2026-09-06T11:25:09.819067Z INFO applied GEMM plan constraint source_operation=0 matching_plans=2 constraint=GemmPlanConstraint { source_operation: 0, orientation: Normal, row_partitions: 4, column_partitions: 30, inner_partitions: 12, result_row_partitions: 12, result_column_partitions: 1, output_column_block: 144, weight_memory_class: Ipu21Interleaved, reduction_staging: Complete, local_weight_staging: Direct } 2026-09-06T11:25:10.181566Z INFO applied GEMM plan constraint source_operation=2 matching_plans=1 constraint=GemmPlanConstraint { source_operation: 2, orientation: Normal, row_partitions: 4, column_partitions: 24, inner_partitions: 15, result_row_partitions: 15, result_column_partitions: 1, output_column_block: 48, weight_memory_class: Ipu21Interleaved, reduction_staging: Complete, local_weight_staging: Direct } 2026-09-06T11:25:10.728909Z INFO retained operator-plan finalist finalist=0 values=7 operations=4 estimated_cycles=372329 estimated_exchange_cycles=147320 exchange_row_bytes=32492 peak_standard_bytes=147180 peak_interleaved_bytes=120832 peak_total_bytes=207596 maximum_standard_allocation_bytes=114688 active_tile_counts={1440} padding_group_counts={} conversion_sources=[Some(OperationId(2))] 2026-09-06T11:25:10.736060Z 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=1186 success=true 2026-09-06T11:25:14.090997Z INFO ipu_codegen.package.build{tile_count=1472 operations=3}:ipu_codegen.package.phase{phase="select_finalist"}: selected analytical operator plan estimated_cycles=327484 estimated_exchange_cycles=104024 2026-09-06T11:25:14.239751Z 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=72 2026-09-06T11:25:14.243511Z 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=3507 success=true 2026-09-06T11:25:14.243560Z INFO ipu_codegen.package.build{tile_count=1472 operations=3}: lowered graph for package construction logical_shards=25596 exchange_phases=4 2026-09-06T11:25:14.243909Z 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 2026-09-06T11:25:14.245503Z 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=1 success=true 2026-09-06T11:25:14.245740Z 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_k96_c128_r182_r184" target=ipu21 2026-09-06T11:25:14.307412Z 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/09138f41a356924f014940dc65508de73998e509791f06c6c220167d1ace81c7/kernel.o metadata=/srv/home/.cache/ipu-stack/kernels/09138f41a356924f014940dc65508de73998e509791f06c6c220167d1ace81c7/kernel.json 2026-09-06T11:25:14.307797Z 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_k96_c144_r182_r184" target=ipu21 2026-09-06T11:25:14.356880Z 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/a465636046ad06fe67743f1575f6d8819848551994552c838b3093ef2fa26fe2/kernel.o metadata=/srv/home/.cache/ipu-stack/kernels/a465636046ad06fe67743f1575f6d8819848551994552c838b3093ef2fa26fe2/kernel.json 2026-09-06T11:25:14.357813Z 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=112 success=true 2026-09-06T11:25:14.361119Z 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 2026-09-06T11:25:14.495880Z 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 2026-09-06T11:25:33.467124Z 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=149026 maximum_identical_destinations=1 2026-09-06T11:25:33.467227Z 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=23277 selected_horizon=20820 endpoint_lower_bound=13824 lower_bound_gap=6996 selected_kind="matching-waves" neighborhood_improvements=0 2026-09-06T11:25:33.576213Z 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=354 row_words=647 horizon=20820 2026-09-06T11:26:05.311937Z 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=6795 ordinary_horizon=47910 paired_horizon=52965 use_paired=false 2026-09-06T11:26:05.537536Z 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=14170 maximum_identical_destinations=1 2026-09-06T11:26:05.537564Z 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=47910 selected_horizon=47910 endpoint_lower_bound=33408 lower_bound_gap=14502 selected_kind="full-duplex" neighborhood_improvements=0 2026-09-06T11:26:05.654008Z 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=875 row_words=583 horizon=47910 2026-09-06T11:26:11.680365Z 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=63720 maximum_identical_destinations=1 2026-09-06T11:26:11.680409Z 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=7991 selected_horizon=7991 endpoint_lower_bound=4704 lower_bound_gap=3287 selected_kind="full-duplex" neighborhood_improvements=0 2026-09-06T11:26:11.747341Z 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=1058 row_words=294 horizon=7991 2026-09-06T11:26:11.768504Z 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=57272 success=true 2026-09-06T11:26:12.124339Z 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 2026-09-06T11:26:12.124378Z INFO ipu_codegen.package.build{tile_count=1472 operations=3}: allocated package support memory linked_end=319176 profile_bytes=132 exchange_table_bytes=5448 host_code_bytes=1316 generated_code_bytes=1636 code_address=320496 standard_ranges=[(409600, 524288)] 2026-09-06T11:26:12.226173Z 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=101 success=true 2026-09-06T11:26:12.234043Z 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 2026-09-06T11:26:13.940547Z 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 2026-09-06T11:26:14.006094Z 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=149026 maximum_identical_destinations=1 2026-09-06T11:26:14.006118Z 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=20820 selected_horizon=20820 endpoint_lower_bound=13824 lower_bound_gap=6996 selected_kind="reused" neighborhood_improvements=0 2026-09-06T11:26:14.181005Z 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=354 row_words=647 horizon=20820 2026-09-06T11:26:53.857308Z 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=6795 ordinary_horizon=47910 paired_horizon=52965 use_paired=false 2026-09-06T11:26:54.003263Z 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=14170 maximum_identical_destinations=1 2026-09-06T11:26:54.003289Z 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=47910 selected_horizon=47910 endpoint_lower_bound=33408 lower_bound_gap=14502 selected_kind="full-duplex" neighborhood_improvements=0 2026-09-06T11:26:54.073638Z 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=875 row_words=583 horizon=47910 2026-09-06T11:26:58.292156Z 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=63720 maximum_identical_destinations=1 2026-09-06T11:26:58.292190Z 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=7991 selected_horizon=7991 endpoint_lower_bound=4704 lower_bound_gap=3287 selected_kind="full-duplex" neighborhood_improvements=0 2026-09-06T11:26:58.332917Z 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=1058 row_words=294 horizon=7991 2026-09-06T11:26:58.347895Z 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=46121 success=true 2026-09-06T11:26:59.651441Z 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 2026-09-06T11:27:01.427259Z 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 2026-09-06T11:27:01.493958Z 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=149026 maximum_identical_destinations=1 2026-09-06T11:27:01.493982Z 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=20820 selected_horizon=20820 endpoint_lower_bound=13824 lower_bound_gap=6996 selected_kind="reused" neighborhood_improvements=0 2026-09-06T11:27:01.665459Z 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=354 row_words=647 horizon=20820 2026-09-06T11:27:52.851993Z 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=6795 ordinary_horizon=39996 paired_horizon=42708 use_paired=false 2026-09-06T11:27:52.996861Z 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=14170 maximum_identical_destinations=1 2026-09-06T11:27:52.996884Z 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=39996 selected_horizon=39996 endpoint_lower_bound=33408 lower_bound_gap=6588 selected_kind="full-duplex" neighborhood_improvements=0 2026-09-06T11:27:53.066872Z 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=875 row_words=583 horizon=39996 2026-09-06T11:27:57.275405Z 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=63720 maximum_identical_destinations=1 2026-09-06T11:27:57.275442Z 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=7991 selected_horizon=7991 endpoint_lower_bound=4704 lower_bound_gap=3287 selected_kind="full-duplex" neighborhood_improvements=0 2026-09-06T11:27:57.317071Z 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=1058 row_words=294 horizon=7991 2026-09-06T11:27:57.343597Z 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=127802769408 score=122593443840 baseline_cycles=83765 candidate_cycles=75851 row_bytes=5440 row_capacity=16320 accepted=true 2026-09-06T11:27:57.360531Z 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=59012 success=true 2026-09-06T11:27:57.471584Z 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 2026-09-06T11:27:57.484100Z 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=12 success=true 2026-09-06T11:27:57.588475Z 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=98 success=true 2026-09-06T11:27:57.644927Z INFO writing application package tiles=1472 2026-09-06T11:27:57.714850Z INFO application package written 2026-09-06T11:27:57.714899Z INFO reading application package 2026-09-06T11:27:57.829588Z INFO application package read tiles=1472 compiler=0.1.0 2026-09-06T11:27:57.832748Z INFO waiting for hardware ownership path=/srv/home/gc-sdk/ipu-stack/artifacts/layout-sweep/device.lock 2026-09-06T11:27:57.832764Z INFO acquired hardware ownership path=/srv/home/gc-sdk/ipu-stack/artifacts/layout-sweep/device.lock 2026-09-06T11:27:57.832776Z INFO opening IPU device path="/dev/ipu0" 2026-09-06T11:27:57.832807Z INFO IPU device attached path=/dev/ipu0 2026-09-06T11:27:57.832812Z INFO initializing IPU device 2026-09-06T11:27:57.840827Z INFO IPU device initialized 2026-09-06T11:27:57.840852Z INFO replaying device configuration register_writes=2003 2026-09-06T11:27:57.847640Z INFO loading application tile_count=1472 final_mark=23 2026-09-06T11:27:58.127306Z INFO application loaded and initial synchronization completed tile_count=1472 2026-09-06T11:27:59.154295Z INFO starting host exchange session startup_mark=23 2026-09-06T11:27:59.154336Z INFO attaching host exchange pages pages=2 2026-09-06T11:27:59.154978Z INFO host exchange session started 2026-09-06T11:27:59.155023Z INFO invoking streaming host exchange call call="initialize" command=0 phases=14 2026-09-06T11:27:59.160891Z INFO invoking streaming host exchange call call="run" command=0 phases=25 profile=artifacts/layout-sweep/up-05/execution.ipuprofile tiles=1472 samples=25414 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=1440 executionTiles=1472 inputBytes=50457600 weightBytes=19869696 cycles=229752 minimumTileCycles=224166 deviceMicroseconds=153.168 effectiveGemmTflops=94.394 peakTflops=282.624 efficiencyPercent=33.40 maximumAbsoluteError=0.011719 package=artifacts/layout-sweep/up-05/model.ipuexe tiles=1472 hardwareTest=PASS