Code

hw: Add SIMD shuffling operations#191

Code

Draft

fischeti wants to merge 55 commits intomainpulp-platform/snitch_cluster:mainfrom shuffling-instrpulp-platform/snitch_cluster:shuffling-instrCopy head branch name to clipboard

+409-10

Commits

Commits on Aug 29, 2024

• doing tutorial

  

  sem24f2 Lucia Luzi (luzil)

  authored and

  fischeti

  committed
  64a517e

  Copy full SHA for 64a517e

• doing tutorial

  

  sem24f2 Lucia Luzi (luzil)

  authored and

  fischeti

  committed
  b03de3c

  Copy full SHA for b03de3c

• tutorial finished

  

  sem24f2 Lucia Luzi (luzil)

  authored and

  fischeti

  committed
  30d168c

  Copy full SHA for 30d168c

• Merge branch 'main' of https://github.com/pulp-platform/snitch_cluster

  

  sem24f2 Lucia Luzi (luzil)

  authored and

  fischeti

  committed
  067dc10

  Copy full SHA for 067dc10

• Updated submodule riscv-opcodes to use my own fork

  

  sem24f2 Lucia Luzi (luzil)

  authored and

  fischeti

  committed
  aabafcd

  Copy full SHA for aabafcd

• added branch to submodule riscv-opcodes

  

  sem24f2 Lucia Luzi (luzil)

  authored and

  fischeti

  committed
  ccdb552

  Copy full SHA for ccdb552

• Added shuffle to opcodes

  

  sem24f2 Lucia Luzi (luzil)

  authored and

  fischeti

  committed
  304bc31

  Copy full SHA for 304bc31

• draft for vfshuffle.s encoding and decoder

  

  sem24f2 Lucia Luzi (luzil)

  authored and

  fischeti

  committed
  edd79a4

  Copy full SHA for edd79a4

• Merge remote-tracking branch 'upstream/main'

  

  sem24f2 Lucia Luzi (luzil)

  authored and

  fischeti

  committed
  931c439

  Copy full SHA for 931c439

• working on decoder

  

  sem24f2 Lucia Luzi (luzil)

  authored and

  fischeti

  committed
  7a65650

  Copy full SHA for 7a65650

• testing the first implementation of vfshuffle.s

  

  sem24f2 Lucia Luzi (luzil)

  authored and

  fischeti

  committed
  1b399b5

  Copy full SHA for 1b399b5

• shuffle implementation

  

  sem24f2 Lucia Luzi (luzil)

  authored and

  fischeti

  committed
  60eaa8c

  Copy full SHA for 60eaa8c

• implementation for vfshuffle.s

  

  sem24f2 Lucia Luzi (luzil)

  authored and

  fischeti

  committed
  1a71ff0

  Copy full SHA for 1a71ff0

• traces: Add MiniFloat representation for RTL traces

  

  Viviane Potocnik

  authored and

  fischeti

  committed
  63af430

  Copy full SHA for 63af430

• gen_trace.py: Fix crash on unsupported SIMD instructions

  

  colluca

  authored and

  fischeti

  committed
  97191b0

  Copy full SHA for 97191b0

• gen_trace.py: Reverse SIMD vectors such that LSB is rightmost

  

  colluca

  authored and

  fischeti

  committed
  bb30748

  Copy full SHA for bb30748

• trying to run vfshuffle.h

  

  sem24f2 Lucia Luzi (luzil)

  authored and

  fischeti

  committed
  c2ddc52

  Copy full SHA for c2ddc52

• done with tests

  

  sem24f2 Lucia Luzi (luzil)

  authored and

  fischeti

  committed
  df8b23a

  Copy full SHA for df8b23a

• finished tests

  

  sem24f2 Lucia Luzi (luzil)

  authored and

  fischeti

  committed
  eeff5ee

  Copy full SHA for eeff5ee

• sw: Add transpose layer

  fischeti

  committed
  2d8abc0

  Copy full SHA for 2d8abc0

• transpose kernel

  

  sem24f2 Lucia Luzi (luzil)

  authored and

  fischeti

  committed
  74b0490

  Copy full SHA for 74b0490

• moved data scripts to /scripts

  

  sem24f2 Lucia Luzi (luzil)

  authored and

  fischeti

  committed
  1f23f6a

  Copy full SHA for 1f23f6a

• pulled fixes from upstream

  

  sem24f2 Lucia Luzi (luzil)

  authored and

  fischeti

  committed
  f8b0786

  Copy full SHA for f8b0786

• kernel done for fp32, fp16, fp8

  

  sem24f2 Lucia Luzi (luzil)

  authored and

  fischeti

  committed
  f727a63

  Copy full SHA for f727a63

• SSR without frep works

  

  sem24f2 Lucia Luzi (luzil)

  authored and

  fischeti

  committed
  8803037

  Copy full SHA for 8803037

• Remove files which should not have been tracked

  fischeti

  committed
  fb2d9d8

  Copy full SHA for fb2d9d8

• Remove C encoding instructions again

  fischeti

  committed
  bd1e9c2

  Copy full SHA for bd1e9c2

• Revert fp32_computation_vector test changes

  fischeti

  committed
  f6052d4

  Copy full SHA for f6052d4

• sw: Fix compilation errors

  fischeti

  committed
  b1cbd78

  Copy full SHA for b1cbd78

• util: Remove generate-opcodes comments

  fischeti

  committed
  8c50a89

  Copy full SHA for 8c50a89

• hw: Clean up shuffle unit

  fischeti

  committed
  716dd4e

  Copy full SHA for 716dd4e

• shfl_unit: Fix bitwidth of masks

  fischeti

  committed
  75fdb6c

  Copy full SHA for 75fdb6c

• wip: Add shuffle tests

  fischeti

  committed
  9289011

  Copy full SHA for 9289011

• REVERT: ignore tranpose_shuffle for now

  fischeti

  committed
  52779f2

  Copy full SHA for 52779f2

• hw: Combine two shuffle variants in decoder

  fischeti

  committed
  46132de

  Copy full SHA for 46132de

• wip: Configure cluster wrapper for shuffle instructions

  fischeti

  committed
  9c17280

  Copy full SHA for 9c17280

• sw: Fix debug target variables

  fischeti

  committed
  9cc3f1d

  Copy full SHA for 9cc3f1d

• util: Add shuffle instructions to format CSV

  fischeti

  committed
  25e8e45

  Copy full SHA for 25e8e45

• hw: Remove shuffle unit in FP SS

  fischeti

  committed
  64ca391

  Copy full SHA for 64ca391

• hw: Clean up

  fischeti

  committed
  f25a4a9

  Copy full SHA for f25a4a9

• hw: Remove shuffle unit

  fischeti

  committed
  9d4b740

  Copy full SHA for 9d4b740

• lint: Add missing license headers

  fischeti

  committed
  2db2d99

  Copy full SHA for 2db2d99

• sw: Whitespace changes

  fischeti

  committed
  b6f8411

  Copy full SHA for b6f8411

• lint: Python sources

  fischeti

  committed
  e0c7d01

  Copy full SHA for e0c7d01

• Fix format encoding

  fischeti

  committed
  3e0e8e9

  Copy full SHA for 3e0e8e9

• lint: Python sources

  fischeti

  committed
  938aa61

  Copy full SHA for 938aa61

• sw: Finalize shuffle test

  fischeti

  committed
  d03f26c

  Copy full SHA for d03f26c

• deps: Bump FPU

  fischeti

  committed
  f78afa6

  Copy full SHA for f78afa6

• Revert IIS setup script changes

  fischeti

  committed
  4e7ffa0

  Copy full SHA for 4e7ffa0

• Remove transpose shuffle stuff

  fischeti

  committed
  913b652

  Copy full SHA for 913b652

• Remove redundant include

  fischeti

  committed
  15697ff

  Copy full SHA for 15697ff

• Remove SW stuff

  fischeti

  committed
  7749def

  Copy full SHA for 7749def

• ci: Run shuffle test in all simulators

  fischeti

  committed
  8ec58c8

  Copy full SHA for 8ec58c8

• Bump opcodes submodule

  fischeti

  committed
  1c350d6

  Copy full SHA for 1c350d6

• Bump fpnew

  fischeti

  committed
  318fd67

  Copy full SHA for 318fd67