1.0.6

What's Changed

See the blog post post that explain this in more detail:

• Shöve 1.0.6 - Sometimes Less is More!

This release focuses on performance improvements across all systems, particularly benefiting devices with integrated graphics:

• Removed custom layer batching: Benchmarking revealed LÖVE's native batching was more efficient
• Implemented canvas pooling: Separate pools for standard and stencil buffers reduce memory allocations
• Optimized shader effect pipeline: New ping-pong buffer system improves multi-effect rendering
• Added stencil state tracking: Completes GPU state management for more efficient rendering
• Improved profiler: Canvas caching for overlays and optimized memory allocation
• Enhanced FPS display: More accurate calculations with lower performance impact

Benchmarks show an average 3.5% FPS improvement on high-end systems and 12.9% improvement on systems with integrated graphics.

• perf(render): defer layer sorting until compositing by @flexiondotorg in #23
• refactor(canvas): optimise canvas creation by @flexiondotorg in #24
• perf(rendering): optimize layer signature generation with numeric hashing by @flexiondotorg in #25
• perf(rendering): optimize layer batching with single-layer fast path by @flexiondotorg in #26
• perf(canvas): implement separate canvas pools for standard and stencil buffers by @flexiondotorg in #27
• perf(effects): optimize shader effect application pipeline by @flexiondotorg in #28
• perf(layers): optimize layer signature calculation algorithm by @flexiondotorg in #29
• perf(renderer): optimize layer batch drawing with hybrid approach by @flexiondotorg in #30
• perf(rendering): optimize shader state changes with caching system by @flexiondotorg in #31
• perf(rendering): optimise blend mode state changes with tracking system by @flexiondotorg in #32
• perf(rendering): implement stencil state tracking system by @flexiondotorg in #33
• fix(profiler): removes redundant state change counts by @flexiondotorg in #34
• perf(profiler): optimize memory allocation in getState function by @flexiondotorg in #35
• feat(profiler): display average FPS by @flexiondotorg in #36
• perf(rendering): remove layer batching optimizations by @flexiondotorg in #37
• feat(rendering): reinstate sequential shader effect chaining by @flexiondotorg in #38
• refactor(rendering): make sharedEffectsTable local to compositeLayersOnScreen by @flexiondotorg in #39
• refactor(profiler): clean up getState() and remove unused layer properties by @flexiondotorg in #40
• perf(profiler): optimize overlay rendering with canvas caching by @flexiondotorg in #41
• refactor(profiler): simplify overlay toggling with unified state management by @flexiondotorg in #42
• perf(overlay) cache FPS overlay and improve content change detection by @flexiondotorg in #43
• docs: updates demo descriptions and order by @flexiondotorg in #44

Full Changelog: 1.0.5...1.0.6