profile-data is a repository that publishes profiling traces and metrics from DeepSeek’s training and inference infrastructure (especially during DeepSeek-V3 / R1 experiments). The profiling data targets insights into computation-communication overlap, pipeline scheduling (e.g. DualPipe), and how MoE / EP / parallelism strategies interact in real systems. The repository contains JSON trace files like train.json, prefill.json, decode.json, and associated assets. Users can load them into tools like Chrome tracing to inspect GPU idle times, overlapping operations, and scheduling alignment. The idea is to bring transparency to internal efficiency tradeoffs, enabling researchers to reproduce, analyze, or improve on DeepSeek’s parallelism strategies. The README explains how trace data corresponds to forward/backward chunks, settings (e.g. EP64, TP1, 4K sequence length), and notes that pipeline communication is excluded for simplicity.
Features
- Profiling traces (JSON) of training, prefill, and inference phases
- Support for Chrome tracing (and similar visualization tools)
- Documentation of experimental settings (e.g. MoE, pipeline shape)
- Transparency into compute-communication overlap and pipeline bubble behavior
- Public access to real system performance data from DeepSeek stack
- Enables external analysis, validation, and benchmarking of parallelism strategies