1. 31 May, 2017 4 commits
  2. 30 May, 2017 7 commits
  3. 26 May, 2017 7 commits
  4. 25 May, 2017 3 commits
  5. 24 May, 2017 6 commits
    • Johann's avatar
      subpel variance neon: reduce stack usage · f3c97ed3
      Johann authored
      Unlike x86, arm does not impose additional alignment restrictions on
      vector loads. For incoming values to the first pass, it uses vld1_u32()
      which typically does impose a 4 byte alignment. However, as the first
      pass operates on user-supplied values we must prepare for unaligned
      values anyway (and have, see mem_neon.h).
      But for the local temporary values there is no stride and the load will
      use vld1_u8 which does not require 4 byte alignment.
      There are 3 temporary structures. In the C, one is uint16_t. The arm
      saturates between passes but still passes tests. If this becomes an
      issue new functions will be needed.
      Change-Id: I3c9d4701bfeb14b77c783d0164608e621bfecfb1
    • Johann's avatar
      Use vdup instead of vmov · d204c4bf
      Johann authored
      Change-Id: Idb6248c1429b55176bb3e9f4e8365ea0ed2be62a
    • Johann Koenig's avatar
      Merge changes Iaab2b9a1,Idfb458d3 · de1a9c77
      Johann Koenig authored
      * changes:
        sub pel avg variance neon: 4x block sizes
        sub pel variance neon: 4x block sizes
    • Johann Koenig's avatar
      Merge changes I31fa6ef8,I228c6f29 · b11a37f5
      Johann Koenig authored
      * changes:
        sub pel avg variance neon: add neon optimizations
        sub pel variance neon: normalize variable names
    • James Zern's avatar
    • James Zern's avatar
      partial_idct_test,InitInput: fix rollover in mult · 566f6d75
      James Zern authored
      promote coeff to signed 64-bit to avoid exceeding integer bounds when
      squaring the value
      Change-Id: If77bef6bc0a6a4c39ca3013e5e2ddb426a1c6e1f
  6. 23 May, 2017 3 commits
  7. 22 May, 2017 10 commits