conferences/cineca2021.mdwn

   1 # Title:  An introduction to Libre-SOC and Cray-style Vectors for OpenPOWER
   2
   3 The current HPC / Supercomputer market has few top contenders:
   4 China ICT Loongson 3, Intel / AMD x86 and IBM POWER9/10.
   5 None of these have the features of one of the most iconic
   6 Supercomputer ISA: Cray-style Vectors, known historically
   7 for compact program size and massive number-crunching
   8 capability.
   9
  10 Libre-SOC is working under the close watchful eye of the OpenPOWER
  11 Foundation to develop Draft Extensions to the Power ISA, extending
  12 the already-proven Supercomputing-grade Power ISA with Cray-style
  13 Vectors.  This talk will give a brief overview of the progress and
  14 roadmap.
  15
  16 # Bio
  17
  18 Luke Kenneth Casson Leighton specialises in Libre Ethical Technology.
  19 He has been using, programming and reverse-engineering computing
  20 devices continuously for 44 years, has a BEng (Hons), ACGI, in
  21 Theory of Computing from Imperial College, and recently put that
  22 education to good use in the form of the Libre-SOC
  23 Project: an entirely Libre-Licensed 3D Hybrid CPU-VPU-GPU based on
  24 OpenPOWER. He writes poetry and has been developing a HEP Physics theory
  25 for the past 36 years in his spare time.
  26
  27 # Links
  28
  29 * <https://twitter.com/OpenPOWERorg/status/1408093712664567809?s=20>
  30 * <https://www.linkedin.com/posts/openpower-foundation_openpower-workshop-at-cineca-activity-6813861142486745088-yWD_>
  31 * June 30th 7.30 am EST to 10.30 am EST
  32 * <https://academy.cineca.it/en/events/openpower-workshop>
  33 * <https://www.hpc.cineca.it/center_news/openpower-workshop-cineca-june-30th-2021>
  34 * Video <https://www.youtube.com/watch?v=gexy0z1YqFY>
  35
  36 # Talk links
  37
  38 * SVP64 Overview <https://libre-soc.org/openpower/sv/overview/>
  39 * SIMD Considered harmful (massive understatement) <https://www.sigarch.org/simd-instructions-considered-harmful/>
  40 * Carnegie course on Vector Processors <https://course.ece.cmu.edu/~ece740/f13/lib/exe/fetch.php?media=seth-740-fall13-module5.1-simd-vector-gpu.pdf>
  41 * IAXPY AVX512 (quite shocking) <https://godbolt.org/z/f8a7PMPWc>
  42 * 250 lines of hand-crafted assembler for VSX strncpy <https://patchwork.ozlabs.org/project/glibc/patch/20200929152103.18564-1-rzinsly@linux.ibm.com/>
  43 * under 20 lines for Vectorised strncpy <https://github.com/plctlab/rvv-benchmark/blob/master/strncpy.s>
  44 * IAXPY for VSX (around 60 lines of assembler) <https://godbolt.org/z/4oGjTe8Ko>
  45 * FFMPEG MP3 code snippet inner loop <https://ffmpeg.org/doxygen/3.1/mpegaudiodsp__template_8c_source.html#l00121>
  46 * FFMPEG MP3 assembler, 450 lines <https://git.libre-soc.org/?p=openpower-isa.git;a=blob;f=media/audio/mp3/mp3_0_apply_window_float.s;hb=HEAD>
  47 * FFMPEG SVP64 MP3, under 100 lines <https://git.libre-soc.org/?p=openpower-isa.git;a=blob;f=media/audio/mp3/mp3_0_apply_window_float_basicsv.s;hb=HEAD>
  48 * Cooley Tukey FFT algorithm <https://en.wikipedia.org/wiki/Cooley%E2%80%93Tukey_FFT_algorithm#Data_reordering,_bit_reversal,_and_in-place_algorithms>
  49 * in-place FFT Butterfky <https://en.wikipedia.org/wiki/File:DIT-FFT-butterfly.png>
  50 * SVP64 bit-reverse LOAD <https://git.libre-soc.org/?p=openpower-isa.git;a=blob;f=src/openpower/decoder/isa/test_caller_svp64_ldst.py;hb=HEAD>
  51 * SVP64 twin +/- Vectorised FMAC <https://git.libre-soc.org/?p=openpower-isa.git;a=blob;f=src/openpower/decoder/isa/test_caller_svp64_fft.py;h=6cb2b522be85a2f86a0b505d1878dbcec645cb90;hb=8dfffc9c2ff7bb91715500160d1b057f9bef3ba0>
  52 * not part of the video, more info about REMAP [[openpower/sv/remap]]