GPU-Accelerated Transit Least Squares (TLS) Implementation#53

Closed

johnh2o2 wants to merge 19 commits intov1.0from

tls-gpu-implementation

Commits on Oct 27, 2025

Phase 1: TLS GPU implementation - Core infrastructure

authored and
John Hoffman
committed
Phase 2: TLS GPU optimization - Advanced features

authored and
John Hoffman
committed
Phase 3: TLS production features - Statistics & usability

authored and
John Hoffman
committed
Fix TLS GPU kernel compilation and add GPU test script

authored and
John Hoffman
committed
Document RunPod GPU testing issues and solutions

authored and
John Hoffman
committed
Fix period grid generation in tls_grids.py

authored and
John Hoffman
committed
Fix critical Ofir period grid generation bugs

authored and
John Hoffman
committed
Fix critical TLS GPU bugs: Ofir grid, duration scaling, and Thrust sorting

authored and
John Hoffman
committed
Consolidate TLS to single performant kernel

authored and
John Hoffman
committed
Add Keplerian-aware duration constraints for TLS

authored and
John Hoffman
committed
Wire up Keplerian TLS Python API

authored and
John Hoffman
committed
Add PR description markdown file for easy copying

authored and
John Hoffman
committed
Clean up TLS test files and update README

authored and
John Hoffman
committed
Reorganize TLS documentation and test files

authored and
John Hoffman
committed
Update cuvarbase/tls_models.py

johnh2o2
and
Copilot
authored
Address PR review comments for TLS implementation

authored and
John Hoffman
committed
Replace insertion sort with bitonic sort for scalability

authored and
John Hoffman
committed

Commits on Feb 7, 2026

Fix TLS to use limb-darkened transit template instead of box model

authored and
John Hoffman
committed
Add automated RunPod pod lifecycle scripts

authored and
John Hoffman
committed