NVIDIA-Merlin / HierarchicalKV
HierarchicalKV is a part of NVIDIA Merlin and provides hierarchical key-value storage to meet RecSys requirements. The key capability of HierarchicalKV is to store key-value feature-embeddings on high-bandwidth memory (HBM) of GPUs and in host memory. It also can be used as a generic key-value storage.
☆144Updated this week
Alternatives and similar repositories for HierarchicalKV:
Users that are interested in HierarchicalKV are comparing it to the libraries listed below
- A high-performance framework for training wide-and-deep recommender systems on heterogeneous cluster☆157Updated last year
- TePDist (TEnsor Program DISTributed) is an HLO-level automatic distributed system for DL models.☆94Updated 2 years ago
- PyTorch distributed training acceleration framework☆48Updated 2 months ago
- NCCL Fast Socket is a transport layer plugin to improve NCCL collective communication performance on Google Cloud.☆116Updated last year
- ☆53Updated last year
- MSCCL++: A GPU-driven communication stack for scalable AI applications☆345Updated this week
- Paella: Low-latency Model Serving with Virtualized GPU Scheduling☆58Updated last year
- ☆148Updated 3 months ago
- A baseline repository of Auto-Parallelism in Training Neural Networks☆144Updated 2 years ago
- ☆79Updated 2 years ago
- A prefill & decode disaggregated LLM serving framework with shared GPU memory and fine-grained compute isolation.☆70Updated last week
- A home for the final text of all TVM RFCs.☆102Updated 7 months ago
- NCCL Profiling Kit☆133Updated 10 months ago
- RDMA and SHARP plugins for nccl library☆191Updated 3 weeks ago
- We invite you to visit and follow our new repository at https://github.com/microsoft/TileFusion. TiledCUDA is a highly efficient kernel …☆181Updated 3 months ago
- PET: Optimizing Tensor Programs with Partially Equivalent Transformations and Automated Corrections☆121Updated 2 years ago
- ☆32Updated last year
- gossip: Efficient Communication Primitives for Multi-GPU Systems☆59Updated 2 years ago
- Synthesizer for optimal collective communication algorithms☆106Updated last year
- An Efficient Pipelined Data Parallel Approach for Training Large Model☆76Updated 4 years ago
- Easy Parallel Library (EPL) is a general and efficient deep learning framework for distributed model training.☆267Updated 2 years ago
- A lightweight parameter server interface☆76Updated 2 years ago
- ☆104Updated last month
- ☆36Updated 4 months ago
- Microsoft Collective Communication Library☆344Updated last year
- A tool for examining GPU scheduling behavior.☆81Updated 8 months ago
- ☆193Updated 2 years ago
- DeepSeek-V3/R1 inference performance simulator☆115Updated last month
- ☆93Updated 7 months ago
- Automated Parallelization System and Infrastructure for Multiple Ecosystems☆78Updated 5 months ago