CGCL-codes / Tensorflow-RDMA
Tensorflow is a computational library using data flow graphs for scalable machine learning, and Tensorflow-RDMA is the implementation over RDMA, which can get about 4.5x speedup on two nodes comparing with TCP/IP.
☆58Updated 2 years ago
Alternatives and similar repositories for Tensorflow-RDMA:
Users that are interested in Tensorflow-RDMA are comparing it to the libraries listed below
- LITE Kernel RDMA Support for Datacenter Applications. SOSP 2017.☆108Updated 4 years ago
- ☆67Updated 7 years ago
- A NUMA-aware Graph-structured Analytics Framework☆42Updated 6 years ago
- 2RDMA_Aware_Programming_user_manual 一书的中文翻译☆78Updated 11 years ago
- verbs profiling library☆22Updated last year
- ☆72Updated 8 years ago
- Utility to simulate cache behavior with Mattson's Stack Algorithm.☆27Updated 3 years ago
- Source code for our OSDI 2016 paper☆110Updated 6 years ago
- Fast In-memory Transaction Processing using Hybrid RDMA Primitives☆66Updated 6 years ago
- Differentiated Computation and Partitioning on Skewed (Natural or Bipartite) Graphs☆66Updated 3 years ago
- A Simple RDMA Wheel☆22Updated 6 years ago
- ☆20Updated 7 years ago
- Cocytus is an efficient and available in-memory K/V-store through hybrid erasure coding and replication☆30Updated 9 years ago
- High Performance Network Library for RDMA☆27Updated 2 years ago
- Fast In-memory Transaction Processing using RDMA and HTM☆57Updated 9 years ago
- Distributed Shared Persistent Memory. SoCC 2017☆69Updated 4 years ago
- A lightweight parameter server interface☆76Updated 2 years ago
- A kernel module to enable RDMA transfers to/from IO, PFN and DAX mapped memory☆10Updated 9 years ago
- Writing RDMA applications on Linux Example programs☆45Updated last year
- Enhanced networking support for TensorFlow. Maintained by SIG-networking.☆98Updated 3 years ago
- Graphene: Fine-Grained IO Management for Graph Computing. FAST'17☆20Updated 8 years ago
- RDMA Optimization on MXNet☆14Updated 7 years ago
- HME a hybrid memory emulator for studying the performance and energy characteristics of upcoming NVM technologies. HME exploits features …☆49Updated 2 years ago
- ☆33Updated 6 years ago
- GPU-specialized parameter server for GPU machine learning.☆101Updated 7 years ago
- Arbitrary offloads for RDMA NICs☆89Updated 3 years ago
- Fine-grained GPU sharing primitives☆141Updated 5 years ago
- An RDMA-enabled Distributed Persistent Memory File System☆157Updated 7 years ago
- ☆25Updated last year
- Prefetching and efficient data path for memory disaggregation☆67Updated 4 years ago