quiver-team/quiver-feature

Issues

• 无法安装

  #47 opened a year ago by xbinglzh
  0

• 非常感兴趣这个项目，请问有未来的RoadMap么，还有还允许加入么？

  #46 opened 2 years ago by cvnovice95
  1

• Support DGL Examples

  #43 opened 2 years ago by eedalong
  0

• Support Float16 Tensor

  #44 opened 2 years ago by eedalong
  0

• TODO for release 0.1.0

  #32 opened 2 years ago by eedalong
  0

• LocalTensorPGAS create from file using shared_loader

  #16 opened 2 years ago by eedalong
  0

• RDMA特征聚合说明文档

  #21 opened 2 years ago by eedalong
  0

• MultiThread + MultiQP in DistTensorClient

  #22 opened 2 years ago by eedalong
  0

• DistHelper Doc

  #31 opened 2 years ago by eedalong
  1

• Auto Placement

  #39 opened 2 years ago by eedalong
  0

• PyPI Package

  #29 opened 2 years ago by eedalong
  0

• GDR Mode Support

  #34 opened 2 years ago by eedalong
  0

• TLB test results

  #17 opened 2 years ago by ZenoTan
  0

• Large memory register in IB

  #13 opened 2 years ago by eedalong
  1

• 特征聚合性能测试数据

  #1 opened 2 years ago by eedalong
  5

• DGL UnifiedTensor随着FeatureSize的性能测试

  #20 opened 2 years ago by eedalong
  4

• DistTensorPGAS load from local data api

  #36 opened 2 years ago by eedalong
  0

• Support IPC for DistTensorPGAS for easier use

  #27 opened 2 years ago by eedalong
  0

• Pipe has implementation bug

  #28 opened 2 years ago by eedalong
  1

• distHelper.sync_all()似乎没有按照预期工作

  #18 opened 2 years ago by eedalong
  0

• Fix max_qp_rd_atom setting

  #25 opened 2 years ago by eedalong
  0

• About RDMA Scatter/ Gather & RC QP's max_rd_atomic

  #24 opened 2 years ago by eedalong
  1

• RDMA TLB 在不同特征维度下的测试

  #19 opened 2 years ago by eedalong
  3

• 测试随着数据集的增大导致的寻址开销带来的影响，以及PageSize带来的影响

  #10 opened 2 years ago by eedalong
  1

• 端到端训练性能测试数据

  #12 opened 2 years ago by eedalong
  1

• Iteration Plan

  #7 opened 2 years ago by eedalong
  1

• Single Machine quiver.Feature TLB optimization

  #9 opened 2 years ago by eedalong
  0

• 更省内存的Server以及进程启动

  #11 opened 2 years ago by eedalong
  2

• 测试paper100M数据的两机collect性能

  #4 opened 2 years ago by eedalong
  4

• 测试全链路的时间耗时，包括前处理，后处理，以及本地collect信息

  #5 opened 2 years ago by eedalong
  7

• 加入缓存和处理

  #2 opened 2 years ago by eedalong
  0