2026-09-06T11:17:38.968168Z INFO applied GEMM plan constraint source_operation=0 matching_plans=1 constraint=GemmPlanConstraint { source_operation: 0, orientation: Normal, row_partitions: 8, column_partitions: 90, inner_partitions: 2, result_row_partitions: 2, result_column_partitions: 1, output_column_block: 48, weight_memory_class: Ipu21Interleaved, reduction_staging: Complete, local_weight_staging: Direct } 2026-09-06T11:17:39.395987Z 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 } 2026-09-06T11:17:39.938063Z INFO retained operator-plan finalist finalist=0 values=7 operations=4 estimated_cycles=329531 estimated_exchange_cycles=135800 exchange_row_bytes=29900 peak_standard_bytes=144588 peak_interleaved_bytes=138240 peak_total_bytes=282828 maximum_standard_allocation_bytes=114688 active_tile_counts={1440} padding_group_counts={} conversion_sources=[Some(OperationId(2))] 2026-09-06T11:17:39.942312Z 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=1267 success=true 2026-09-06T11:17:45.121658Z INFO ipu_codegen.package.build{tile_count=1472 operations=3}:ipu_codegen.package.phase{phase="select_finalist"}: selected analytical operator plan estimated_cycles=261009 estimated_exchange_cycles=79560 2026-09-06T11:17:45.205265Z 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 2026-09-06T11:17:45.207003Z 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=5264 success=true 2026-09-06T11:17:45.207023Z INFO ipu_codegen.package.build{tile_count=1472 operations=3}: lowered graph for package construction logical_shards=25776 exchange_phases=4 2026-09-06T11:17:45.207223Z 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 2026-09-06T11:17:45.207748Z 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 2026-09-06T11:17:45.207891Z 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_k576_c32_r90_r92" target=ipu21 2026-09-06T11:17:45.263187Z 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/fcd845e24b89bbc6ff8d99ada6211cdf89afe58e9633ced5ead69f4bfbb433d9/kernel.o metadata=/srv/home/.cache/ipu-stack/kernels/fcd845e24b89bbc6ff8d99ada6211cdf89afe58e9633ced5ead69f4bfbb433d9/kernel.json 2026-09-06T11:17:45.263464Z 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_k576_c48_r90_r92" target=ipu21 2026-09-06T11:17:45.311842Z 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/b7983b2c1c94f444b9c0fcf51a9d1f63822727840b73011d81b9ee0ab8886862/kernel.o metadata=/srv/home/.cache/ipu-stack/kernels/b7983b2c1c94f444b9c0fcf51a9d1f63822727840b73011d81b9ee0ab8886862/kernel.json 2026-09-06T11:17:45.312149Z 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=104 success=true 2026-09-06T11:17:45.315253Z 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 2026-09-06T11:17:45.439389Z 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=124 success=true 2026-09-06T11:17:45.557181Z 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=5380 maximum_identical_destinations=1 2026-09-06T11:17:45.557232Z 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=1288 selected_horizon=1256 endpoint_lower_bound=1152 lower_bound_gap=104 selected_kind="critical-neighborhood" neighborhood_improvements=1 2026-09-06T11:17:45.560629Z 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=1193 row_words=31 horizon=1256 2026-09-06T11:17:52.810780Z 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=3876 ordinary_horizon=50542 paired_horizon=61867 use_paired=false 2026-09-06T11:17:52.888331Z 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=5562 maximum_identical_destinations=1 2026-09-06T11:17:52.888358Z 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=50542 selected_horizon=50542 endpoint_lower_bound=33408 lower_bound_gap=17134 selected_kind="full-duplex" neighborhood_improvements=0 2026-09-06T11:17:52.928628Z 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=1149 row_words=223 horizon=50542 2026-09-06T11:17:58.934185Z 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 2026-09-06T11:17:58.934252Z 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 2026-09-06T11:17:58.994361Z 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 2026-09-06T11:17:59.013851Z 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=13574 success=true 2026-09-06T11:17:59.603898Z 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 2026-09-06T11:17:59.603935Z INFO ipu_codegen.package.build{tile_count=1472 operations=3}: allocated package support memory linked_end=319128 profile_bytes=136 exchange_table_bytes=2112 host_code_bytes=2860 generated_code_bytes=1612 code_address=321992 standard_ranges=[(409600, 524288)] 2026-09-06T11:17:59.714634Z 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=110 success=true 2026-09-06T11:17:59.728004Z 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 2026-09-06T11:17:59.760267Z 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 2026-09-06T11:17:59.762289Z 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=5380 maximum_identical_destinations=1 2026-09-06T11:17:59.762301Z 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=1256 selected_horizon=1256 endpoint_lower_bound=1152 lower_bound_gap=104 selected_kind="reused" neighborhood_improvements=0 2026-09-06T11:17:59.765990Z 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=1193 row_words=31 horizon=1256 2026-09-06T11:18:07.591313Z 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=3876 ordinary_horizon=50542 paired_horizon=61867 use_paired=false 2026-09-06T11:18:07.674195Z 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=5562 maximum_identical_destinations=1 2026-09-06T11:18:07.674265Z 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=50542 selected_horizon=50542 endpoint_lower_bound=33408 lower_bound_gap=17134 selected_kind="full-duplex" neighborhood_improvements=0 2026-09-06T11:18:07.716350Z 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=1149 row_words=223 horizon=50542 2026-09-06T11:18:14.367527Z 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 2026-09-06T11:18:14.367593Z 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 2026-09-06T11:18:14.429839Z 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 2026-09-06T11:18:14.455068Z 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=14740 success=true 2026-09-06T11:18:15.806804Z 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 2026-09-06T11:18:15.844785Z 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 2026-09-06T11:18:15.847407Z 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=5380 maximum_identical_destinations=1 2026-09-06T11:18:15.847430Z 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=1256 selected_horizon=1256 endpoint_lower_bound=1152 lower_bound_gap=104 selected_kind="reused" neighborhood_improvements=0 2026-09-06T11:18:15.851327Z 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=1193 row_words=31 horizon=1256 2026-09-06T11:18:24.167247Z 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=3876 ordinary_horizon=76021 paired_horizon=59011 use_paired=true 2026-09-06T11:18:24.252359Z 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=5562 maximum_identical_destinations=1 2026-09-06T11:18:24.252398Z 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=59011 selected_horizon=59011 endpoint_lower_bound=23544 lower_bound_gap=35467 selected_kind="full-duplex" neighborhood_improvements=0 2026-09-06T11:18:24.295498Z 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=1384 row_words=271 horizon=59011 2026-09-06T11:18:31.928375Z 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 2026-09-06T11:18:31.928462Z 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 2026-09-06T11:18:31.991507Z 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 2026-09-06T11:18:32.028385Z INFO ipu_codegen.package.build{tile_count=1472 operations=3}:ipu_codegen.package.phase{phase="optimize_exchange_placement"}: evaluated exchange placement finalist offset=4096 baseline_score=448685899776 score=173497614336 baseline_cycles=73738 candidate_cycles=82207 row_bytes=2136 row_capacity=16320 accepted=false 2026-09-06T11:18:32.041837Z 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=17586 success=true 2026-09-06T11:18:32.357448Z 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 2026-09-06T11:18:32.370174Z 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 2026-09-06T11:18:32.480033Z 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=103 success=true 2026-09-06T11:18:32.548630Z INFO writing application package tiles=1472 2026-09-06T11:18:32.613096Z INFO application package written 2026-09-06T11:18:32.617068Z INFO reading application package 2026-09-06T11:18:32.728988Z INFO application package read tiles=1472 compiler=0.1.0 2026-09-06T11:18:32.732722Z INFO opening IPU device path="/dev/ipu0" 2026-09-06T11:18:32.732780Z INFO IPU device attached path=/dev/ipu0 2026-09-06T11:18:32.732787Z INFO initializing IPU device 2026-09-06T11:18:32.740789Z INFO IPU device initialized 2026-09-06T11:18:32.740797Z INFO replaying device configuration register_writes=2003 2026-09-06T11:18:32.751090Z INFO loading application tile_count=1472 final_mark=23 2026-09-06T11:18:33.076762Z INFO application loaded and initial synchronization completed tile_count=1472 2026-09-06T11:18:35.775004Z INFO starting host exchange session startup_mark=23 2026-09-06T11:18:35.775084Z INFO attaching host exchange pages pages=2 2026-09-06T11:18:35.775749Z INFO host exchange session started 2026-09-06T11:18:35.775852Z INFO invoking streaming host exchange call call="initialize" command=0 phases=28 2026-09-06T11:18:35.785453Z INFO invoking streaming host exchange call call="run" command=0 phases=59 profile=artifacts/layout-sweep/up-00/execution.ipuprofile tiles=1472 samples=25676 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=151372800 weightBytes=19869696 cycles=229164 minimumTileCycles=212256 deviceMicroseconds=152.776 effectiveGemmTflops=94.636 peakTflops=282.624 efficiencyPercent=33.48 maximumAbsoluteError=0.011719 package=artifacts/layout-sweep/up-00/model.ipuexe tiles=1472 hardwareTest=PASS