Intel Parallel Studio XE 2017 更新重點

New in this release:

  • All components updated to current versions
  • Addition of Japanese localization for online tutorials and documentation
  • General improvements to all components in terms of stability, performance and bug fixes for latest Intel® Architectures
  • Default installation includes components for 32-bit targets on Linux*
  • Intel® Cluster Checker:
    • Added Intel® Scalable System Framework support
    • Added additional support for Intel® Xeon Phi™ Product Family x200 processors
    • Removed heartbeat functionality
  • Intel® MPI Library:
    • PMI-2 support for SLURM*
    • Deprecating support for cross-OS launches, and for DAPL, TMI, and OFA fabrics
  • Intel® Trace Analyzer and Collector:
    • Introduced mouse wheel zooming support for timelines
    • MPI Performance Snapshot adds a new diagram Node-to-Node Data Transfers
    • MPI Performance Snapshot adds support for non-MPI applications
  • Intel® Advisor: Extended recommendations for virtual methods in vectorized loops
  • Intel® Data Analytics Acceleration Library:
    • Added distributed neural network training
    • Added KNN algorithm for batch computing mode
    • Added min-max normalization
  • Intel® Math Kernel Library:
    • Added support of non-square cores of convolution
    • Improved performance of ?GETRF, ?GETRS and ?GETRI for very small matrices via MKL_DIRECT_CALL
    • Improved single thread SGEMM/DGEMM performance on Intel® Advanced Vector Extensions 2 (Intel® AVX2), Intel® Advanced Vector Extensions 512 (Intel® AVX-512), and Intel® Xeon® for Intel® Many Integrated Core Architecture
  • Intel® Integrated Performance Primitives:
    • Added functions for the finite field GF(p) arithmetic, and the elliptic curves over the finite field GF(p)
    • Added ippsECCPBindGxyTblStd functions that allow to control memory size for the elliptic curves over GF(p)
  • Intel® VTune™ Amplifier XE:
    • Support for locator hardware event metrics for the General Exploration analysis results in the Source/Assembly view that enable you to filter the data by a metric of interest and identify performance-critical code lines/instructions
    • Summary view of the General Exploration analysis extended to explicitly display measure for the hardware metrics: Clockticks vs. Pipeline Slots
    • Command line summary report for the HPC Performance Characterization analysis extended to show metrics for CPU, Memory, and FPU performance aspects including issue descriptions for metrics that exceed the predefined threshold