rocm/tensile issues and pull requests

#100 - Assembly Generator

Pull Request - State: closed - Opened by guacamoleo over 7 years ago

#99 - Finer thread assignments

Issue - State: closed - Opened by guacamoleo over 7 years ago - 1 comment

#98 - Multi-dimensional tensor

Issue - State: closed - Opened by guacamoleo over 7 years ago - 1 comment
Labels: bug

#97 - Beta kernels need to check B==0

Issue - State: closed - Opened by guacamoleo over 7 years ago - 1 comment
Labels: bug

#96 - fixing GSU with prefetching

Pull Request - State: closed - Opened by guacamoleo over 7 years ago

#95 - changing LocalWrite to GlobalRead for ease, de-linting from pyflakes

Pull Request - State: closed - Opened by guacamoleo over 7 years ago - 1 comment

#94 - explicitly define work-group

Pull Request - State: closed - Opened by guacamoleo over 7 years ago

#93 - fixed recursive solution selection logic

Pull Request - State: closed - Opened by guacamoleo over 7 years ago

#92 - Semantic Versioning and GlobalSplitU

Pull Request - State: closed - Opened by guacamoleo over 7 years ago

#91 - v3.0.0 improving benchmarking thoroughness

Pull Request - State: closed - Opened by guacamoleo over 7 years ago

#90 - fixed bug for global read prefetch for NN, TT

Pull Request - State: closed - Opened by guacamoleo over 7 years ago

#89 - updated configs and DeepBench benchmarking script

Pull Request - State: closed - Opened by guacamoleo over 7 years ago

#88 - re-enabling large thread tile since register spilling bug no longer a…

Pull Request - State: closed - Opened by guacamoleo over 7 years ago

#87 - unroll mem_fence is configurable and off by default

Pull Request - State: closed - Opened by guacamoleo over 7 years ago

#86 - enabling Tensile on rocm-opencl

Pull Request - State: closed - Opened by guacamoleo over 7 years ago

#85 - short-vectors use verbose register initialization

Pull Request - State: closed - Opened by guacamoleo over 7 years ago

#84 - Enabling Half-precision

Pull Request - State: closed - Opened by guacamoleo over 7 years ago

#83 - kernel writer is abstract base class

Pull Request - State: closed - Opened by guacamoleo over 7 years ago

#82 - kernels with LoopUnroll<2 are invalid

Pull Request - State: closed - Opened by guacamoleo over 7 years ago

#81 - prefetching support

Pull Request - State: closed - Opened by guacamoleo over 7 years ago

#80 - Adding kernel timer capabilities

Pull Request - State: closed - Opened by kknox over 7 years ago

#79 - fixed global increments and updating configs

Pull Request - State: closed - Opened by guacamoleo over 7 years ago

#78 - Vector-Shifting complete

Pull Request - State: closed - Opened by guacamoleo over 7 years ago

#77 - short-vectors [mostly] working

Pull Request - State: closed - Opened by guacamoleo over 7 years ago

#76 - merging in SplitU and work-group mapping

Pull Request - State: closed - Opened by guacamoleo over 7 years ago

#75 - v2.2 new solution selection logic

Pull Request - State: closed - Opened by guacamoleo over 7 years ago

#74 - updating rocblas config.yaml files

Pull Request - State: closed - Opened by guacamoleo over 7 years ago

#73 - fixes critical bug in LibraryClient for multiple problem types

Pull Request - State: closed - Opened by guacamoleo over 7 years ago

#72 - Tests

Pull Request - State: closed - Opened by pfultz2 over 7 years ago

#71 - Install cmake relative to install dir

Pull Request - State: closed - Opened by pfultz2 over 7 years ago

#70 - filter out failing solution

Pull Request - State: closed - Opened by guacamoleo over 7 years ago

#69 - fixing Jenkinsfile for v2

Pull Request - State: closed - Opened by guacamoleo over 7 years ago

#68 - updating to version 2

Pull Request - State: closed - Opened by guacamoleo over 7 years ago

#67 - remove debug writes

Pull Request - State: closed - Opened by amcamd almost 8 years ago

#66 - fixing rocBLAS bug for small k; when updating fallback_pspu1 num load…

Pull Request - State: closed - Opened by guacamoleo almost 8 years ago

#65 - applying unroll fix to exact-only branch

Pull Request - State: closed - Opened by guacamoleo almost 8 years ago

#64 - fixed batched gemm

Pull Request - State: closed - Opened by guacamoleo almost 8 years ago

#63 - Develop

Pull Request - State: closed - Opened by guacamoleo almost 8 years ago

#62 - debugging hangs

Pull Request - State: closed - Opened by guacamoleo almost 8 years ago

#61 - Check if policy exists first in cmake

Pull Request - State: closed - Opened by pfultz2 almost 8 years ago

#60 - Fix warnings and add additional node to jenkins

Pull Request - State: closed - Opened by pfultz2 almost 8 years ago - 8 comments

#59 - fixing newline with clang pragmas

Pull Request - State: closed - Opened by guacamoleo almost 8 years ago

#58 - ignore clang warnings when not using clang

Pull Request - State: closed - Opened by guacamoleo almost 8 years ago

#57 - turning off debug prints

Pull Request - State: closed - Opened by guacamoleo almost 8 years ago

#56 - allowing Tensile to be built without clients

Pull Request - State: closed - Opened by guacamoleo almost 8 years ago

#55 - Develop

Pull Request - State: closed - Opened by guacamoleo almost 8 years ago

#54 - removing test dependency on opencl and fixing benchmarking for tiny s…

Pull Request - State: closed - Opened by guacamoleo about 8 years ago

#53 - fixing logical bug in library generation

Pull Request - State: closed - Opened by guacamoleo about 8 years ago

#52 - adding MIT license to every file

Pull Request - State: closed - Opened by guacamoleo about 8 years ago

#51 - updating test.py and Tensile.py to use sepperate build and outputSolutions paths

Pull Request - State: closed - Opened by guacamoleo about 8 years ago

#50 - changing Cobalt to Tensile

Pull Request - State: closed - Opened by guacamoleo about 8 years ago

#49 - eliminating and correcting compiler warnings

Pull Request - State: closed - Opened by guacamoleo about 8 years ago

#48 - fixing benchmarking protocol to enable automated test.py

Pull Request - State: closed - Opened by guacamoleo about 8 years ago

#47 - Adding the initial Jenkinsfile to enable CI builds

Pull Request - State: closed - Opened by kknox about 8 years ago

#46 - fixed bug in batched and strided kernels

Pull Request - State: closed - Opened by guacamoleo about 8 years ago

#45 - fixing bugs for new fast branch for unusual tile sizes

Pull Request - State: closed - Opened by guacamoleo about 8 years ago

#44 - adding more tests

Pull Request - State: closed - Opened by guacamoleo about 8 years ago

#43 - fixing kernels for unroll=1

Pull Request - State: closed - Opened by guacamoleo about 8 years ago

#42 - Add missing xml file for tests

Pull Request - State: closed - Opened by pfultz2 about 8 years ago

#41 - fixing visual studio warnings

Pull Request - State: closed - Opened by guacamoleo about 8 years ago

#40 - Add initial python script to test cobalt

Pull Request - State: closed - Opened by pfultz2 about 8 years ago

#39 - bug fixes for disabling fp16

Pull Request - State: closed - Opened by guacamoleo about 8 years ago

#38 - fixing compiler warnings

Pull Request - State: closed - Opened by guacamoleo about 8 years ago

#37 - Develop

Pull Request - State: closed - Opened by kknox about 8 years ago

#36 - Improved branching in kernels and library generation

Pull Request - State: closed - Opened by guacamoleo about 8 years ago

#35 - New cmake structure for cobalt

Pull Request - State: closed - Opened by pfultz2 about 8 years ago - 13 comments

#34 - adding copyright headers

Pull Request - State: closed - Opened by guacamoleo about 8 years ago

#33 - using only one unroll per tile size and one tile size for unroll=1; r…

Pull Request - State: closed - Opened by guacamoleo about 8 years ago

#32 - bug fixes from David's fork

Pull Request - State: closed - Opened by guacamoleo about 8 years ago

#31 - Linux build improvements against HiP backend

Pull Request - State: closed - Opened by kknox about 8 years ago

#30 - fixing ppdLeadingStrides

Pull Request - State: closed - Opened by guacamoleo about 8 years ago

#29 - Adding export install targets for Cobalt library

Pull Request - State: closed - Opened by kknox about 8 years ago

#28 - Updating FindHIP.cmake & FindHCC.cmake file

Pull Request - State: closed - Opened by kknox about 8 years ago

#27 - the layout of rocm dir has change; fix FindHCC.cmake to help cmake detect rocm

Pull Request - State: closed - Opened by tingxingdong about 8 years ago - 1 comment

#26 - Master

Pull Request - State: closed - Opened by guacamoleo about 8 years ago

#25 - incrementing version number

Pull Request - State: closed - Opened by guacamoleo about 8 years ago

#24 - documentation fixes

Pull Request - State: closed - Opened by guacamoleo about 8 years ago

#23 - documentation

Pull Request - State: closed - Opened by guacamoleo over 8 years ago

#22 - Master

Pull Request - State: closed - Opened by guacamoleo over 8 years ago

#21 - enumerate devices pulls from backend

Pull Request - State: closed - Opened by guacamoleo over 8 years ago

#20 - fixed solution selection log; partitioned benchmark into groups

Pull Request - State: closed - Opened by guacamoleo over 8 years ago

#19 - support for DNN

Pull Request - State: closed - Opened by guacamoleo over 8 years ago

#18 - fixed logical bug for load corner cases

Pull Request - State: closed - Opened by guacamoleo over 8 years ago

#17 - enabled 7D tensor contraction for convolutions

Pull Request - State: closed - Opened by guacamoleo over 8 years ago

#16 - rocBLAS library can build with ExternalProject_add

Pull Request - State: closed - Opened by kknox over 8 years ago

#15 - bug fixes, faster kernels

Pull Request - State: closed - Opened by guacamoleo over 8 years ago

#14 - fixed opencl bug for kernel grids larger than 16 kernels

Pull Request - State: closed - Opened by guacamoleo over 8 years ago

#13 - new load algorithm and HIP enabled

Pull Request - State: closed - Opened by guacamoleo over 8 years ago

#12 - Re-order work-groups

Issue - State: closed - Opened by guacamoleo over 8 years ago - 1 comment

#11 - Linux Compatibility

Pull Request - State: closed - Opened by guacamoleo over 8 years ago

#10 - Higher-dimensional tensor contractions

Issue - State: closed - Opened by guacamoleo over 8 years ago - 1 comment

#9 - Tensor Contractions in OpenCL

Pull Request - State: closed - Opened by guacamoleo over 8 years ago

#8 - Skinny Tensors

Issue - State: closed - Opened by guacamoleo over 8 years ago - 1 comment

#7 - Solution Selection Logic doesn't support tolerance

Issue - State: closed - Opened by guacamoleo over 8 years ago - 1 comment

#6 - Higher Accumulation Precision

Issue - State: closed - Opened by guacamoleo over 8 years ago - 1 comment

#5 - Convolution Support

Issue - State: closed - Opened by guacamoleo over 8 years ago - 1 comment

#4 - What to benchmark for clBLAS

Issue - State: closed - Opened by guacamoleo over 8 years ago - 1 comment

GitHub / rocm/tensile issues and pull requests