hard disk drives are hence more compact (25% less height), or have more platters (7 instead of 5) and hence larger capacity (50% more), quicker and faster, and consume less power (50% less) due to the lesser friction. Today's Helium filled 3.5" HDDs offer 12 TB, and hermetically sealed during manufacturing

tld image.jpg

UNDERWATER DATA CENTERS

Microsoft recently sank a data center the size of a shipping container in the Orkney Islands of Scotland. Underwater data centers get cheap power and free cooling. A data center on land takes between18 and 24 months to complete, while the portable underwater data centers can be delivered in 3 months, and serve for 5 to 10 years. Underwater data centers can run autonomously, scaled easily, and they are environment friendly