Post has attachment

Here are some of the factors to inform an archiving strategy for your HPC system. It’s vital to consider the benefits and drawbacks of both tape and disks as storage options. http://bit.ly/2rLM7Ju

Post has attachment

Post has attachment
For enterprises searching for the right data platform, IBM Spectrum Scale, formerly known as GPFS, offers several key advantages, thanks in part to a number of updates made in recent years. Here’s our review. http://bit.ly/2ocgYQJ

Post has attachment

Hi, dear members of HPC,
I have a question.
So I am working on a project which requires HPC cluster to run a simulation. But my code is not always right (a lot of times in fact), so I need to debug the C++ program on cluster.
While when I run the program using PBS job, I only get very limited error information including the core dump file. And sometimes I prefer to use Totalview or other GUI based debugger with xwin to analyze the code. But the HPC will kill my process after it runs a while on front end. The problem is my program runs a long until an error comes out.
So I am looking for some practical ways to improve this debugging process, any suggestion would be appreciated!

Post has attachment
CMake is an excellent, open source option for software development and testing. Here’s guidance on how to use it to manage your software builds and improve portability. http://bit.ly/2jUbNOw

Post has attachment
The new 3D XPoint from Intel and Micron is here. And this high-speed, non-volatile storage solution could reshape the HPC computing world as we know it today. http://bit.ly/2fEt7JL

Post has attachment
We've made performance testing of pre-release version of Tempesta FW, an open-source application delivery controller (ADC) built into Linux. The benchnarks show that the system reaches 1.8M HTTP requests per second on cheap 4 cores CPU. Details of the performance testing are available at our Wiki: https://github.com/tempesta-tech/tempesta/wiki/Tempesta-FW-benchmark .

Post has attachment
High-performance computing environments require ongoing, sophisticated performance monitoring to detect the kinds of problems that degrade performance in subtle ways. http://bit.ly/2fmuxoO
Wait while more posts are being loaded