Commits

Branch selector

 master

User selector

All users

Datepicker

All time

Commit History

Commits on Apr 30, 2020

• Update README.md

  ebugger

  authored
  9734b9f

  Copy full SHA for 9734b9f

Commits on Apr 18, 2020

• Update README.md

  ebugger

  authored
  d53b278

  Copy full SHA for d53b278

Commits on Nov 8, 2019

• moved make_kernel() out of launchKernel()

  dwdoerf

  committed
  96c4bbb

  Copy full SHA for 96c4bbb

Commits on Aug 27, 2019

• finalized configuration files for odinson.lbl.gov (AMD Vega 10)

  dwdoerf

  committed
  a266e22

  Copy full SHA for a266e22

• Changed roll of ERT_TRIALS_MIN from minimum number of trials (i.e. for t=ERT_TRIALS_MIN, t<ntrials) to the minimum maximum number of trials (i.e. for t=1, t<ntrials, where ntrials >= ERT_TRIALS_MIN…

  Show description for 660b908
  dwdoerf

  committed
  660b908

  Copy full SHA for 660b908

• HIP version now uses HIP timers; OpenCL version tested on AMD Vega 10 GPU

  dwdoerf

  committed
  2cbd932

  Copy full SHA for 2cbd932

• fixed logic error in parsing ERT_OCL_SIZES

  dwdoerf

  committed
  508ab33

  Copy full SHA for 508ab33

Commits on Aug 24, 2019

• Updated OpenCL version to specify Global and Local sizes; Integrated OpenCL version into driver1.cxx; Added FP32 and FP16 (FP16 untested) support to kernel1.cl; Updated Python scripts accordingly

  dwdoerf

  committed
  6d23b7d

  Copy full SHA for 6d23b7d

• fixed bug that ignored FP types other than FP64

  dwdoerf

  committed
  2b6270e

  Copy full SHA for 2b6270e

Commits on Aug 23, 2019

• adding support for parsing fp32 and fp16

  dwdoerf

  committed
  915e0e9

  Copy full SHA for 915e0e9

Commits on Aug 9, 2019

• Merged HIP version into same source as CPU and CUDA versions

  dwdoerf

  committed
  10582ab

  Copy full SHA for 10582ab

• Merging OpenCL version, now compatible with new precision format

  dwdoerf

  committed
  465727a

  Copy full SHA for 465727a

Commits on Jul 20, 2019

• Code formatting

  PointKernel

  committed
  7a440e0

  Copy full SHA for 7a440e0

Commits on Jul 19, 2019

• Reorganize python scripts

  PointKernel

  committed
  e29aac8

  Copy full SHA for e29aac8

• Remove C files

  PointKernel

  committed
  8eb019e

  Copy full SHA for 8eb019e

• Minor correction & update README

  PointKernel

  committed
  d8eec18

  Copy full SHA for d8eec18

• Merge with master

  PointKernel

  committed
  0e13eb4

  Copy full SHA for 0e13eb4

Commits on Jul 18, 2019

• Add tensor core kernel - by using wmma

  PointKernel

  committed
  25e7ac5

  Copy full SHA for 25e7ac5

• Minor correction

  PointKernel

  committed
  40074b8

  Copy full SHA for 40074b8

• Update config files & README

  PointKernel

  committed
  a28126b

  Copy full SHA for a28126b

• Updates:

  Show description for 1249398
  PointKernel

  committed
  1249398

  Copy full SHA for 1249398

• Merge branch 'master' of github.com:PointKernel/cs-roofline-toolkit

  PointKernel

  committed
  6c4f2e8

  Copy full SHA for 6c4f2e8

• Add precision control flag

  PointKernel

  committed
  3e8d8df

  Copy full SHA for 3e8d8df

• Add gitignore file

  PointKernel

  committed
  4b23db8

  Copy full SHA for 4b23db8

• fixed error introduced by new parameter seetings

  dwdoerf

  committed
  d69957c

  Copy full SHA for d69957c

Commits on Jul 16, 2019

• Add config file for maeve

  PointKernel

  committed
  b25628b

  Copy full SHA for b25628b

• Minor update

  PointKernel

  committed
  76cf86c

  Copy full SHA for 76cf86c

• Updates: fix cpu bugs - low SP peak to investigate

  PointKernel

  committed
  548851b

  Copy full SHA for 548851b

Commits on Jul 10, 2019

• Adding HIP and OpenCL drivers/kernels. Also changed the kernel calculation to avoid floating point overflow

  dwdoerf

  committed
  7599646

  Copy full SHA for 7599646

Commits on Jun 27, 2019

• Use uint32_t in GPU kernels

  PointKernel

  committed
  bb2503e

  Copy full SHA for bb2503e

Commits on Jun 26, 2019

• Fix bugs

  PointKernel

  committed
  20a5d36

  Copy full SHA for 20a5d36

Commits on Jun 25, 2019

• Add half-precision peak

  PointKernel

  committed
  0e2f5c0

  Copy full SHA for 0e2f5c0

Commits on Jun 20, 2019

• adding example configuration for HIP usage using odinson.lbl.gov

  dwdoerf

  committed
  6ac90e4

  Copy full SHA for 6ac90e4

• Adding HIP driver and kernel

  dwdoerf

  committed
  193b75b

  Copy full SHA for 193b75b

• The input buffer initialization is now inside the working set size loop, this is done to ensure the calculations don't overflow to inifinity when executing multiple trials with large ERT_FLOP values

  dwdoerf

  committed
  96db902

  Copy full SHA for 96db902

Pagination

Previous

Next