2026-09-22T14:51:10.281649Z INFO ipu_codegen.package.build{tile_count=1472 operations=3}:compile_runtime: close time.busy=205µs time.idle=7.08µs 2026-09-22T14:51:35.452243Z INFO ipu_codegen.package.build{tile_count=1472 operations=3}:construct_mid:gemm_candidates: close time.busy=25.2s time.idle=5.27µs 2026-09-22T14:52:19.852397Z INFO ipu_codegen.package.build{tile_count=1472 operations=3}:construct_mid:gemm_candidates: close time.busy=44.4s time.idle=4.39µs 2026-09-22T14:52:21.871305Z INFO ipu_codegen.package.build{tile_count=1472 operations=3}:construct_mid: close time.busy=71.6s time.idle=2.49µs 2026-09-22T14:52:21.955643Z INFO ipu_codegen.package.build{tile_count=1472 operations=3}: wrote planner memory profile path=artifacts/planner-memory-20260922/fp8-b1-n1-default/memory/baseline-1170923-0000.json scope="baseline" peak_step=4 effective_peak_bytes=388352 2026-09-22T14:52:23.027059Z INFO ipu_codegen.package.build{tile_count=1472 operations=3}:group_exchanges: close time.busy=88.2µs time.idle=2.78µs 2026-09-22T14:52:23.028066Z INFO ipu_codegen.package.build{tile_count=1472 operations=3}:merge_copies: close time.busy=803µs time.idle=1.02µs 2026-09-22T14:52:23.060542Z INFO ipu_codegen.package.build{tile_count=1472 operations=3}:eliminate: eliminated padding clears using kernel access contracts removed=0 2026-09-22T14:52:23.061704Z INFO ipu_codegen.package.build{tile_count=1472 operations=3}:eliminate: close time.busy=33.6ms time.idle=880ns 2026-09-22T14:52:23.105366Z INFO ipu_codegen.package.build{tile_count=1472 operations=3}:place_provisional_storage: close time.busy=43.1ms time.idle=2.56µs 2026-09-22T14:52:23.397961Z INFO ipu_codegen.package.build{tile_count=1472 operations=3}:schedule_provisional_exchanges: retrying exchange schedule with incremental instruction-alignment validation transfers=50933 2026-09-22T14:52:23.898594Z INFO ipu_codegen.package.build{tile_count=1472 operations=3}:schedule_provisional_exchanges: retrying exchange schedule with incremental instruction-alignment validation transfers=21111 2026-09-22T14:52:23.906841Z INFO ipu_codegen.package.build{tile_count=1472 operations=3}:schedule_provisional_exchanges: finished incremental instruction-alignment validation transfers=50933 elapsed_ms=508 success=true 2026-09-22T14:52:24.411825Z INFO ipu_codegen.package.build{tile_count=1472 operations=3}:schedule_provisional_exchanges: prepared large physical exchange phase phase=6 transfers=50933 maximum_identical_destinations=1 2026-09-22T14:52:24.411852Z INFO ipu_codegen.package.build{tile_count=1472 operations=3}:schedule_provisional_exchanges: optimized physical exchange schedule phase=6 initial_horizon=474 selected_horizon=474 endpoint_lower_bound=288 lower_bound_gap=186 selected_kind="full-duplex" neighborhood_improvements=0 2026-09-22T14:52:24.976117Z INFO ipu_codegen.package.build{tile_count=1472 operations=3}:schedule_provisional_exchanges: finished incremental instruction-alignment validation transfers=21111 elapsed_ms=1077 success=true 2026-09-22T14:52:26.841721Z INFO ipu_codegen.package.build{tile_count=1472 operations=3}:schedule_provisional_exchanges: compared ordinary and paired exchange schedules phase=3 candidates=20310 ordinary_horizon=8554 paired_horizon=8114 use_paired=true 2026-09-22T14:52:26.902857Z INFO ipu_codegen.package.build{tile_count=1472 operations=3}:schedule_provisional_exchanges: prepared large physical exchange phase phase=3 transfers=21111 maximum_identical_destinations=1 2026-09-22T14:52:26.902878Z INFO ipu_codegen.package.build{tile_count=1472 operations=3}:schedule_provisional_exchanges: optimized physical exchange schedule phase=3 initial_horizon=8114 selected_horizon=8114 endpoint_lower_bound=3362 lower_bound_gap=4752 selected_kind="full-duplex" neighborhood_improvements=0 2026-09-22T14:52:27.657336Z INFO ipu_codegen.package.build{tile_count=1472 operations=3}:schedule_provisional_exchanges: prepared large physical exchange phase phase=1 transfers=94930 maximum_identical_destinations=1 2026-09-22T14:52:27.657370Z INFO ipu_codegen.package.build{tile_count=1472 operations=3}:schedule_provisional_exchanges: optimized physical exchange schedule phase=1 initial_horizon=46966 selected_horizon=46966 endpoint_lower_bound=42048 lower_bound_gap=4918 selected_kind="full-duplex" neighborhood_improvements=0 2026-09-22T14:52:30.324321Z INFO ipu_codegen.package.build{tile_count=1472 operations=3}:schedule_provisional_exchanges: retrying exchange schedule with incremental instruction-alignment validation transfers=195879 2026-09-22T14:52:32.554752Z INFO ipu_codegen.package.build{tile_count=1472 operations=3}:schedule_provisional_exchanges: retrying exchange schedule with incremental instruction-alignment validation transfers=454896 2026-09-22T14:52:39.929416Z INFO ipu_codegen.package.build{tile_count=1472 operations=3}:schedule_provisional_exchanges: finished incremental instruction-alignment validation transfers=195879 elapsed_ms=9605 success=true 2026-09-22T14:52:43.815968Z INFO ipu_codegen.package.build{tile_count=1472 operations=3}:schedule_provisional_exchanges: finished incremental instruction-alignment validation transfers=454896 elapsed_ms=11261 success=true 2026-09-22T14:52:44.858114Z INFO ipu_codegen.package.build{tile_count=1472 operations=3}:schedule_provisional_exchanges: prepared large physical exchange phase phase=2 transfers=195879 maximum_identical_destinations=1 2026-09-22T14:52:44.858144Z INFO ipu_codegen.package.build{tile_count=1472 operations=3}:schedule_provisional_exchanges: optimized physical exchange schedule phase=2 initial_horizon=4058 selected_horizon=4058 endpoint_lower_bound=1168 lower_bound_gap=2890 selected_kind="full-duplex" neighborhood_improvements=0 2026-09-22T14:52:45.160799Z INFO ipu_codegen.package.build{tile_count=1472 operations=3}:schedule_provisional_exchanges: prepared large physical exchange phase phase=4 transfers=3096 maximum_identical_destinations=1 2026-09-22T14:52:45.160840Z INFO ipu_codegen.package.build{tile_count=1472 operations=3}:schedule_provisional_exchanges: optimized physical exchange schedule phase=4 initial_horizon=7943 selected_horizon=7943 endpoint_lower_bound=7680 lower_bound_gap=263 selected_kind="full-duplex" neighborhood_improvements=0 2026-09-22T14:52:45.598299Z INFO ipu_codegen.package.build{tile_count=1472 operations=3}:schedule_provisional_exchanges: retrying exchange schedule with incremental instruction-alignment validation transfers=28976 2026-09-22T14:52:46.631069Z INFO ipu_codegen.package.build{tile_count=1472 operations=3}:schedule_provisional_exchanges: finished incremental instruction-alignment validation transfers=28976 elapsed_ms=1032 success=true 2026-09-22T14:52:47.716616Z INFO ipu_codegen.package.build{tile_count=1472 operations=3}:schedule_provisional_exchanges: retrying exchange schedule with incremental instruction-alignment validation transfers=28976 2026-09-22T14:52:48.819009Z INFO ipu_codegen.package.build{tile_count=1472 operations=3}:schedule_provisional_exchanges: finished incremental instruction-alignment validation transfers=28976 elapsed_ms=1102 success=true 2026-09-22T14:52:49.400633Z INFO ipu_codegen.package.build{tile_count=1472 operations=3}:schedule_provisional_exchanges: compared ordinary and paired exchange schedules phase=0 candidates=18257 ordinary_horizon=17447 paired_horizon=18832 use_paired=false 2026-09-22T14:52:49.449623Z INFO ipu_codegen.package.build{tile_count=1472 operations=3}:schedule_provisional_exchanges: prepared large physical exchange phase phase=0 transfers=28976 maximum_identical_destinations=1 2026-09-22T14:52:49.449640Z INFO ipu_codegen.package.build{tile_count=1472 operations=3}:schedule_provisional_exchanges: optimized physical exchange schedule phase=0 initial_horizon=17447 selected_horizon=17447 endpoint_lower_bound=9288 lower_bound_gap=8159 selected_kind="full-duplex" neighborhood_improvements=0 2026-09-22T14:52:53.576695Z INFO ipu_codegen.package.build{tile_count=1472 operations=3}:schedule_provisional_exchanges: prepared large physical exchange phase phase=5 transfers=454896 maximum_identical_destinations=1 2026-09-22T14:52:53.576725Z INFO ipu_codegen.package.build{tile_count=1472 operations=3}:schedule_provisional_exchanges: optimized physical exchange schedule phase=5 initial_horizon=9273 selected_horizon=9273 endpoint_lower_bound=7488 lower_bound_gap=1785 selected_kind="full-duplex" neighborhood_improvements=0 2026-09-22T14:52:54.361296Z INFO ipu_codegen.package.build{tile_count=1472 operations=3}:schedule_provisional_exchanges: close time.busy=31.3s time.idle=4.93µs 2026-09-22T14:52:54.385075Z INFO ipu_codegen.package.build{tile_count=1472 operations=3}:plan_kernels: close time.busy=23.7ms time.idle=3.07µs 2026-09-22T14:52:54.386423Z INFO ipu_codegen.package.build{tile_count=1472 operations=3}:compile_kernels: close time.busy=1.33ms time.idle=966ns 2026-09-22T14:52:54.386643Z INFO ipu_codegen.package.build{tile_count=1472 operations=3}:link_runtime: close time.busy=137µs time.idle=987ns 2026-09-22T14:52:54.498527Z INFO ipu_codegen.package.build{tile_count=1472 operations=3}: exchange table and repeat patch storage exchange_table_bytes=12272 maximum_uncompressed_repeat_patch_bytes=0 maximum_elided_arithmetic_patch_bytes=0 2026-09-22T14:52:54.702109Z INFO ipu_codegen.package.build{tile_count=1472 operations=3}:size_tile_code: close time.busy=29.1ms time.idle=4.95µs 2026-09-22T14:52:54.702167Z INFO ipu_codegen.package.build{tile_count=1472 operations=3}: placing generated tile programs linked_end=319432 host_code_base=376832 host_code_bytes=620 generated_code_bytes=1152 exchange_table_bytes=12272 executable_ranges=[(377520, 524288)] 2026-09-22T14:52:54.702282Z INFO ipu_codegen.package.build{tile_count=1472 operations=3}: allocated package support memory linked_end=319432 profile_bytes=92 exchange_table_bytes=12272 host_code_bytes=620 host_data_bytes=404 generated_code_bytes=1152 code_address=377520 available_ranges=[(327680, 360448), (362612, 376832), (409600, 949168)] 2026-09-22T14:52:54.761844Z INFO ipu_codegen.package.build{tile_count=1472 operations=3}:place_storage: close time.busy=48.0ms time.idle=4.23µs 2026-09-22T14:52:54.894906Z INFO ipu_codegen.package.build{tile_count=1472 operations=3}:lower_exchanges: reused exchange optimization after validating relocated rows phase=4 2026-09-22T14:52:54.899946Z INFO ipu_codegen.package.build{tile_count=1472 operations=3}:lower_exchanges: prepared large physical exchange phase phase=4 transfers=3096 maximum_identical_destinations=1 2026-09-22T14:52:54.899960Z INFO ipu_codegen.package.build{tile_count=1472 operations=3}:lower_exchanges: optimized physical exchange schedule phase=4 initial_horizon=7943 selected_horizon=7943 endpoint_lower_bound=7680 lower_bound_gap=263 selected_kind="reused" neighborhood_improvements=0 2026-09-22T14:52:55.406581Z INFO ipu_codegen.package.build{tile_count=1472 operations=3}:lower_exchanges: reused exchange optimization after validating relocated rows phase=1 2026-09-22T14:52:55.426913Z INFO ipu_codegen.package.build{tile_count=1472 operations=3}:lower_exchanges: prepared large physical exchange phase phase=1 transfers=94930 maximum_identical_destinations=1 2026-09-22T14:52:55.426930Z INFO ipu_codegen.package.build{tile_count=1472 operations=3}:lower_exchanges: optimized physical exchange schedule phase=1 initial_horizon=46966 selected_horizon=46966 endpoint_lower_bound=42048 lower_bound_gap=4918 selected_kind="reused" neighborhood_improvements=0 2026-09-22T14:52:55.458871Z INFO ipu_codegen.package.build{tile_count=1472 operations=3}:lower_exchanges: reused exchange optimization after validating relocated rows phase=3 2026-09-22T14:52:55.479400Z INFO ipu_codegen.package.build{tile_count=1472 operations=3}:lower_exchanges: prepared large physical exchange phase phase=3 transfers=21111 maximum_identical_destinations=1 2026-09-22T14:52:55.479414Z INFO ipu_codegen.package.build{tile_count=1472 operations=3}:lower_exchanges: optimized physical exchange schedule phase=3 initial_horizon=8114 selected_horizon=8114 endpoint_lower_bound=3362 lower_bound_gap=4752 selected_kind="reused" neighborhood_improvements=0 2026-09-22T14:52:55.747019Z INFO ipu_codegen.package.build{tile_count=1472 operations=3}:lower_exchanges: retrying exchange schedule with incremental instruction-alignment validation transfers=50933 2026-09-22T14:52:56.244617Z INFO ipu_codegen.package.build{tile_count=1472 operations=3}:lower_exchanges: finished incremental instruction-alignment validation transfers=50933 elapsed_ms=497 success=true 2026-09-22T14:52:56.764105Z INFO ipu_codegen.package.build{tile_count=1472 operations=3}:lower_exchanges: prepared large physical exchange phase phase=6 transfers=50933 maximum_identical_destinations=1 2026-09-22T14:52:56.764133Z INFO ipu_codegen.package.build{tile_count=1472 operations=3}:lower_exchanges: optimized physical exchange schedule phase=6 initial_horizon=474 selected_horizon=474 endpoint_lower_bound=288 lower_bound_gap=186 selected_kind="full-duplex" neighborhood_improvements=0 2026-09-22T14:52:57.049503Z INFO ipu_codegen.package.build{tile_count=1472 operations=3}:lower_exchanges: retrying exchange schedule with incremental instruction-alignment validation transfers=28976 2026-09-22T14:52:58.399081Z INFO ipu_codegen.package.build{tile_count=1472 operations=3}:lower_exchanges: finished incremental instruction-alignment validation transfers=28976 elapsed_ms=1349 success=true 2026-09-22T14:52:59.179944Z INFO ipu_codegen.package.build{tile_count=1472 operations=3}:lower_exchanges: reused exchange optimization after validating relocated rows phase=2 2026-09-22T14:52:59.228610Z INFO ipu_codegen.package.build{tile_count=1472 operations=3}:lower_exchanges: prepared large physical exchange phase phase=2 transfers=195879 maximum_identical_destinations=1 2026-09-22T14:52:59.228630Z INFO ipu_codegen.package.build{tile_count=1472 operations=3}:lower_exchanges: optimized physical exchange schedule phase=2 initial_horizon=4058 selected_horizon=4058 endpoint_lower_bound=1168 lower_bound_gap=2890 selected_kind="reused" neighborhood_improvements=0 2026-09-22T14:52:59.875700Z INFO ipu_codegen.package.build{tile_count=1472 operations=3}:lower_exchanges: retrying exchange schedule with incremental instruction-alignment validation transfers=28976 2026-09-22T14:53:01.447374Z INFO ipu_codegen.package.build{tile_count=1472 operations=3}:lower_exchanges: finished incremental instruction-alignment validation transfers=28976 elapsed_ms=1571 success=true 2026-09-22T14:53:01.985015Z INFO ipu_codegen.package.build{tile_count=1472 operations=3}:lower_exchanges: compared ordinary and paired exchange schedules phase=0 candidates=7124 ordinary_horizon=10094 paired_horizon=11675 use_paired=false 2026-09-22T14:53:02.037544Z INFO ipu_codegen.package.build{tile_count=1472 operations=3}:lower_exchanges: prepared large physical exchange phase phase=0 transfers=28976 maximum_identical_destinations=1 2026-09-22T14:53:02.037566Z INFO ipu_codegen.package.build{tile_count=1472 operations=3}:lower_exchanges: optimized physical exchange schedule phase=0 initial_horizon=10094 selected_horizon=10094 endpoint_lower_bound=9288 lower_bound_gap=806 selected_kind="full-duplex" neighborhood_improvements=0 2026-09-22T14:53:05.663395Z INFO ipu_codegen.package.build{tile_count=1472 operations=3}:lower_exchanges: reused exchange optimization after validating relocated rows phase=5 2026-09-22T14:53:05.805548Z INFO ipu_codegen.package.build{tile_count=1472 operations=3}:lower_exchanges: prepared large physical exchange phase phase=5 transfers=454896 maximum_identical_destinations=1 2026-09-22T14:53:05.805568Z INFO ipu_codegen.package.build{tile_count=1472 operations=3}:lower_exchanges: optimized physical exchange schedule phase=5 initial_horizon=9273 selected_horizon=9273 endpoint_lower_bound=7488 lower_bound_gap=1785 selected_kind="reused" neighborhood_improvements=0 2026-09-22T14:53:06.415249Z INFO ipu_codegen.package.build{tile_count=1472 operations=3}:lower_exchanges: close time.busy=11.7s time.idle=3.62µs 2026-09-22T14:53:06.452595Z INFO ipu_codegen.package.build{tile_count=1472 operations=3}: costed final placed program final_cycles=250282 final_exchange=91122 2026-09-22T14:53:06.794465Z INFO ipu_codegen.package.build{tile_count=1472 operations=3}:prepare_tile_code: close time.busy=67.8ms time.idle=4.68µs 2026-09-22T14:53:06.816874Z INFO ipu_codegen.package.build{tile_count=1472 operations=3}:emit_tile_code: close time.busy=22.4ms time.idle=1.86µs 2026-09-22T14:53:06.933041Z INFO ipu_codegen.package.build{tile_count=1472 operations=3}:build_tile_images: close time.busy=111ms time.idle=2.27µs 2026-09-22T14:53:07.102380Z INFO ipu_codegen.package.build{tile_count=1472 operations=3}: wrote exact tile placement profile path=artifacts/planner-memory-20260922/fp8-b1-n1-default/memory/placement-1170923.html 2026-09-22T14:53:07.262678Z INFO ipu_codegen.package.build{tile_count=1472 operations=3}: close time.busy=117s time.idle=5.05µs 2026-09-22T14:53:07.265644Z INFO writing application package tiles=1472 2026-09-22T14:53:07.414529Z INFO application package written 2026-09-22T14:53:07.417713Z INFO reading application package 2026-09-22T14:53:07.543846Z INFO application package read tiles=1472 compiler=0.1.0 2026-09-22T14:53:07.548467Z INFO waiting for hardware ownership path=/tmp/ipu-stack-device.lock 2026-09-22T14:53:07.548479Z INFO acquired hardware ownership path=/tmp/ipu-stack-device.lock 2026-09-22T14:53:07.548483Z INFO opening IPU device path="/dev/ipu0" 2026-09-22T14:53:07.548508Z INFO IPU device attached path=/dev/ipu0 2026-09-22T14:53:07.548510Z INFO initializing IPU device 2026-09-22T14:53:07.556279Z INFO IPU device initialized 2026-09-22T14:53:07.556282Z INFO replaying device configuration register_writes=2003 2026-09-22T14:53:07.561162Z INFO loading application tile_count=1472 final_mark=23 2026-09-22T14:53:07.724943Z INFO application loaded and initial synchronization completed tile_count=1472 2026-09-22T14:53:08.449962Z INFO prepared reference inputs elapsed_ms=721 2026-09-22T14:53:08.727292Z INFO configured host reference BLAS threads=4 referencePrecision=device 2026-09-22T14:53:09.307460Z INFO evaluated host reference elapsed_ms=857 2026-09-22T14:53:09.313110Z INFO starting host exchange session startup_mark=23 2026-09-22T14:53:09.313123Z INFO attaching host exchange pages pages=2 2026-09-22T14:53:09.313913Z INFO host exchange session started 2026-09-22T14:53:09.314021Z INFO invoking host exchange call call="initialize" command=0 phases=4 2026-09-22T14:53:09.320802Z INFO invoking host exchange call call="run" command=0 phases=13 referenceInference=1 numericalTest=PASS profile=artifacts/planner-memory-20260922/fp8-b1-n1-default/profile.capnp tiles=1472 samples=26537 workload=siglip-mlp-f16-b1-t729-d1152-h4304-n1 benchmark=siglip-mlp-f16 batch=1 tokens=729 rows=729 dimension=1152 hiddenDimension=4304 blocks=1 biases=false activeTiles=1469 executionTiles=1472 inputBytes=839808 weightBytes=9934848 cycles=225672 minimumTileCycles=221646 deviceMicroseconds=150.448 effectiveGemmTflops=96.101 peakTflops=282.624 efficiencyPercent=34.00 maximumAbsoluteError=0.007385 package=artifacts/planner-memory-20260922/fp8-b1-n1-default/model.ipuexe tiles=1472 hardwareTest=PASS