US DoE (Argonne) to acquire AMD+Nvidia supercomputer as testbed for delayed Intel-based exascale supercomputer

From Reuters: The Nvidia and AMD machine, to be called Polaris, will not be a replacement for the Intel-based Aurora machine slated for the Argonne National Lab near Chicago, which was poised to be the nation's fastest computer when announced in 2019.

Instead, Polaris, which will come online this year, will be a test machine for Argonne to start readying its software for the Intel machine, the people familiar with the matter said.

New Intel speculative execution vulnerability: "Microarchitectural Data Sampling" (MDS)

Another day, another hardware bug that has security implications. Like the recent Meltdown and Spectre bugs, this new bug called Microarchitectural Data Sampling (MDS) leaks data. Ars Technica has a nice write-up: "MDS attacks perform speculation based on a stale value from one of these [CPU] buffers."

Red Hat's more technical summary, and more detailed video explainers here and here.

Optimized zlib

I wasn't aware of optimized versions of zlib, the free patent-unencumbered compression library, until today. I ran across Juho Snellman's comparison benchmarks of vanilla zlib, CloudFlare zlib, Intel zlib, and zlib-ng. The upshot is that CloudFlare's optimizations seem to be the best performing. And its decompression times were 75% of the vanilla version, while Intel and zlib-ng ran at 98% and 99% respectively. This would be a clear win for read-intensive workflows, such as some bioinformatics workflows.

Read more about how CloudFlare was contacted by the Institute of Cancer Research in London to help improve zlib at this blog post by Vlad Krasnov. Intel has an overview of zlib in its Intel Performance Primitives (IPP) product. And this is zlib-ng's GitHub repo.

Disabling diagnostic remarks in Intel Compilers

So, I was just trying to compile R, a statistical analysis package, on RedHat EL 4 using Intel Compilers 11.1 + MKL 10.1. Unfortunately, the configure step kept barfing on a diagnostic message produced by the "-ipo" (interprocedural optimization) flag. The configure script was interpreting it as an error message when testing for certain features.

The online help for icc (the "-help" flag) does not show how to turn off IPO diagnostics. This Intel forum post gives the answer:

-diag-disable ipo

I will have more to say about compiling R with Intel and linking to the MKL later on.