2026-09-06T11:45:43.347427Z 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 } 2026-09-06T11:45:43.814091Z INFO applied GEMM plan constraint source_operation=2 matching_plans=2 constraint=GemmPlanConstraint { source_operation: 2, orientation: Normal, row_partitions: 16, column_partitions: 23, inner_partitions: 4, result_row_partitions: 4, result_column_partitions: 1, output_column_block: 64, weight_memory_class: Ipu21Interleaved, reduction_staging: Complete, local_weight_staging: Direct } 2026-09-06T11:45:44.004446Z INFO retained operator-plan finalist finalist=0 values=7 operations=4 estimated_cycles=428231 estimated_exchange_cycles=181080 exchange_row_bytes=40088 peak_standard_bytes=154776 peak_interleaved_bytes=278528 peak_total_bytes=433304 maximum_standard_allocation_bytes=114688 active_tile_counts={1472} padding_group_counts={} conversion_sources=[Some(OperationId(2))] 2026-09-06T11:45:44.007653Z 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=951 success=true 2026-09-06T11:45:52.425326Z INFO ipu_codegen.package.build{tile_count=1472 operations=3}:ipu_codegen.package.phase{phase="select_finalist"}: selected analytical operator plan estimated_cycles=327786 estimated_exchange_cycles=100584 2026-09-06T11:45:52.458999Z 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=345 2026-09-06T11:45:52.460587Z 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=8452 success=true 2026-09-06T11:45:52.460606Z INFO ipu_codegen.package.build{tile_count=1472 operations=3}: lowered graph for package construction logical_shards=27002 exchange_phases=4 2026-09-06T11:45:52.460796Z 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 2026-09-06T11:45:52.461419Z 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 2026-09-06T11:45:52.461690Z INFO ipu_codegen.package.build{tile_count=1472 operations=3}:ipu_codegen.package.phase{phase="compile_kernels"}: compiling kernel source=/srv/home/gc-sdk/ipu-stack/crates/ipu-tests/../../device/gemm_f16_amp.S name="gemm_f16_interleaved_k1088_c48_r44_r46" target=ipu21 2026-09-06T11:45:52.514850Z INFO ipu_codegen.package.build{tile_count=1472 operations=3}:ipu_codegen.package.phase{phase="compile_kernels"}: kernel artifact created object=/srv/home/.cache/ipu-stack/kernels/43fdef9bf3eea5a7f2f13ab980a4f60032d4cbc89082a3808960927a21da9198/kernel.o metadata=/srv/home/.cache/ipu-stack/kernels/43fdef9bf3eea5a7f2f13ab980a4f60032d4cbc89082a3808960927a21da9198/kernel.json 2026-09-06T11:45:52.515174Z INFO ipu_codegen.package.build{tile_count=1472 operations=3}:ipu_codegen.package.phase{phase="compile_kernels"}: compiling kernel source=/srv/home/gc-sdk/ipu-stack/crates/ipu-tests/../../device/gemm_f16_amp.S name="gemm_f16_interleaved_k1088_c64_r44_r46" target=ipu21 2026-09-06T11:45:52.566621Z INFO ipu_codegen.package.build{tile_count=1472 operations=3}:ipu_codegen.package.phase{phase="compile_kernels"}: kernel artifact created object=/srv/home/.cache/ipu-stack/kernels/896c1c8c4807d36f6c8caf0e9857c74867ef4a56482144962aac4761a583d367/kernel.o metadata=/srv/home/.cache/ipu-stack/kernels/896c1c8c4807d36f6c8caf0e9857c74867ef4a56482144962aac4761a583d367/kernel.json 2026-09-06T11:45:52.566963Z 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=105 success=true 2026-09-06T11:45:52.569964Z 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 2026-09-06T11:45:52.693186Z 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=123 success=true 2026-09-06T11:45:53.061141Z 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 2026-09-06T11:45:53.061169Z 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 2026-09-06T11:45:53.071589Z 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 2026-09-06T11:45:53.862683Z 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=53 ordinary_horizon=65118 paired_horizon=61163 use_paired=true 2026-09-06T11:45:53.887000Z 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=2347 maximum_identical_destinations=1 2026-09-06T11:45:53.887021Z 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=67962 selected_horizon=61163 endpoint_lower_bound=58736 lower_bound_gap=2427 selected_kind="critical-neighborhood" neighborhood_improvements=2 2026-09-06T11:45:53.901711Z 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=1425 row_words=170 horizon=61163 2026-09-06T11:45:54.582883Z 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=18432 maximum_identical_destinations=1 2026-09-06T11:45:54.582917Z 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=2183 selected_horizon=2183 endpoint_lower_bound=1536 lower_bound_gap=647 selected_kind="full-duplex" neighborhood_improvements=0 2026-09-06T11:45:54.598512Z 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=966 row_words=98 horizon=2183 2026-09-06T11:45:54.604100Z 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=1910 success=true 2026-09-06T11:45:55.199195Z 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=5 success=true 2026-09-06T11:45:55.199249Z INFO ipu_codegen.package.build{tile_count=1472 operations=3}: allocated package support memory linked_end=321080 profile_bytes=332 exchange_table_bytes=1100 host_code_bytes=4132 generated_code_bytes=2732 code_address=409600 standard_ranges=[(425984, 524288)] 2026-09-06T11:45:55.306299Z 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=107 success=true 2026-09-06T11:45:55.317396Z 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 2026-09-06T11:45:55.414566Z 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 2026-09-06T11:45:55.420572Z 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 2026-09-06T11:45:55.420587Z 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 2026-09-06T11:45:55.430293Z 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 2026-09-06T11:45:56.107952Z INFO ipu_codegen.package.build{tile_count=1472 operations=3}:ipu_codegen.package.phase{phase="lower_exchanges"}: compared ordinary and paired exchange schedules phase=2 candidates=39 ordinary_horizon=65118 paired_horizon=61163 use_paired=true 2026-09-06T11:45:56.133632Z 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=2347 maximum_identical_destinations=1 2026-09-06T11:45:56.133651Z 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=65049 selected_horizon=61163 endpoint_lower_bound=58736 lower_bound_gap=2427 selected_kind="critical-neighborhood" neighborhood_improvements=1 2026-09-06T11:45:56.145247Z 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=1404 row_words=173 horizon=61163 2026-09-06T11:45:56.865054Z 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=18432 maximum_identical_destinations=1 2026-09-06T11:45:56.865087Z 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=2183 selected_horizon=2183 endpoint_lower_bound=1536 lower_bound_gap=647 selected_kind="full-duplex" neighborhood_improvements=0 2026-09-06T11:45:56.880743Z 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=966 row_words=98 horizon=2183 2026-09-06T11:45:56.886670Z 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=1580 success=true 2026-09-06T11:45:58.006413Z 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 2026-09-06T11:45:58.095886Z 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 2026-09-06T11:45:58.102017Z 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 2026-09-06T11:45:58.102032Z 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 2026-09-06T11:45:58.112176Z 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 2026-09-06T11:45:58.904710Z 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=40 ordinary_horizon=65118 paired_horizon=63699 use_paired=true 2026-09-06T11:45:58.929901Z 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=2347 maximum_identical_destinations=1 2026-09-06T11:45:58.929918Z 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=65049 selected_horizon=63699 endpoint_lower_bound=58736 lower_bound_gap=4963 selected_kind="critical-neighborhood" neighborhood_improvements=1 2026-09-06T11:45:58.941302Z 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=1404 row_words=174 horizon=63699 2026-09-06T11:45:59.661768Z 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=18432 maximum_identical_destinations=1 2026-09-06T11:45:59.661795Z 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=2183 selected_horizon=2183 endpoint_lower_bound=1536 lower_bound_gap=647 selected_kind="full-duplex" neighborhood_improvements=0 2026-09-06T11:45:59.677279Z 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=966 row_words=98 horizon=2183 2026-09-06T11:45:59.688015Z 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=641471152128 score=609471627264 baseline_cycles=74173 candidate_cycles=76709 row_bytes=1116 row_capacity=16320 accepted=false 2026-09-06T11:45:59.694429Z 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=2807 success=true 2026-09-06T11:46:00.003109Z 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=14 success=true 2026-09-06T11:46:00.014500Z 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 2026-09-06T11:46:00.138312Z 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=119 success=true 2026-09-06T11:46:00.198252Z INFO writing application package tiles=1472 2026-09-06T11:46:00.236059Z INFO application package written 2026-09-06T11:46:00.237916Z INFO reading application package 2026-09-06T11:46:00.330677Z INFO application package read tiles=1472 compiler=0.1.0 2026-09-06T11:46:00.333370Z INFO waiting for hardware ownership path=/srv/home/gc-sdk/ipu-stack/artifacts/layout-sweep/device.lock 2026-09-06T11:46:00.333381Z INFO acquired hardware ownership path=/srv/home/gc-sdk/ipu-stack/artifacts/layout-sweep/device.lock 2026-09-06T11:46:00.333386Z INFO opening IPU device path="/dev/ipu0" 2026-09-06T11:46:00.333418Z INFO IPU device attached path=/dev/ipu0 2026-09-06T11:46:00.333421Z INFO initializing IPU device 2026-09-06T11:46:00.341259Z INFO IPU device initialized 2026-09-06T11:46:00.341264Z INFO replaying device configuration register_writes=2003 2026-09-06T11:46:00.350351Z INFO loading application tile_count=1472 final_mark=23 2026-09-06T11:46:00.584167Z INFO application loaded and initial synchronization completed tile_count=1472 2026-09-06T11:46:03.466718Z INFO starting host exchange session startup_mark=23 2026-09-06T11:46:03.466748Z INFO attaching host exchange pages pages=2 2026-09-06T11:46:03.467495Z INFO host exchange session started 2026-09-06T11:46:03.467594Z INFO invoking streaming host exchange call call="initialize" command=0 phases=68 2026-09-06T11:46:03.487227Z INFO invoking streaming host exchange call call="run" command=0 phases=59 profile=artifacts/layout-sweep/joint-result-rows-4/execution.ipuprofile tiles=1472 samples=27055 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=1472 executionTiles=1472 inputBytes=154736640 weightBytes=19943424 cycles=270600 minimumTileCycles=260304 deviceMicroseconds=180.400 effectiveGemmTflops=80.145 peakTflops=282.624 efficiencyPercent=28.36 maximumAbsoluteError=0.017578 package=artifacts/layout-sweep/joint-result-rows-4/model.ipuexe tiles=1472 hardwareTest=PASS