Add MGE imaging profiling: A100 + RTX 2060 + CPU sweep

jax_profiling/results/jit/imaging/mge/hpc_a100_fp64.json

@@ -0,0 +1,46 @@
+{
+  "config_name": "hpc_a100_fp64",
+  "use_mixed_precision": false,
+  "autolens_version": "2026.5.1.4",
+  "instrument": "hst",
+  "configuration": {
+    "pixel_scale_arcsec": 0.05,
+    "mask_radius_arcsec": 3.5,
+    "image_pixels_masked": 15361,
+    "over_sampled_pixels": 17980,
+    "linear_gaussians": 0
+  },
+  "device": {
+    "backend": "gpu",
+    "device": "cuda:0",
+    "nvidia_smi": "NVIDIA A100 80GB PCIe, 19790 MiB, 81920 MiB"
+  },
+  "log_likelihood": {
+    "eager_numpy": -159736.35504220804,
+    "step_by_step_jit": -159734.65625,
+    "full_pipeline_jit": -159734.59375,
+    "vmap_first_element": -159737.8125
+  },
+  "steps": {
+    "Ray-trace grids": 0.000232636698638089,
+    "Mapping matrix": 0.001340581598924473,
+    "Blurred mapping matrix": 0.003814887901535258,
+    "Data vector (D)": 0.00023779570183251052,
+    "Curvature matrix (F)": 0.00028616949857678263,
+    "Reconstruction (NNLS)": 0.00199669889989309,
+    "Mapped reconstructed image": 0.00021696779876947402,
+    "Chi-squared & log likelihood": 0.00023722680052742361
+  },
+  "total_step_by_step": 0.0083629648986971,
+  "full_pipeline_per_call": 0.005700614600209519,
+  "vmap": {
+    "batch_size": 3,
+    "batch_time": 0.0071246926992898805,
+    "per_call": 0.00237489756642996,
+    "speedup_vs_single_jit": 2.4
+  },
+  "memory_mb": {
+    "output": 1.1444091796875e-05,
+    "temp": 24.970977783203125
+  }
+}

jax_profiling/results/jit/imaging/mge/hpc_a100_mp.json

@@ -0,0 +1,46 @@
+{
+  "config_name": "hpc_a100_mp",
+  "use_mixed_precision": true,
+  "autolens_version": "2026.5.1.4",
+  "instrument": "hst",
+  "configuration": {
+    "pixel_scale_arcsec": 0.05,
+    "mask_radius_arcsec": 3.5,
+    "image_pixels_masked": 15361,
+    "over_sampled_pixels": 17980,
+    "linear_gaussians": 0
+  },
+  "device": {
+    "backend": "gpu",
+    "device": "cuda:0",
+    "nvidia_smi": "NVIDIA A100 80GB PCIe, 19790 MiB, 81920 MiB"
+  },
+  "log_likelihood": {
+    "eager_numpy": -159736.35504220804,
+    "step_by_step_jit": -159734.65625,
+    "full_pipeline_jit": -159734.59375,
+    "vmap_first_element": -159737.8125
+  },
+  "steps": {
+    "Ray-trace grids": 0.0002060166996670887,
+    "Mapping matrix": 0.0012590978003572673,
+    "Blurred mapping matrix": 0.0036326707020634787,
+    "Data vector (D)": 0.00020699769956991076,
+    "Curvature matrix (F)": 0.0002976489980937913,
+    "Reconstruction (NNLS)": 0.0019418963987845928,
+    "Mapped reconstructed image": 0.00022419560118578375,
+    "Chi-squared & log likelihood": 0.0002323824999621138
+  },
+  "total_step_by_step": 0.008000906399684028,
+  "full_pipeline_per_call": 0.00544624489848502,
+  "vmap": {
+    "batch_size": 3,
+    "batch_time": 0.006934574301703833,
+    "per_call": 0.002311524767234611,
+    "speedup_vs_single_jit": 2.36
+  },
+  "memory_mb": {
+    "output": 1.1444091796875e-05,
+    "temp": 24.970977783203125
+  }
+}

jax_profiling/results/jit/imaging/mge/local_cpu_fp64.json

@@ -0,0 +1,45 @@
+{
+  "config_name": "local_cpu_fp64",
+  "use_mixed_precision": false,
+  "autolens_version": "2026.5.8.2",
+  "instrument": "hst",
+  "configuration": {
+    "pixel_scale_arcsec": 0.05,
+    "mask_radius_arcsec": 3.5,
+    "image_pixels_masked": 15361,
+    "over_sampled_pixels": 17980,
+    "linear_gaussians": 0
+  },
+  "device": {
+    "backend": "cpu",
+    "device": "cpu"
+  },
+  "log_likelihood": {
+    "eager_numpy": -159736.35504220804,
+    "step_by_step_jit": -159736.35504220804,
+    "full_pipeline_jit": -159736.35504220793,
+    "vmap_first_element": -159736.355042208
+  },
+  "steps": {
+    "Ray-trace grids": 0.0029358899999806453,
+    "Mapping matrix": 0.15586809000001267,
+    "Blurred mapping matrix": 0.3403517600000214,
+    "Data vector (D)": 0.00046917999998186133,
+    "Curvature matrix (F)": 0.005527689999962604,
+    "Reconstruction (NNLS)": 0.0008506399999987479,
+    "Mapped reconstructed image": 0.00048670999999558264,
+    "Chi-squared & log likelihood": 0.00028616999998121173
+  },
+  "total_step_by_step": 0.5067761299999347,
+  "full_pipeline_per_call": 0.3082785000000058,
+  "vmap": {
+    "batch_size": 3,
+    "batch_time": 0.7007651299999906,
+    "per_call": 0.23358837666666354,
+    "speedup_vs_single_jit": 1.32
+  },
+  "memory_mb": {
+    "output": 2.288818359375e-05,
+    "temp": 42.27388000488281
+  }
+}

jax_profiling/results/jit/imaging/mge/local_cpu_fp64_pre_fix.json

@@ -0,0 +1,35 @@
+{
+  "config_name": "local_cpu_fp64_pre_fix",
+  "use_mixed_precision": false,
+  "autolens_version": "unknown",
+  "instrument": "hst",
+  "configuration": {},
+  "device": {
+    "backend": "cpu",
+    "device": "cpu",
+    "note": "pre-fix"
+  },
+  "log_likelihood": {
+    "full_pipeline_jit": 27379.38890685523
+  },
+  "steps": {
+    "Ray-trace grids": 0.004253,
+    "Mapping matrix": 0.113323,
+    "Blurred mapping matrix": 0.200123,
+    "Data vector (D)": 0.000315,
+    "Curvature matrix (F)": 0.001596,
+    "Reconstruction (NNLS)": 0.00061,
+    "Mapped reconstructed image": 0.000272,
+    "Chi-squared & log likelihood": 0.000269
+  },
+  "total_step_by_step": 0.32076099999999996,
+  "full_pipeline_per_call": 0.208569,
+  "vmap": {
+    "batch_size": 3,
+    "per_call": 0.128317,
+    "speedup_vs_single_jit": 1.6,
+    "batch_time": null
+  },
+  "memory_mb": {},
+  "_provenance": "ingested from pre-FFT-fix /tmp profiling artifacts"
+}

jax_profiling/results/jit/imaging/mge/local_cpu_mp.json

@@ -0,0 +1,45 @@
+{
+  "config_name": "local_cpu_mp",
+  "use_mixed_precision": true,
+  "autolens_version": "2026.5.8.2",
+  "instrument": "hst",
+  "configuration": {
+    "pixel_scale_arcsec": 0.05,
+    "mask_radius_arcsec": 3.5,
+    "image_pixels_masked": 15361,
+    "over_sampled_pixels": 17980,
+    "linear_gaussians": 0
+  },
+  "device": {
+    "backend": "cpu",
+    "device": "cpu"
+  },
+  "log_likelihood": {
+    "eager_numpy": -159736.35504220804,
+    "step_by_step_jit": -159736.35504220804,
+    "full_pipeline_jit": -159736.362460121,
+    "vmap_first_element": -159736.35937145964
+  },
+  "steps": {
+    "Ray-trace grids": 0.003994999999986248,
+    "Mapping matrix": 0.18811679999998887,
+    "Blurred mapping matrix": 0.28807799000001066,
+    "Data vector (D)": 0.0006640400000378577,
+    "Curvature matrix (F)": 0.005379319999974541,
+    "Reconstruction (NNLS)": 0.0007877799999732815,
+    "Mapped reconstructed image": 0.00032731999999668915,
+    "Chi-squared & log likelihood": 0.00034153999999944065
+  },
+  "total_step_by_step": 0.48768978999996754,
+  "full_pipeline_per_call": 0.32217029999997066,
+  "vmap": {
+    "batch_size": 3,
+    "batch_time": 0.6968236900000193,
+    "per_call": 0.23227456333333976,
+    "speedup_vs_single_jit": 1.39
+  },
+  "memory_mb": {
+    "output": 2.288818359375e-05,
+    "temp": 28.210952758789062
+  }
+}

jax_profiling/results/jit/imaging/mge/local_cpu_mp_pre_fix.json

@@ -0,0 +1,41 @@
+{
+  "config_name": "local_cpu_mp_pre_fix",
+  "use_mixed_precision": true,
+  "autolens_version": "2026.5.1.4",
+  "instrument": "hst",
+  "configuration": {
+    "pixel_scale_arcsec": 0.05,
+    "mask_radius_arcsec": 3.5,
+    "image_pixels_masked": 15361,
+    "over_sampled_pixels": 17980,
+    "linear_gaussians": 0
+  },
+  "device": {
+    "backend": "cpu",
+    "device": "cpu",
+    "note": "pre-fix"
+  },
+  "log_likelihood": {
+    "full_pipeline_jit": 27379.388781919046
+  },
+  "steps": {
+    "Ray-trace grids": 0.002722230000017589,
+    "Mapping matrix": 0.18460562999998728,
+    "Blurred mapping matrix": 0.189258239999981,
+    "Data vector (D)": 0.0002965399999993679,
+    "Curvature matrix (F)": 0.0015331399999922724,
+    "Reconstruction (NNLS)": 0.0007736699999441043,
+    "Mapped reconstructed image": 0.00019877000004271396,
+    "Chi-squared & log likelihood": 0.0002946099999462604
+  },
+  "total_step_by_step": 0.37968282999991054,
+  "full_pipeline_per_call": 0.19330503000001045,
+  "vmap": {
+    "batch_size": 3,
+    "per_call": 0.12038505666669153,
+    "speedup_vs_single_jit": 1.6,
+    "batch_time": 0.3611551700000746
+  },
+  "memory_mb": {},
+  "_provenance": "ingested from pre-FFT-fix /tmp profiling artifacts"
+}

jax_profiling/results/jit/imaging/mge/local_gpu_fp64.json

@@ -0,0 +1,46 @@
+{
+  "config_name": "local_gpu_fp64",
+  "use_mixed_precision": false,
+  "autolens_version": "2026.5.8.2",
+  "instrument": "hst",
+  "configuration": {
+    "pixel_scale_arcsec": 0.05,
+    "mask_radius_arcsec": 3.5,
+    "image_pixels_masked": 15361,
+    "over_sampled_pixels": 17980,
+    "linear_gaussians": 0
+  },
+  "device": {
+    "backend": "gpu",
+    "device": "cuda:0",
+    "nvidia_smi": "NVIDIA GeForce RTX 2060 with Max-Q Design, 1194 MiB, 6144 MiB"
+  },
+  "log_likelihood": {
+    "eager_numpy": -159736.35504220804,
+    "step_by_step_jit": -159736.35504220804,
+    "full_pipeline_jit": -159736.355042208,
+    "vmap_first_element": -159736.355042208
+  },
+  "steps": {
+    "Ray-trace grids": 0.0031467600000269157,
+    "Mapping matrix": 0.0273685199999818,
+    "Blurred mapping matrix": 0.03429440999998405,
+    "Data vector (D)": 0.00038411999998970713,
+    "Curvature matrix (F)": 0.0017256699999961711,
+    "Reconstruction (NNLS)": 0.01145449999999073,
+    "Mapped reconstructed image": 0.00036515000001600126,
+    "Chi-squared & log likelihood": 0.00043246000000181085
+  },
+  "total_step_by_step": 0.07917158999998718,
+  "full_pipeline_per_call": 0.043708280000009835,
+  "vmap": {
+    "batch_size": 3,
+    "batch_time": 0.07181166000000302,
+    "per_call": 0.023937220000001008,
+    "speedup_vs_single_jit": 1.83
+  },
+  "memory_mb": {
+    "output": 2.288818359375e-05,
+    "temp": 59.006126403808594
+  }
+}

jax_profiling/results/jit/imaging/mge/local_gpu_fp64_pre_fix.json

@@ -0,0 +1,41 @@
+{
+  "config_name": "local_gpu_fp64_pre_fix",
+  "use_mixed_precision": false,
+  "autolens_version": "2026.5.1.4",
+  "instrument": "hst",
+  "configuration": {
+    "pixel_scale_arcsec": 0.05,
+    "mask_radius_arcsec": 3.5,
+    "image_pixels_masked": 15361,
+    "over_sampled_pixels": 17980,
+    "linear_gaussians": 0
+  },
+  "device": {
+    "backend": "gpu",
+    "device": "gpu",
+    "note": "pre-fix"
+  },
+  "log_likelihood": {
+    "full_pipeline_jit": 27379.38890685539
+  },
+  "steps": {
+    "Ray-trace grids": 0.0031153899999935676,
+    "Mapping matrix": 0.022293959999979053,
+    "Blurred mapping matrix": 0.01880874000003132,
+    "Data vector (D)": 0.00037680999998883634,
+    "Curvature matrix (F)": 0.0015178300000115996,
+    "Reconstruction (NNLS)": 0.019577590000017152,
+    "Mapped reconstructed image": 0.00032828000003064516,
+    "Chi-squared & log likelihood": 0.000306780000028084
+  },
+  "total_step_by_step": 0.06632538000008026,
+  "full_pipeline_per_call": 0.03705118000002585,
+  "vmap": {
+    "batch_size": 3,
+    "per_call": 0.01763577000000017,
+    "speedup_vs_single_jit": 2.1,
+    "batch_time": 0.052907310000000506
+  },
+  "memory_mb": {},
+  "_provenance": "ingested from pre-FFT-fix /tmp profiling artifacts"
+}