Commits

Branch selector

 master

User selector

zhengzuohe

Datepicker

All time

Commit History

Commits on Mar 12, 2022

• fix the bug in npu_isl_emitter::EmitGemmRangeInfoBackpropFiltet

  zhengzuohe

  committed
  f67db2e

  Copy full SHA for f67db2e

Commits on Mar 9, 2022

• fix warnings in gpu_isl_emitter, gpu_tensor_core_emit_pass, isl_emitter_csr, cpu_isl_emitter, npu_isl_emitter

  zhengzuohe

  committed
  a60444e

  Copy full SHA for a60444e

Commits on Feb 23, 2022

• split yolov3_darknet53 tensorcore testcases into 4 parts

  zhengzuohe

  committed
  3b2e978

  Copy full SHA for 3b2e978

Commits on Feb 8, 2022

• exclude constant batch axes while analysing matmul stmt

  zhengzuohe

  committed
  1eec62c

  Copy full SHA for 1eec62c

Commits on Nov 16, 2021

• Add tiling repo for GNN-GCN to fix precision and compiling error

  zhengzuohe

  committed
  cbfd65a

  Copy full SHA for cbfd65a

Commits on Oct 15, 2021

• remove g_attrs as the parameter of AutoPoly and the related functions

  zhengzuohe

  committed
  7dc4764

  Copy full SHA for 7dc4764

Commits on Sep 24, 2021

• add attrs in repository_gpu for certain cases of bert, bert-crf and tinybert

  zhengzuohe

  committed
  ba3ec41

  Copy full SHA for ba3ec41

Commits on Sep 22, 2021

• Disable loop_reversal only on CCE cases

  zhengzuohe

  committed
  c7dba3b

  Copy full SHA for c7dba3b

Commits on Sep 14, 2021

• remove PermuteOuterBand

  zhengzuohe

  committed
  b037993

  Copy full SHA for b037993

• Add a constrain of enabling reducelib to avoid isl exception error caused by huge tensor shapes in GNN-GAT

  zhengzuohe

  committed
  067b01a

  Copy full SHA for 067b01a

Commits on Sep 11, 2021

• code refactoring of reduce-lib usage (fix the bug of enabling reduce lib on elemwise OPs, which causes GPU performance degradation)

  zhengzuohe

  committed
  ad2af02

  Copy full SHA for ad2af02

Commits on Aug 20, 2021

• Add gnn CI cases

  Show description for bb95b5b
  zhengzuohe

  committed
  bb95b5b

  Copy full SHA for bb95b5b

Commits on Aug 18, 2021

• Split tensorcore-related info files out~

  zhengzuohe

  committed
  42a7a1c

  Copy full SHA for 42a7a1c

Commits on Aug 12, 2021

• Add CI test cases of MatMul, Conv, Pad and Unpad OPs in networks

  zhengzuohe

  committed
  668da84

  Copy full SHA for 668da84

Commits on Jul 21, 2021

• Add CumSum/CumProd OPs in topi.py. Add test cases at tests/operators/gpu. Complete the tiling strategy for this op type

  zhengzuohe

  committed
  a047424

  Copy full SHA for a047424

Commits on Jul 7, 2021

• reconstruct and merge the DSL of conv and conv_tensorcore

  Show description for be3cf4e
  zhengzuohe

  committed
  be3cf4e

  Copy full SHA for be3cf4e

Commits on Jun 30, 2021

• add constraints of auto-tiling for conv

  Show description for 772b4f7
  zhengzuohe

  committed
  772b4f7

  Copy full SHA for 772b4f7

Commits on Jun 22, 2021

• Add a precheck function for fused elemwise ops

  zhengzuohe

  committed
  a49f462

  Copy full SHA for a49f462

Commits on Jun 16, 2021

• remove testing cases that have numpy precision error

  zhengzuohe

  committed
  8ac5609

  Copy full SHA for 8ac5609

Commits on Jun 11, 2021

• Add info files of all networks enabled Graph-Kernel into CI testing cases

  zhengzuohe

  committed
  57a8f6e

  Copy full SHA for 57a8f6e

Commits on Jun 10, 2021

• avoid updating attributes repeatedly when parallel fusion enabled

  zhengzuohe

  committed
  3130185

  Copy full SHA for 3130185

Commits on Jun 1, 2021

• Use speedup tiling algorithm when target is cuda

  zhengzuohe

  committed
  7970ec2

  Copy full SHA for 7970ec2