Frontier sets fluid dynamics speed record, cutting simulation time from 38.5 to 1.5 hours on AMD GPUs
Using GPU cores instead of CPUs provides a massive boost to computational speed
Zeus GPU from startup Bolt Graphics promises to run circles around the competition
Vaporware... or should Nvidia be worried?
DeepSeek open-sources file system, claims it runs AI models faster and more efficiently
What just happened? In response to Western organizations calling it "shady and untrustworthy," DeepSeek launched "Open Source Week." During last week's event, the company released several repositories to the open-source community, including a highly efficient file system. Many AI specialists reviewing the code have come away impressed.
AMD-powered supercomputer El Capitan takes the global performance crown
Do It Super: AMD's resurgence in the chip industry is influencing the specialized high-performance computing market as well. A few years after unveiling the world's fastest supercomputer, the US chipmaker is once again making headlines by bringing another HPC powerhouse online.
Elon Musk's supercomputer with 100,000 Nvidia GPUs uses proprietary Spectrum-X networking platform
All these GPUs need a proper switching technology for lag-free AI hallucinations
Fujitsu introduces dynamic resource allocator for AI servers and HPC systems
The Japanese giant thinks it has a working solution to the global GPU shortage as well
Japanese supercomputer Fugaku is now a software stack for virtual HPC instances on AWS
Super Cloud: The Fugaku supercomputer was at the forefront of high-performance computing development just a few years ago. Now, the Japanese technology has been turned into a software stack that could "democratize" supercomputing applications around the world.
GPUs poised for major performance enhancements under preliminary HBM4 specs
AI applications in particular stand to benefit
US military can now remotely access powerful supercomputer resources in the cloud
Scientific research and analysis today, HPC electronic warfare tomorrow?
Intel says its optical interconnect chiplet technology is a milestone in high-speed data transmission
It can already support 4 terabits per second bidirectional data transfer across 100 meters
Cheyenne supercomputer sold to the highest bidder for $480,000
The super-system has been sold "as is," with water leaks, packaged cable bits, and everything else
Ultra Ethernet gains momentum as tech giants join for AI and HPC network innovation
Newcomers to the group include IBM, Nokia, Dell, Baidu, Huawei, Lenovo, Supermicro, and Tencent
India to build a 10,000 GPU supercomputer for self-sufficient AI development
A wide-reaching initiative to foster New Delhi's tech sovereignty in the new AI era
Nvidia is reportedly investing $30 billion in new custom chip unit
In context: The booming AI sector has propelled Nvidia's financials to new heights, as its AI acceleration GPUs have become highly sought after. However, these general-purpose components are facing increasing competition, prompting the company to develop more efficient chips tailored for specific clients and tasks.
Nvidia H100 AI GPUs might cost 3 to 6 times as much as AMD's MI300s
Enormous price difference seen between AMD and Nvidia data center GPUs
Europe first exaFLOPS supercomputer should fire up in less than a year
Modular construction will speed up contract completion and allow easy future upgrades
AMD says its MI300X AI accelerator is faster than Nvidia's H100
Two GPU companies fighting over each other's benchmark results
China's new Tianhe supercomputer doubles performance with homegrown tech
Beijing is reportedly joining the exascale club without Western chip technology
Sunway SW26010-Pro is China's most powerful supercomputer chip to date
A lot of computing power hindered by poor memory interface choices?
Google GraphCast AI can predict weather conditions faster than conventional methods
Machine learning won't replace supercomputer-based forecasting anytime soon, though
Fujitsu claims new tech can optimize CPU & GPU allocation for AI and HPC workloads
Working on a practical solution for the global GPU shortage, allegedly
Socionext is working with TSMC on a 2nm Arm chiplet for data centers
A power-optimized Arm CPU for HPC and hyperscale applications
Europe's first exascale supercomputer will use Arm processors with Nvidia GPU nodes
Arm power efficiency is good enough for next-gen HPC systems, it seems
US government is renting out GPU power at half price for a limited time
TL;DR: Private companies are in desperate need for additional GPU computing capabilities to train their new generative AI services, yet they are encountering significant challenges in obtaining them. In contrast, researchers involved in US supercomputing projects currently have access to potent Nvidia GPU nodes at a substantial discount, albeit for a limited time.
Nvidia generates up to 1,000% profit for each H100 GPU sold
Hopper-based HPC accelerator truly is a money-printing machine