2026-09-07T09:53:41.010293Z  INFO retained operator-plan finalist finalist=0 values=7 operations=4 estimated_cycles=241328 estimated_exchange_cycles=106840 exchange_row_bytes=23384 peak_standard_bytes=224088 peak_interleaved_bytes=53248 peak_total_bytes=272216 maximum_standard_allocation_bytes=180224 active_tile_counts={1458, 1470} padding_group_counts={} conversion_sources=[Some(OperationId(2))]
2026-09-07T09:53:41.016273Z  INFO retained operator-plan finalist finalist=1 values=7 operations=4 estimated_cycles=243004 estimated_exchange_cycles=113400 exchange_row_bytes=24860 peak_standard_bytes=160028 peak_interleaved_bytes=79360 peak_total_bytes=239388 maximum_standard_allocation_bytes=114688 active_tile_counts={1458, 1472} padding_group_counts={} conversion_sources=[Some(OperationId(2))]
2026-09-07T09:53:41.022276Z  INFO retained operator-plan finalist finalist=2 values=7 operations=4 estimated_cycles=251408 estimated_exchange_cycles=111480 exchange_row_bytes=24428 peak_standard_bytes=227180 peak_interleaved_bytes=55296 peak_total_bytes=275308 maximum_standard_allocation_bytes=180224 active_tile_counts={1425, 1470} padding_group_counts={} conversion_sources=[Some(OperationId(2))]
2026-09-07T09:53:41.028187Z  INFO retained operator-plan finalist finalist=3 values=7 operations=4 estimated_cycles=251538 estimated_exchange_cycles=115160 exchange_row_bytes=25256 peak_standard_bytes=225960 peak_interleaved_bytes=53248 peak_total_bytes=274088 maximum_standard_allocation_bytes=180224 active_tile_counts={1458, 1470} padding_group_counts={} conversion_sources=[Some(OperationId(2))]
2026-09-07T09:53:41.371252Z  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=12003 success=true
2026-09-07T09:53:42.906361Z  INFO eliminated finite GEMM padding clears using load-time SRAM initialization removed=36
2026-09-07T09:53:43.502683Z  INFO eliminated finite GEMM padding clears using load-time SRAM initialization removed=36
2026-09-07T09:53:43.875301Z  INFO eliminated finite GEMM padding clears using load-time SRAM initialization removed=36
2026-09-07T09:53:43.920407Z  INFO eliminated finite GEMM padding clears using load-time SRAM initialization removed=38
2026-09-07T09:53:47.680999Z  INFO screened tile mappings by exchange resource load candidates=80 baseline_cycles=45376 candidate_cycles=45376 baseline_pressure=30251 candidate_pressure=30251
2026-09-07T09:53:47.736099Z  INFO modelled expanded operator plan finalist=3 modelled_cycles=177268 estimated_cycles=190364 estimated_exchange_cycles=60872
2026-09-07T09:53:48.223138Z  INFO screened tile mappings by exchange resource load candidates=74 baseline_cycles=41920 candidate_cycles=41920 baseline_pressure=29556 candidate_pressure=29556
2026-09-07T09:53:48.278135Z  INFO modelled expanded operator plan finalist=0 modelled_cycles=171638 estimated_cycles=209310 estimated_exchange_cycles=81992
2026-09-07T09:53:49.400686Z  INFO screened tile mappings by exchange resource load candidates=62 baseline_cycles=43744 candidate_cycles=43744 baseline_pressure=39600 candidate_pressure=39600
2026-09-07T09:53:49.487863Z  INFO modelled expanded operator plan finalist=2 modelled_cycles=177800 estimated_cycles=210992 estimated_exchange_cycles=79336
2026-09-07T09:53:49.615123Z  INFO screened tile mappings by exchange resource load candidates=52 baseline_cycles=37152 candidate_cycles=37152 baseline_pressure=32228 candidate_pressure=32228
2026-09-07T09:53:49.723446Z  INFO modelled expanded operator plan finalist=1 modelled_cycles=162650 estimated_cycles=210306 estimated_exchange_cycles=87208
2026-09-07T09:53:51.064973Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:ipu_codegen.package.phase{phase="select_finalist"}: prepared large physical exchange phase phase=1 transfers=32280 maximum_identical_destinations=1
2026-09-07T09:53:51.065027Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:ipu_codegen.package.phase{phase="select_finalist"}: optimized physical exchange schedule phase=1 initial_horizon=9994 selected_horizon=9600 endpoint_lower_bound=7776 lower_bound_gap=1824 selected_kind="critical-neighborhood" neighborhood_improvements=1
2026-09-07T09:53:51.094167Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:ipu_codegen.package.phase{phase="select_finalist"}: finished large physical exchange phase phase=1 tile=705 row_words=140 horizon=9600
2026-09-07T09:53:55.532452Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:ipu_codegen.package.phase{phase="select_finalist"}: compared ordinary and paired exchange schedules phase=2 candidates=4417 ordinary_horizon=29091 paired_horizon=22790 use_paired=true
2026-09-07T09:53:55.622722Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:ipu_codegen.package.phase{phase="select_finalist"}: prepared large physical exchange phase phase=2 transfers=5814 maximum_identical_destinations=1
2026-09-07T09:53:55.622774Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:ipu_codegen.package.phase{phase="select_finalist"}: optimized physical exchange schedule phase=2 initial_horizon=26271 selected_horizon=22790 endpoint_lower_bound=17280 lower_bound_gap=5510 selected_kind="critical-neighborhood" neighborhood_improvements=3
2026-09-07T09:53:55.662554Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:ipu_codegen.package.phase{phase="select_finalist"}: finished large physical exchange phase phase=2 tile=934 row_words=227 horizon=22790
2026-09-07T09:54:34.016943Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:ipu_codegen.package.phase{phase="select_finalist"}: prepared large physical exchange phase phase=3 transfers=155880 maximum_identical_destinations=1
2026-09-07T09:54:34.016985Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:ipu_codegen.package.phase{phase="select_finalist"}: 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
2026-09-07T09:54:34.138920Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:ipu_codegen.package.phase{phase="select_finalist"}: finished large physical exchange phase phase=3 tile=447 row_words=633 horizon=15700
2026-09-07T09:54:34.189034Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:ipu_codegen.package.phase{phase="select_finalist"}: scheduled operator-plan finalist finalist=1 mapped=false analytical_cycles=210306 analytical_exchange_cycles=87208 scheduled_exchange_cycles=54160 refined_cycles=177258
2026-09-07T09:54:34.252971Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:ipu_codegen.package.phase{phase="select_finalist"}: selected physically scheduled operator-plan finalist selected=1 mapped=false
2026-09-07T09:54:34.252989Z  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=52881 success=true
2026-09-07T09:54:34.253009Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}: lowered graph for package construction logical_shards=26518 exchange_phases=4
2026-09-07T09:54:34.253436Z  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
2026-09-07T09:54:34.254347Z  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
2026-09-07T09:54:34.255491Z  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_k144_c32_r364_r366" target=ipu21
2026-09-07T09:54:34.312094Z  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/93562a1f5fd25a261a2410d3e84e88e55f967d5460b30dc5391a03608306eb3f/kernel.o metadata=/srv/home/.cache/ipu-stack/kernels/93562a1f5fd25a261a2410d3e84e88e55f967d5460b30dc5391a03608306eb3f/kernel.json
2026-09-07T09:54:34.312479Z  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_k144_c48_r364_r366" target=ipu21
2026-09-07T09:54:34.358866Z  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/f944eb668c99389f3c11e7d34619a0817133fabfd161a5a1a8be3b364cac632c/kernel.o metadata=/srv/home/.cache/ipu-stack/kernels/f944eb668c99389f3c11e7d34619a0817133fabfd161a5a1a8be3b364cac632c/kernel.json
2026-09-07T09:54:34.359692Z  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
2026-09-07T09:54:34.361181Z  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
2026-09-07T09:54:34.761307Z  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=10 success=true
2026-09-07T09:54:34.761373Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}: allocated package support memory linked_end=319380 profile_bytes=100 exchange_table_bytes=3652 host_code_bytes=2272 generated_code_bytes=1408 code_address=321720 standard_ranges=[(409600, 524288)]
2026-09-07T09:54:35.179282Z  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=417 success=true
2026-09-07T09:54:35.182671Z  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
2026-09-07T09:54:35.339734Z  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
2026-09-07T09:54:35.348185Z  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=32280 maximum_identical_destinations=1
2026-09-07T09:54:35.348198Z  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=9600 selected_horizon=9600 endpoint_lower_bound=7776 lower_bound_gap=1824 selected_kind="reused" neighborhood_improvements=0
2026-09-07T09:54:35.364431Z  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=705 row_words=140 horizon=9600
2026-09-07T09:54:35.787091Z  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
2026-09-07T09:54:35.799245Z  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
2026-09-07T09:54:35.799258Z  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=22790 selected_horizon=22790 endpoint_lower_bound=17280 lower_bound_gap=5510 selected_kind="reused" neighborhood_improvements=0
2026-09-07T09:54:35.820820Z  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=934 row_words=227 horizon=22790
2026-09-07T09:54:45.571794Z  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=3
2026-09-07T09:54:45.624974Z  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
2026-09-07T09:54:45.624994Z  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="reused" neighborhood_improvements=0
2026-09-07T09:54:45.736400Z  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
2026-09-07T09:54:45.768980Z  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=10589 success=true
2026-09-07T09:54:49.285823Z  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
2026-09-07T09:54:49.437712Z  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
2026-09-07T09:54:49.446028Z  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=32280 maximum_identical_destinations=1
2026-09-07T09:54:49.446040Z  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=9600 selected_horizon=9600 endpoint_lower_bound=7776 lower_bound_gap=1824 selected_kind="reused" neighborhood_improvements=0
2026-09-07T09:54:49.462270Z  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=705 row_words=140 horizon=9600
2026-09-07T09:54:51.315523Z  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=28429 paired_horizon=21776 use_paired=true
2026-09-07T09:54:51.358534Z  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
2026-09-07T09:54:51.358553Z  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=21776 selected_horizon=21776 endpoint_lower_bound=17280 lower_bound_gap=4496 selected_kind="full-duplex" neighborhood_improvements=0
2026-09-07T09:54:51.382687Z  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=878 row_words=233 horizon=21776
2026-09-07T09:55:00.720908Z  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=3
2026-09-07T09:55:00.766890Z  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
2026-09-07T09:55:00.766911Z  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="reused" neighborhood_improvements=0
2026-09-07T09:55:00.881748Z  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
2026-09-07T09:55:00.923198Z  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=51760 candidate_cycles=50746 row_bytes=3652 row_capacity=16320 accepted=true
2026-09-07T09:55:00.938081Z  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=15169 success=true
2026-09-07T09:55:00.942186Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}: costed final placed program final_cycles=176244 final_exchange=53146
2026-09-07T09:55:01.142144Z  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=16 success=true
2026-09-07T09:55:01.153631Z  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
2026-09-07T09:55:01.223868Z  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=66 success=true
2026-09-07T09:55:01.274876Z  INFO writing application package tiles=1472
2026-09-07T09:55:01.303388Z  INFO application package written
2026-09-07T09:55:01.303736Z  INFO reading application package
2026-09-07T09:55:01.375883Z  INFO application package read tiles=1472 compiler=0.1.0
2026-09-07T09:55:01.376093Z  INFO waiting for hardware ownership path=artifacts/layout-sweep/device.lock
2026-09-07T09:55:01.376100Z  INFO acquired hardware ownership path=artifacts/layout-sweep/device.lock
2026-09-07T09:55:01.376104Z  INFO opening IPU device path="/dev/ipu0"
2026-09-07T09:55:01.376125Z  INFO IPU device attached path=/dev/ipu0
2026-09-07T09:55:01.376127Z  INFO initializing IPU device
2026-09-07T09:55:01.384110Z  INFO IPU device initialized
2026-09-07T09:55:01.384133Z  INFO replaying device configuration register_writes=2003
2026-09-07T09:55:01.392123Z  INFO loading application tile_count=1472 final_mark=23
2026-09-07T09:55:01.602911Z  INFO application loaded and initial synchronization completed tile_count=1472
2026-09-07T09:55:03.179763Z  INFO starting host exchange session startup_mark=23
2026-09-07T09:55:03.179794Z  INFO attaching host exchange pages pages=2
2026-09-07T09:55:03.180228Z  INFO host exchange session started
2026-09-07T09:55:03.180275Z  INFO invoking streaming host exchange call call="initialize" command=0 phases=10
2026-09-07T09:55:03.185427Z  INFO invoking streaming host exchange call call="run" command=0 phases=59
profile=artifacts/useful-work/mlp/execution.ipuprofile tiles=1472 samples=24023
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=176478 minimumTileCycles=154908 deviceMicroseconds=117.652 effectiveGemmTflops=122.889 peakTflops=282.624 efficiencyPercent=43.48 maximumAbsoluteError=0.001221
package=artifacts/useful-work/mlp/model.ipuexe tiles=1472 hardwareTest=PASS