06-14-2017 12:09 AM
Check the post to learn how to join the Beta program, and the provide your feedback.
What's New in Intel® MKL 2018 Beta Update 1:
- Addressed an early release buffer issue in threaded *GEMV
- Improved TBB *GEMM performance for small m and n while k is large
- Improved performance on Intel® Xeon Phi™ processor x200 (formerly Knights Landing)
- Improved convolution performance for Intel(R) Xeon Phi™ processors based on Intel® Advanced Vector Extensions 512 (Intel® AVX-512) with support of AVX512_4FMAPS and AVX512_4VNNIW instruction groups
- Improved performance of 3D FFT complex-to-real and real-to-complex scaled and non-scaled problems on Intel® Xeon Phi™ processor 72** (formerly Knights Landing)
- Improved performance of 2D FFT complex-to-complex problems with scale on Intel® Xeon Phi™ processor 72** (formerly Knights Landing) and on Intel® Xeon processors E3-xxxx V5 (formerly SkyLake)
- Completed alignment with Netlib* LAPACK 3.7.0 by integrating all new routines and bug fixes. Notable new features are:
- Optimized factorization, solve and inverse routines with rook pivoting: ?sytrf_rk/?hetrf_rk, ?sytrs_rk/?hetrs_rk and ?sytri_3/?hetri_3
- Added LAPACKE interfaces for all new routines including Aasen factorization and solve routines
- Improved ?gesvd performance for tall-and-skinny matrices.
- Improved ?gelq/?gemlq performance for short-and-wide matrices.
- Improved performance of vdPowx and vsPowx functions for certain exponent values (0,5,6,7,8,9 )