Accessing LAPACK from the MKL library

Many user programs reference the LAPACK (Linear Algebra PACKage) library. The ARC software library does not include a standalone compiled version of this library. However, a highly tuned version of the LAPACK library is included in the ATLAS library (along with the BLAS, FFT functions, and many others). The optimized versions of LAPACK functions available through the MKL library typically offer better performance than a user could get from a source code library of LAPACK. For details, refer to https://secure.hosting.vt.edu/www.arc.vt.edu/userguide/lapack_mkl/ .

Accessing LAPACK from the ATLAS library

Many user programs reference the LAPACK (Linear Algebra PACKage) library. The ARC software library does not include a standalone compiled version of this library. However, some of the most useful LAPACK functions are available in the ATLAS library, which contains automatically tuned versions of those functions, as well as every BLAS function. Automatic tuning typically offers better performance than a user could get from a source code library of LAPACK. For details, refer to https://secure.hosting.vt.edu/www.arc.vt.edu/userguide/lapack_atlas/ .

Accessing BLAS functions in the MKL Library

Many user programs reference the BLAS (Basic Linear Algebra Subprograms) library. The ARC software library does not include a standalone compiled version of this library. However, the BLAS are available in the Intel MKL library, which contains automatically tuned versions of every BLAS function, offering better performance than a user could get from a source code library of BLAS. For details, refer to https://secure.hosting.vt.edu/www.arc.vt.edu/userguide/blas_mkl/ .

Accessing BLAS functions by using ATLAS

Many user programs reference the BLAS (Basic Linear Algebra Subprograms) library. The ARC software library does not include a standalone compiled version of this library. However, the BLAS are available in the ATLAS library, which contains automatically tuned versions of every BLAS function, offering better performance than a user could get from a source code library of BLAS. For details, refer to https://secure.hosting.vt.edu/www.arc.vt.edu/userguide/blas_atlas/ .

ARC System Access Requires Two Factor Authorization

All users need two factor authorization (2FA) to access VT ARC systems.

Details about setting up and using 2FA are in:

https://secure.hosting.vt.edu/www.arc.vt.edu/2fa/

If you are using the DUO app on a cell phone for your 2FA, then you
can click on the key next to “VT”, and a code will pop up. Then, when
you log into any 2FA system, you can follow your password by a comma and
then the code:

      password,code
    

This alternative is useful for situations where cell phone service is
an issue, or for dealing with systems with a fast timeout, such as ETX.

P100 GPU Nodes added to NewRiver

ARC is happy to announce the addition of 39 new GPU nodes to the NewRiver cluster. Each of these nodes is equipped with two Intel Xeon E5-2680v4 (Broadwell) 2.4GHz GPU (28 cores/node in all), 512 GB memory, and two NVIDIA P100 GPUs. Each GPU is capable of up to 4.7 TeraFLOPS of double-precision performance, so including CPU and GPU these nodes add over 400 TFLOPS of peak double-precision throughput to ARC’s resources.
Continue reading P100 GPU Nodes added to NewRiver