2026-09-06T11:15:42.054762Z  INFO applied GEMM plan constraint source_operation=0 matching_plans=1 constraint=GemmPlanConstraint { source_operation: 0, orientation: Normal, row_partitions: 4, column_partitions: 92, inner_partitions: 4, result_row_partitions: 4, result_column_partitions: 1, output_column_block: 48, weight_memory_class: Ipu21Interleaved, reduction_staging: Complete, local_weight_staging: Direct }
2026-09-06T11:15:42.586724Z  INFO applied GEMM plan constraint source_operation=2 matching_plans=2 constraint=GemmPlanConstraint { source_operation: 2, orientation: Normal, row_partitions: 3, column_partitions: 18, inner_partitions: 27, result_row_partitions: 27, result_column_partitions: 1, output_column_block: 64, weight_memory_class: Ipu21Standard, reduction_staging: Complete, local_weight_staging: Direct }
2026-09-06T11:15:43.365006Z  INFO retained operator-plan finalist finalist=0 values=7 operations=4 estimated_cycles=303439 estimated_exchange_cycles=111000 exchange_row_bytes=24320 peak_standard_bytes=159488 peak_interleaved_bytes=60416 peak_total_bytes=219904 maximum_standard_allocation_bytes=114688 active_tile_counts={1458, 1472} padding_group_counts={} conversion_sources=[Some(OperationId(2))]
2026-09-06T11:15:43.372695Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:ipu_codegen.package.phase{phase="lower_mid"}: package build phase finished phase="lower_mid" elapsed_ms=1614 success=true
2026-09-06T11:15:46.123297Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:ipu_codegen.package.phase{phase="select_finalist"}: selected analytical operator plan estimated_cycles=275103 estimated_exchange_cycles=85384
2026-09-06T11:15:46.184636Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:ipu_codegen.package.phase{phase="select_finalist"}: eliminated finite GEMM padding clears using load-time SRAM initialization removed=36
2026-09-06T11:15:46.186803Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:ipu_codegen.package.phase{phase="select_finalist"}: package build phase finished phase="select_finalist" elapsed_ms=2813 success=true
2026-09-06T11:15:46.186837Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}: lowered graph for package construction logical_shards=26426 exchange_phases=4
2026-09-06T11:15:46.187025Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:ipu_codegen.package.phase{phase="compile_runtime"}: package build phase finished phase="compile_runtime" elapsed_ms=0 success=true
2026-09-06T11:15:46.187758Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:ipu_codegen.package.phase{phase="plan_kernels"}: package build phase finished phase="plan_kernels" elapsed_ms=0 success=true
2026-09-06T11:15:46.187975Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:ipu_codegen.package.phase{phase="compile_kernels"}: package build phase finished phase="compile_kernels" elapsed_ms=0 success=true
2026-09-06T11:15:46.190011Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:ipu_codegen.package.phase{phase="link_runtime"}: package build phase finished phase="link_runtime" elapsed_ms=0 success=true
2026-09-06T11:15:46.324926Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:ipu_codegen.package.phase{phase="plan_exchange_storage"}: package build phase finished phase="plan_exchange_storage" elapsed_ms=134 success=true
2026-09-06T11:15:46.591087Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:ipu_codegen.package.phase{phase="lower_exchanges_provisional"}: prepared large physical exchange phase phase=1 transfers=15064 maximum_identical_destinations=1
2026-09-06T11:15:46.591126Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:ipu_codegen.package.phase{phase="lower_exchanges_provisional"}: optimized physical exchange schedule phase=1 initial_horizon=4180 selected_horizon=3779 endpoint_lower_bound=3408 lower_bound_gap=371 selected_kind="matching-waves" neighborhood_improvements=0
2026-09-06T11:15:46.598637Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:ipu_codegen.package.phase{phase="lower_exchanges_provisional"}: finished large physical exchange phase phase=1 tile=690 row_words=75 horizon=3779
2026-09-06T11:15:49.778517Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:ipu_codegen.package.phase{phase="lower_exchanges_provisional"}: compared ordinary and paired exchange schedules phase=2 candidates=4417 ordinary_horizon=29099 paired_horizon=23031 use_paired=true
2026-09-06T11:15:49.850027Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:ipu_codegen.package.phase{phase="lower_exchanges_provisional"}: prepared large physical exchange phase phase=2 transfers=5814 maximum_identical_destinations=1
2026-09-06T11:15:49.850067Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:ipu_codegen.package.phase{phase="lower_exchanges_provisional"}: optimized physical exchange schedule phase=2 initial_horizon=25465 selected_horizon=23031 endpoint_lower_bound=17280 lower_bound_gap=5751 selected_kind="critical-neighborhood" neighborhood_improvements=1
2026-09-06T11:15:49.888996Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:ipu_codegen.package.phase{phase="lower_exchanges_provisional"}: finished large physical exchange phase phase=2 tile=691 row_words=230 horizon=23031
2026-09-06T11:16:24.906707Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:ipu_codegen.package.phase{phase="lower_exchanges_provisional"}: prepared large physical exchange phase phase=3 transfers=155880 maximum_identical_destinations=1
2026-09-06T11:16:24.906754Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:ipu_codegen.package.phase{phase="lower_exchanges_provisional"}: optimized physical exchange schedule phase=3 initial_horizon=15700 selected_horizon=15700 endpoint_lower_bound=8640 lower_bound_gap=7060 selected_kind="full-duplex" neighborhood_improvements=0
2026-09-06T11:16:25.016908Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:ipu_codegen.package.phase{phase="lower_exchanges_provisional"}: finished large physical exchange phase phase=3 tile=447 row_words=633 horizon=15700
2026-09-06T11:16:25.050461Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:ipu_codegen.package.phase{phase="lower_exchanges_provisional"}: package build phase finished phase="lower_exchanges_provisional" elapsed_ms=38725 success=true
2026-09-06T11:16:25.428063Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:ipu_codegen.package.phase{phase="size_tile_code"}: package build phase finished phase="size_tile_code" elapsed_ms=4 success=true
2026-09-06T11:16:25.428096Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}: allocated package support memory linked_end=319176 profile_bytes=104 exchange_table_bytes=3388 host_code_bytes=2396 generated_code_bytes=1440 code_address=321576 standard_ranges=[(409600, 524288)]
2026-09-06T11:16:25.539246Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:ipu_codegen.package.phase{phase="place_storage"}: package build phase finished phase="place_storage" elapsed_ms=111 success=true
2026-09-06T11:16:25.546863Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:ipu_codegen.package.phase{phase="lower_exchanges"}: reused exchange optimization after validating relocated rows phase=0
2026-09-06T11:16:25.640284Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:ipu_codegen.package.phase{phase="lower_exchanges"}: reused exchange optimization after validating relocated rows phase=1
2026-09-06T11:16:25.646260Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:ipu_codegen.package.phase{phase="lower_exchanges"}: prepared large physical exchange phase phase=1 transfers=15064 maximum_identical_destinations=1
2026-09-06T11:16:25.646277Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:ipu_codegen.package.phase{phase="lower_exchanges"}: optimized physical exchange schedule phase=1 initial_horizon=3779 selected_horizon=3779 endpoint_lower_bound=3408 lower_bound_gap=371 selected_kind="reused" neighborhood_improvements=0
2026-09-06T11:16:25.656297Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:ipu_codegen.package.phase{phase="lower_exchanges"}: finished large physical exchange phase phase=1 tile=690 row_words=75 horizon=3779
2026-09-06T11:16:26.265139Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:ipu_codegen.package.phase{phase="lower_exchanges"}: reused exchange optimization after validating relocated rows phase=2
2026-09-06T11:16:26.284218Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:ipu_codegen.package.phase{phase="lower_exchanges"}: prepared large physical exchange phase phase=2 transfers=5814 maximum_identical_destinations=1
2026-09-06T11:16:26.284235Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:ipu_codegen.package.phase{phase="lower_exchanges"}: optimized physical exchange schedule phase=2 initial_horizon=23031 selected_horizon=23031 endpoint_lower_bound=17280 lower_bound_gap=5751 selected_kind="reused" neighborhood_improvements=0
2026-09-06T11:16:26.319290Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:ipu_codegen.package.phase{phase="lower_exchanges"}: finished large physical exchange phase phase=2 tile=691 row_words=230 horizon=23031
2026-09-06T11:16:57.745614Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:ipu_codegen.package.phase{phase="lower_exchanges"}: prepared large physical exchange phase phase=3 transfers=155880 maximum_identical_destinations=1
2026-09-06T11:16:57.745657Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:ipu_codegen.package.phase{phase="lower_exchanges"}: optimized physical exchange schedule phase=3 initial_horizon=15700 selected_horizon=15700 endpoint_lower_bound=8640 lower_bound_gap=7060 selected_kind="full-duplex" neighborhood_improvements=0
2026-09-06T11:16:57.860941Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:ipu_codegen.package.phase{phase="lower_exchanges"}: finished large physical exchange phase phase=3 tile=447 row_words=633 horizon=15700
2026-09-06T11:16:57.896065Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:ipu_codegen.package.phase{phase="lower_exchanges"}: package build phase finished phase="lower_exchanges" elapsed_ms=32356 success=true
2026-09-06T11:16:59.109119Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:ipu_codegen.package.phase{phase="optimize_exchange_placement"}: reused exchange optimization after validating relocated rows phase=0
2026-09-06T11:16:59.204564Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:ipu_codegen.package.phase{phase="optimize_exchange_placement"}: reused exchange optimization after validating relocated rows phase=1
2026-09-06T11:16:59.210652Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:ipu_codegen.package.phase{phase="optimize_exchange_placement"}: prepared large physical exchange phase phase=1 transfers=15064 maximum_identical_destinations=1
2026-09-06T11:16:59.210669Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:ipu_codegen.package.phase{phase="optimize_exchange_placement"}: optimized physical exchange schedule phase=1 initial_horizon=3779 selected_horizon=3779 endpoint_lower_bound=3408 lower_bound_gap=371 selected_kind="reused" neighborhood_improvements=0
2026-09-06T11:16:59.220804Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:ipu_codegen.package.phase{phase="optimize_exchange_placement"}: finished large physical exchange phase phase=1 tile=690 row_words=75 horizon=3779
2026-09-06T11:17:01.718174Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:ipu_codegen.package.phase{phase="optimize_exchange_placement"}: compared ordinary and paired exchange schedules phase=2 candidates=4417 ordinary_horizon=28446 paired_horizon=22392 use_paired=true
2026-09-06T11:17:01.784474Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:ipu_codegen.package.phase{phase="optimize_exchange_placement"}: prepared large physical exchange phase phase=2 transfers=5814 maximum_identical_destinations=1
2026-09-06T11:17:01.784503Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:ipu_codegen.package.phase{phase="optimize_exchange_placement"}: optimized physical exchange schedule phase=2 initial_horizon=22392 selected_horizon=22392 endpoint_lower_bound=17280 lower_bound_gap=5112 selected_kind="full-duplex" neighborhood_improvements=0
2026-09-06T11:17:01.822410Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:ipu_codegen.package.phase{phase="optimize_exchange_placement"}: finished large physical exchange phase phase=2 tile=691 row_words=240 horizon=22392
2026-09-06T11:17:34.978140Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:ipu_codegen.package.phase{phase="optimize_exchange_placement"}: prepared large physical exchange phase phase=3 transfers=155880 maximum_identical_destinations=1
2026-09-06T11:17:34.978194Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:ipu_codegen.package.phase{phase="optimize_exchange_placement"}: optimized physical exchange schedule phase=3 initial_horizon=15700 selected_horizon=15700 endpoint_lower_bound=8640 lower_bound_gap=7060 selected_kind="full-duplex" neighborhood_improvements=0
2026-09-06T11:17:35.165236Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:ipu_codegen.package.phase{phase="optimize_exchange_placement"}: finished large physical exchange phase phase=3 tile=447 row_words=633 horizon=15700
2026-09-06T11:17:35.231758Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:ipu_codegen.package.phase{phase="optimize_exchange_placement"}: evaluated exchange placement finalist offset=28672 baseline_score=127530434560 score=100812718080 baseline_cycles=49558 candidate_cycles=48919 row_bytes=3388 row_capacity=16320 accepted=true
2026-09-06T11:17:35.261461Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:ipu_codegen.package.phase{phase="optimize_exchange_placement"}: package build phase finished phase="optimize_exchange_placement" elapsed_ms=37365 success=true
2026-09-06T11:17:35.578990Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:ipu_codegen.package.phase{phase="prepare_tile_code"}: package build phase finished phase="prepare_tile_code" elapsed_ms=8 success=true
2026-09-06T11:17:35.590234Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:ipu_codegen.package.phase{phase="emit_tile_code"}: package build phase finished phase="emit_tile_code" elapsed_ms=11 success=true
2026-09-06T11:17:35.696545Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:ipu_codegen.package.phase{phase="build_tile_images"}: package build phase finished phase="build_tile_images" elapsed_ms=101 success=true
2026-09-06T11:17:35.775024Z  INFO writing application package tiles=1472
2026-09-06T11:17:35.849927Z  INFO application package written
2026-09-06T11:17:35.851529Z  INFO reading application package
2026-09-06T11:17:35.959654Z  INFO application package read tiles=1472 compiler=0.1.0
2026-09-06T11:17:35.962408Z  INFO opening IPU device path="/dev/ipu0"
2026-09-06T11:17:35.962447Z  INFO IPU device attached path=/dev/ipu0
2026-09-06T11:17:35.962452Z  INFO initializing IPU device
2026-09-06T11:17:35.970715Z  INFO IPU device initialized
2026-09-06T11:17:35.970740Z  INFO replaying device configuration register_writes=2003
2026-09-06T11:17:35.979493Z  INFO loading application tile_count=1472 final_mark=23
2026-09-06T11:17:36.250008Z  INFO application loaded and initial synchronization completed tile_count=1472
2026-09-06T11:17:38.098511Z  INFO starting host exchange session startup_mark=23
2026-09-06T11:17:38.098549Z  INFO attaching host exchange pages pages=2
2026-09-06T11:17:38.099076Z  INFO host exchange session started
2026-09-06T11:17:38.099120Z  INFO invoking streaming host exchange call call="initialize" command=0 phases=14
2026-09-06T11:17:38.105438Z  INFO invoking streaming host exchange call call="run" command=0 phases=59
profile=artifacts/layout-sweep/default/execution.ipuprofile tiles=1472 samples=24115
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=1458 executionTiles=1472 inputBytes=154736640 weightBytes=19869696 cycles=197394 minimumTileCycles=164634 deviceMicroseconds=131.596 effectiveGemmTflops=109.868 peakTflops=282.624 efficiencyPercent=38.87 maximumAbsoluteError=0.011719
package=artifacts/layout-sweep/default/model.ipuexe tiles=1472 hardwareTest=PASS