raylib perf - comparison

rlgl vs rlmtl_mesa | GPU Apple M5 (24576 MB) | Driver MoltenVK 1.4.1 | OS macOS 26.5.1 | 10 runs × 60000 ms full speed

Green = best backend for that example/metric, red = worst. Frame time & FPS are the representative run; software (rlsw) has no GPU so its VRAM is ~0. Shader-heavy examples may not execute custom shaders on the software backend.

° = at that backend's floor (median within 15% of its own bench_idle median): the scene finishes its real work faster than one present/loop turnaround, so the number measures presentation pacing or loop overhead, not rendering cost, and cross-backend ratios there compare present plumbing.

rlmtl_mesa = the native Metal backend + Mesa's shader compiler. This column runs rlmtl's stock pipeline (GLSL → shaderc → SPIR-V → SPIRV-Cross → MSL) on every scene except performance_stress_test_direct, whose fragment shader is precompiled offline by raylib/tools/nir2msl through Mesa's NIR→MSL compiler (the KosmicKrisp backend's kosmicomp, MIT) and injected via RLMTL_MSL_OVERRIDE. On that scene SPIRV-Cross's flattened-SSA output defeats Metal's optimizer; Mesa's NIR pipeline is the only automatic GLSL translation measured faster than Apple's own GL compiler — 231.6 ±14 vs 247.4 ±9 ms over 10 interleaved cycles (2026-08-18), ahead of hand-written MSL (250), ANGLE (281), Slang (289), naga (322) and stock SPIRV-Cross (442). Pixel drift vs SPIRV-Cross is ULP-class only (84% of differing channels off by 1, max 17/255; Mesa lowers sin/cos to conformant polynomials).

speedup = rlmtl_mesa relative to rlgl, per metric direction (>1x favors rlmtl_mesa_macos_apple/bench_instanced/run_16.rini); ° on a ratio means at least one side sits at its floor, so it compares present plumbing.

Frames per second (higher is better)

Sustained FPS at full speed (uncapped).
Example	rlgl	rlmtl_mesa	rlmtl_mesa speedup
shapes_basic_shapes	6801°	17938	2.64x°
textures_tiled_drawing	6380°	15115	2.37x°
textures_particles_blending	5956°	9896	1.66x°
shapes_starfield_effect	6605°	18824	2.85x°
models_loading	2469	13760°	5.57x°
models_heightmap_rendering	2157	12564	5.83x
models_skybox_rendering	2590	15785	6.09x
models_waving_cubes	249	1237	4.97x
models_first_person_maze	7757°	13734°	1.77x°
core_3d_camera_free	2591	17905	6.91x
shaders_basic_lighting	3619	9026	2.49x
shaders_basic_pbr	3352	4915	1.47x
shaders_mandelbrot_set	1158	4146	3.58x
shaders_raymarching_rendering	276	232	0.84x
performance_stress_test	83	163	1.97x
performance_stress_test_direct	4	4	1.09x
bench_idle	5779°	14756°	2.55x°
bench_drawcalls	120	2151	17.90x
bench_instanced	517	8199	15.86x

Median frame time, ms (lower is better)

Median per-frame CPU wall time.
Example	rlgl	rlmtl_mesa	rlmtl_mesa speedup
shapes_basic_shapes	0.134°	0.041	3.27x°
textures_tiled_drawing	0.145°	0.040	3.61x°
textures_particles_blending	0.164°	0.071	2.32x°
shapes_starfield_effect	0.142°	0.042	3.41x°
models_loading	0.368	0.033°	11.30x°
models_heightmap_rendering	0.401	0.047	8.43x
models_skybox_rendering	0.359	0.039	9.28x
models_waving_cubes	4.039	0.803	5.03x
models_first_person_maze	0.123°	0.035°	3.51x°
core_3d_camera_free	0.359	0.041	8.82x
shaders_basic_lighting	0.269	0.082	3.26x
shaders_basic_pbr	0.287	0.177	1.62x
shaders_mandelbrot_set	0.827	0.236	3.50x
shaders_raymarching_rendering	3.500	4.014	0.87x
performance_stress_test	11.668	5.965	1.96x
performance_stress_test_direct	247.350	231.606	1.07x
bench_idle	0.157°	0.031°	5.00x°
bench_drawcalls	7.958	0.474	16.79x
bench_instanced	1.667	0.114	14.58x

Average frame time, ms (lower is better)

Mean per-frame wall time.
Example	rlgl	rlmtl_mesa	rlmtl_mesa speedup
shapes_basic_shapes	0.148°	0.056	2.64x°
textures_tiled_drawing	0.158°	0.066	2.39x°
textures_particles_blending	0.172°	0.101	1.70x°
shapes_starfield_effect	0.152°	0.053	2.87x°
models_loading	0.405	0.073°	5.58x°
models_heightmap_rendering	0.465	0.080	5.83x
models_skybox_rendering	0.386	0.063	6.09x
models_waving_cubes	4.038	0.809	4.99x
models_first_person_maze	0.129°	0.073°	1.77x°
core_3d_camera_free	0.386	0.056	6.91x
shaders_basic_lighting	0.278	0.111	2.50x
shaders_basic_pbr	0.301	0.204	1.48x
shaders_mandelbrot_set	0.864	0.241	3.58x
shaders_raymarching_rendering	3.631	4.306	0.84x
performance_stress_test	12.089	6.147	1.97x
performance_stress_test_direct	248.378	228.219	1.09x
bench_idle	0.175°	0.068°	2.59x°
bench_drawcalls	8.320	0.465	17.88x
bench_instanced	1.935	0.122	15.86x

CPU utilization, %

Average process CPU as percent of the whole machine.
Example	rlgl	rlmtl_mesa	rlmtl_mesa speedup
shapes_basic_shapes	12.7	13.5	0.94x
textures_tiled_drawing	12.5	13.4	0.93x
textures_particles_blending	12.8	16.6	0.77x
shapes_starfield_effect	12.5	15.2	0.83x
models_loading	5.6	10.0	0.56x
models_heightmap_rendering	5.0	12.4	0.40x
models_skybox_rendering	5.7	11.7	0.49x
models_waving_cubes	8.3	11.6	0.72x
models_first_person_maze	12.8	10.4	1.22x
core_3d_camera_free	5.9	12.8	0.46x
shaders_basic_lighting	12.8	14.9	0.86x
shaders_basic_pbr	11.9	8.5	1.40x
shaders_mandelbrot_set	4.0	8.4	0.48x
shaders_raymarching_rendering	1.0	0.5	1.88x
performance_stress_test	10.2	10.4	0.98x
performance_stress_test_direct	0.0	0.0	1.12x
bench_idle	12.4	9.3	1.33x
bench_drawcalls	10.0	12.1	0.82x
bench_instanced	6.9	16.8	0.41x

RAM, MB

Average working-set memory.
Example	rlgl	rlmtl_mesa	rlmtl_mesa speedup
shapes_basic_shapes	87.4	83.8	1.04x
textures_tiled_drawing	87.9	84.3	1.04x
textures_particles_blending	87.7	83.9	1.05x
shapes_starfield_effect	87.1	83.9	1.04x
models_loading	103.5	99.6	1.04x
models_heightmap_rendering	92.8	90.1	1.03x
models_skybox_rendering	148.0	161.9	0.91x
models_waving_cubes	87.9	100.7	0.87x
models_first_person_maze	88.8	84.9	1.05x
core_3d_camera_free	86.7	83.7	1.03x
shaders_basic_lighting	94.4	102.2	0.92x
shaders_basic_pbr	132.9	140.8	0.94x
shaders_mandelbrot_set	86.6	100.8	0.86x
shaders_raymarching_rendering	87.3	119.1	0.73x
performance_stress_test	122.8	164.1	0.75x
performance_stress_test_direct	86.5	102.9	0.84x
bench_idle	86.9	83.4	1.04x
bench_drawcalls	89.5	84.8	1.06x
bench_instanced	88.3	103.9	0.85x

GPU VRAM, MB