CHANGELOG.md

Changelog

v0.3.0

Research / simulator

• added prefill/decode disaggregation with independent accelerator profiles and worker counts
• added a genuine event-driven P/D pipeline with overlapping prefill, transfer, and decode activity
• added explicit serialized KV-transfer modeling with bandwidth/base-latency controls
• added P/D telemetry: role utilization, transfer volume, mean/p95 transfer latency, and transfer utilization
• added P/D-specific bottleneck diagnoses for prefill-pool, decode-pool, and KV-transfer pressure
• added deterministic shared-prefix reuse with independent cache RNG so enabling cache preserves the workload trace
• added shared-prefix KV accounting and saved-prefill-token telemetry
• added four-way colocated/P-D x cache-off/on comparison
• added bounded design-space exploration and goodput/p95-TTFT Pareto-frontier marking

UI / usability

• removed the product-style footer
• simplified top branding toward a research-prototype presentation
• added explicit Worst repetition and Target columns to Capacity Planner tables/exports
• removed area fill from the capacity line chart
• changed line-chart legends to line swatches rather than rectangle blocks
• added PNG download controls to every chart
• added meaningful experiment-aware PNG filenames instead of generic chart names
• retained full-screen chart expansion and Escape-to-close behavior
• added Modern Serving and Design Explorer sections
• added live topology, shared-prefix, P/D worker, and interconnect controls to Serving Lab

Validation

• expanded tests to 17 deterministic cases
• release check now validates P/D transfer telemetry, design-space output, PNG controls, planner evidence columns, UI ASCII safety, and absence of the old footer

v0.2.0

• auto-hide runtime header after Pyodide becomes ready; hover/focus at the top edge reveals it
• redesigned the interface around a flatter engineering-dashboard visual system
• replaced decorative/non-ASCII ranking glyphs with explicit text labels
• added chart expansion with Escape-to-close support
• added copy-to-clipboard and CSV export for Scheduler Arena and Capacity Planner tables
• added JSON copy alongside download for individual simulations
• exposed burst multiplier and burst-period controls
• added component TTFT/E2E SLO-attainment accounting and failure breakdowns
• added heuristic simulator-side bottleneck diagnosis with evidence and next-step guidance
• made capacity feasibility conservative across repetitions: every repetition must meet the target and fully drain
• added minimum/mean/maximum repetition evidence to capacity traces
• changed the capacity plot to a true linear request-rate x-axis and added worst-repetition evidence

v0.1.0

Initial browser-native Python release with workload generation, static/continuous scheduling, chunked prefill, KV-cache accounting, analytical latency profiles, Scheduler Arena, and SLO capacity search.