Update README with GPU acceleration and v2.0.0 docs

[joeljose]

Summary

• Added GPU Acceleration section documenting two-pass batched architecture, chunked cuFFT, auto-tuning, and design decisions table
• Added Performance section with benchmarks (RTX 4050) and hardware requirements
• Added all missing CLI flags to usage table: --gpu, --device, --biort, --qshift
• Added GPU Docker setup section (Dockerfile.gpu, docker-build-gpu.sh, nvidia-container-toolkit)
• Added v2.0.0 Breaking Changes section (default filter change, how to restore old behavior)
• Updated project structure tree with all missing files
• Updated development section with GPU test documentation (tests/test_motion_mag_gpu.py, ./test.sh gpu)
• Added ASCII pipeline diagram to Algorithm Pipeline section

Fixes #19

Test plan

• Docs-only change — no code modified
• Verify README renders correctly on GitHub

🤖 Generated with Claude Code