-
- UC San Diego To Build Cyberinfrastructure for NASA ResearchThe San Diego Supercomputer Center (SDSC) and UC San Diego’s Scripps Institution of Oceanography have been awarded a NASA grant to...
- Old Supercomputers Never Die, They Just Get RepurposedInstead of retiring its five-year-old Gordon supercomputer, the San Diego Supercomputer Center (SDSC) has struck a deal with a New York-based...
- Intel Fuels Wave Of Training To Stave Off Data Scientist ShortageFocus on AI, Machine Learning application development yields base for next generation of data scientists The shortage of data scientists that...
- New Intel Hybrid Cloud Datacenter Security Aims at “Shadow IT” UsersThe increasing complexity of datacenter infrastructures and the challenge of protecting them is shifting InfoSec professionals away from best-of-breed tools in...
- Supercomputing Expected To Help Increase Life Expectancy By A DecadeSupercomputers will give rise to advances in personalized medicine, which will improve healthcare and potentially increase people’s life expectancy by five...
- Supercomputer Models Speed New Materials DiscoveryMachine Learning helps confirm properties of new “nanomaterials” 10x+ faster Researchers using supercomputers are taking advantage of machine-learning algorithms to accurately...
Blogroll
Case Studies
Evaluation Guides
Parallel Universe Magazine
Research Papers
Tutorials
Videos
White Papers
- UC San Diego To Build Cyberinfrastructure for NASA Research
-
Most Recent Verify Posts RSS
Profiling Python Code with Intel Vtune Amplifier
Even experienced Python programmers fall into coding traps once in a while. To help find out what’s really going on, watch this video where Intel’s Lead Technical Consulting Engineer Kevin O’Leary provides a hands-on demo showing how to use Intel VTune Amplifier to find hot spots in Python code that …
Machine Learning is Just Like it Sounds
Underpinnings of automation enables functions thought impossible just a few short years ago. Though it sounds quite vague, machine learning (ML) is actually what it says. It lets computers acquire the ability to learn without solutions being explicitly programmed. From a programmer’s point of view, it’s a little bit Meta. …
Object-Oriented FORTRAN (as Seen by Other Languages)
Though FORTRAN has existed since the late 1950s, it’s only since Fortran 2003 that proper object-oriented programming became possible. Having said that, Fortran 90 introduced modules, and there was a concerted effort to write code in an object-oriented type way. See this, for example. In this article I’ll touch on …
Read Full Post
Posted in Verify
Tagged C#, C++, FORTRAN, Java, Parallelization, Vectorization
Leave a comment
Using MKL for Fast Vector Math
Last semester I taught a 4000-level computer graphics class. The entire class focused around 3D objects, their animation, and their interaction. Fortunately everyone in the class had taken Calculus 3, Differential Equations and Linear Algebra. The reason this is important is because the math to manipulate 3D objects is crazy …
Making Sense of Parallel Sorts
Sorting is a common procedure in computer science. I spend a week teaching the most common sort of algorithms to my computer science class. I could spend an entire semester teaching all of the sorts in existence since there are so many. But in most situations, there are six common …
Which is Better: Parallelization or Vectorization?
Parallelization or Vectorization? It’s a slightly odd question, as ideally you would want both. But sometimes, you have to make a choice. Parallelization is where your application uses background threads to do the heavy lifting. Vectorization is where the special vector registers that are typically 128 or 256 bits long …
Modernizing Code for Tomorrow’s HPC Problem Solving
Code modernization–and the development of future software–must meet the changing demands of an economy that’s increasingly based on combining modeling, simulation, and virtual prototyping with large volumes of unstructured data and the Internet. This article shares tips on code modernization that are proven valuable for dedicated HPC software developers, domain …
When Performance Tuning an App, Do You Have the Right Tools?
To fully utilize today’s hardware, you must modernize your software using software development tools and libraries. This article in the newest Intel Parallel Universe Magazine recommends a methodology for getting the most performance out of your hardware though threading, vectorization, your memory hierarchy, and the tuning of cross-node parallelism using …
Read Full Post
Posted in Verify
Tagged Cluster, MPI, Parallelism, Threading, Vectorization
Leave a comment
Using Intel Trace Analyzer and Collector to Tune MPI
Most of my blogs here at Go Parallel are about parallelization. This includes techniques of employing multiple cores and processors, and taking advantage of processor vectorization. Parallelization is not new; the well-publicized case of Big Blue beating Kasparov in 1997 using parallelization among other techniques underscores its longevity. Since so …
Read Full Post
Posted in Verify
Tagged ITAC, MPI, OpenMP, Parallel Studio, Vectorization
Leave a comment
A New Tool for Vectorization Advice
This article in the newest Intel Parallel Universe Magazine demonstrates how to use the Intel Advisor XE tool on real code examples to optimize vector codes. Intel Advisor XE combines dynamic analysis, static binary analysis and compiler reports with recommendations for fixing performance bottlenecks. Read More >>