2026-09-22T15:41:17.091180Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:compile_runtime: close time.busy=289µs time.idle=5.21µs
2026-09-22T15:41:35.897184Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:construct_mid:gemm_candidates: close time.busy=18.8s time.idle=4.73µs
2026-09-22T15:42:08.299749Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:construct_mid:gemm_candidates: close time.busy=32.4s time.idle=4.69µs
2026-09-22T15:42:10.285038Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:construct_mid: close time.busy=53.2s time.idle=2.71µs
2026-09-22T15:42:10.334670Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}: wrote planner memory profile path=artifacts/kernel-cost-mlp-20260922/fp8-b1-n1-default/memory/baseline-1176494-0000.json scope="baseline" peak_step=2 effective_peak_bytes=154624
2026-09-22T15:42:11.014095Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:group_exchanges: close time.busy=83.6µs time.idle=2.84µs
2026-09-22T15:42:11.015106Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:merge_copies: close time.busy=843µs time.idle=1.00µs
2026-09-22T15:42:11.036102Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:eliminate: eliminated padding clears using kernel access contracts removed=0
2026-09-22T15:42:11.037247Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:eliminate: close time.busy=22.1ms time.idle=766ns
2026-09-22T15:42:11.069030Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:place_provisional_storage: close time.busy=31.3ms time.idle=1.63µs
2026-09-22T15:42:11.411366Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:schedule_provisional_exchanges: prepared large physical exchange phase phase=4 transfers=3096 maximum_identical_destinations=1
2026-09-22T15:42:11.411400Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:schedule_provisional_exchanges: optimized physical exchange schedule phase=4 initial_horizon=7943 selected_horizon=7943 endpoint_lower_bound=7680 lower_bound_gap=263 selected_kind="full-duplex" neighborhood_improvements=0
2026-09-22T15:42:11.476047Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:schedule_provisional_exchanges: retrying exchange schedule with incremental instruction-alignment validation transfers=50933
2026-09-22T15:42:11.958889Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:schedule_provisional_exchanges: finished incremental instruction-alignment validation transfers=50933 elapsed_ms=482 success=true
2026-09-22T15:42:12.114596Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:schedule_provisional_exchanges: retrying exchange schedule with incremental instruction-alignment validation transfers=131766
2026-09-22T15:42:12.220322Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:schedule_provisional_exchanges: retrying exchange schedule with incremental instruction-alignment validation transfers=21111
2026-09-22T15:42:12.465398Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:schedule_provisional_exchanges: prepared large physical exchange phase phase=6 transfers=50933 maximum_identical_destinations=1
2026-09-22T15:42:12.465424Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:schedule_provisional_exchanges: optimized physical exchange schedule phase=6 initial_horizon=474 selected_horizon=474 endpoint_lower_bound=288 lower_bound_gap=186 selected_kind="full-duplex" neighborhood_improvements=0
2026-09-22T15:42:13.482521Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:schedule_provisional_exchanges: retrying exchange schedule with incremental instruction-alignment validation transfers=196058
2026-09-22T15:42:13.575181Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:schedule_provisional_exchanges: retrying exchange schedule with incremental instruction-alignment validation transfers=15488
2026-09-22T15:42:13.602546Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:schedule_provisional_exchanges: finished incremental instruction-alignment validation transfers=131766 elapsed_ms=1487 success=true
2026-09-22T15:42:13.698863Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:schedule_provisional_exchanges: finished incremental instruction-alignment validation transfers=21111 elapsed_ms=1478 success=true
2026-09-22T15:42:15.391459Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:schedule_provisional_exchanges: prepared large physical exchange phase phase=1 transfers=131766 maximum_identical_destinations=1
2026-09-22T15:42:15.391488Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:schedule_provisional_exchanges: optimized physical exchange schedule phase=1 initial_horizon=8448 selected_horizon=8448 endpoint_lower_bound=6912 lower_bound_gap=1536 selected_kind="full-duplex" neighborhood_improvements=0
2026-09-22T15:42:15.549767Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:schedule_provisional_exchanges: finished incremental instruction-alignment validation transfers=15488 elapsed_ms=1974 success=true
2026-09-22T15:42:16.366297Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:schedule_provisional_exchanges: compared ordinary and paired exchange schedules phase=3 candidates=20310 ordinary_horizon=7868 paired_horizon=8729 use_paired=false
2026-09-22T15:42:16.443470Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:schedule_provisional_exchanges: prepared large physical exchange phase phase=3 transfers=21111 maximum_identical_destinations=1
2026-09-22T15:42:16.443494Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:schedule_provisional_exchanges: optimized physical exchange schedule phase=3 initial_horizon=7868 selected_horizon=7868 endpoint_lower_bound=6480 lower_bound_gap=1388 selected_kind="full-duplex" neighborhood_improvements=0
2026-09-22T15:42:16.941418Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:schedule_provisional_exchanges: finished incremental instruction-alignment validation transfers=196058 elapsed_ms=3458 success=true
2026-09-22T15:42:17.569394Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:schedule_provisional_exchanges: retrying exchange schedule with incremental instruction-alignment validation transfers=15488
2026-09-22T15:42:19.141919Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:schedule_provisional_exchanges: retrying exchange schedule with incremental instruction-alignment validation transfers=454896
2026-09-22T15:42:19.516607Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:schedule_provisional_exchanges: finished incremental instruction-alignment validation transfers=15488 elapsed_ms=1947 success=true
2026-09-22T15:42:20.175520Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:schedule_provisional_exchanges: prepared large physical exchange phase phase=2 transfers=196058 maximum_identical_destinations=1
2026-09-22T15:42:20.175551Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:schedule_provisional_exchanges: optimized physical exchange schedule phase=2 initial_horizon=1705 selected_horizon=1705 endpoint_lower_bound=1152 lower_bound_gap=553 selected_kind="full-duplex" neighborhood_improvements=0
2026-09-22T15:42:20.349363Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:schedule_provisional_exchanges: compared ordinary and paired exchange schedules phase=0 candidates=23 ordinary_horizon=15608 paired_horizon=15657 use_paired=false
2026-09-22T15:42:20.450509Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:schedule_provisional_exchanges: prepared large physical exchange phase phase=0 transfers=15488 maximum_identical_destinations=1
2026-09-22T15:42:20.450537Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:schedule_provisional_exchanges: optimized physical exchange schedule phase=0 initial_horizon=15608 selected_horizon=15608 endpoint_lower_bound=14832 lower_bound_gap=776 selected_kind="full-duplex" neighborhood_improvements=0
2026-09-22T15:42:28.423107Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:schedule_provisional_exchanges: finished incremental instruction-alignment validation transfers=454896 elapsed_ms=9281 success=true
2026-09-22T15:42:37.154006Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:schedule_provisional_exchanges: prepared large physical exchange phase phase=5 transfers=454896 maximum_identical_destinations=1
2026-09-22T15:42:37.154040Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:schedule_provisional_exchanges: optimized physical exchange schedule phase=5 initial_horizon=9273 selected_horizon=9273 endpoint_lower_bound=7488 lower_bound_gap=1785 selected_kind="full-duplex" neighborhood_improvements=0
2026-09-22T15:42:37.742728Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:schedule_provisional_exchanges: close time.busy=26.7s time.idle=6.71µs
2026-09-22T15:42:37.794545Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:plan_kernels: close time.busy=51.7ms time.idle=5.80µs
2026-09-22T15:42:37.795533Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:compile_kernels: close time.busy=972µs time.idle=900ns
2026-09-22T15:42:37.795710Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:link_runtime: close time.busy=118µs time.idle=687ns
2026-09-22T15:42:37.914791Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}: exchange table and repeat patch storage exchange_table_bytes=12900 maximum_uncompressed_repeat_patch_bytes=0 maximum_elided_arithmetic_patch_bytes=0
2026-09-22T15:42:38.120874Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:size_tile_code: close time.busy=18.6ms time.idle=2.13µs
2026-09-22T15:42:38.120894Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}: placing generated tile programs linked_end=318984 host_code_base=376832 host_code_bytes=620 generated_code_bytes=1168 exchange_table_bytes=12900 executable_ranges=[(377520, 524288)]
2026-09-22T15:42:38.120916Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}: allocated package support memory linked_end=318984 profile_bytes=96 exchange_table_bytes=12900 host_code_bytes=620 host_data_bytes=404 generated_code_bytes=1168 code_address=377520 available_ranges=[(327680, 360448), (362612, 376832), (409600, 949168)]
2026-09-22T15:42:38.175850Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:place_storage: close time.busy=45.6ms time.idle=2.04µs
2026-09-22T15:42:38.318559Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:lower_exchanges: reused exchange optimization after validating relocated rows phase=4
2026-09-22T15:42:38.323416Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:lower_exchanges: prepared large physical exchange phase phase=4 transfers=3096 maximum_identical_destinations=1
2026-09-22T15:42:38.323433Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:lower_exchanges: optimized physical exchange schedule phase=4 initial_horizon=7943 selected_horizon=7943 endpoint_lower_bound=7680 lower_bound_gap=263 selected_kind="reused" neighborhood_improvements=0
2026-09-22T15:42:39.076769Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:lower_exchanges: retrying exchange schedule with incremental instruction-alignment validation transfers=50933
2026-09-22T15:42:39.620028Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:lower_exchanges: reused exchange optimization after validating relocated rows phase=3
2026-09-22T15:42:39.640096Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:lower_exchanges: prepared large physical exchange phase phase=3 transfers=21111 maximum_identical_destinations=1
2026-09-22T15:42:39.640111Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:lower_exchanges: optimized physical exchange schedule phase=3 initial_horizon=7868 selected_horizon=7868 endpoint_lower_bound=6480 lower_bound_gap=1388 selected_kind="reused" neighborhood_improvements=0
2026-09-22T15:42:39.687191Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:lower_exchanges: finished incremental instruction-alignment validation transfers=50933 elapsed_ms=610 success=true
2026-09-22T15:42:40.444415Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:lower_exchanges: prepared large physical exchange phase phase=6 transfers=50933 maximum_identical_destinations=1
2026-09-22T15:42:40.444446Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:lower_exchanges: optimized physical exchange schedule phase=6 initial_horizon=474 selected_horizon=474 endpoint_lower_bound=288 lower_bound_gap=186 selected_kind="full-duplex" neighborhood_improvements=0
2026-09-22T15:42:40.558366Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:lower_exchanges: reused exchange optimization after validating relocated rows phase=0
2026-09-22T15:42:40.608870Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:lower_exchanges: prepared large physical exchange phase phase=0 transfers=15488 maximum_identical_destinations=1
2026-09-22T15:42:40.608893Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:lower_exchanges: optimized physical exchange schedule phase=0 initial_horizon=15608 selected_horizon=15608 endpoint_lower_bound=14832 lower_bound_gap=776 selected_kind="reused" neighborhood_improvements=0
2026-09-22T15:42:40.770635Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:lower_exchanges: reused exchange optimization after validating relocated rows phase=1
2026-09-22T15:42:40.812845Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:lower_exchanges: prepared large physical exchange phase phase=1 transfers=131766 maximum_identical_destinations=1
2026-09-22T15:42:40.812867Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:lower_exchanges: optimized physical exchange schedule phase=1 initial_horizon=8448 selected_horizon=8448 endpoint_lower_bound=6912 lower_bound_gap=1536 selected_kind="reused" neighborhood_improvements=0
2026-09-22T15:42:41.410333Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:lower_exchanges: reused exchange optimization after validating relocated rows phase=2
2026-09-22T15:42:41.452884Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:lower_exchanges: prepared large physical exchange phase phase=2 transfers=196058 maximum_identical_destinations=1
2026-09-22T15:42:41.452900Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:lower_exchanges: optimized physical exchange schedule phase=2 initial_horizon=1705 selected_horizon=1705 endpoint_lower_bound=1152 lower_bound_gap=553 selected_kind="reused" neighborhood_improvements=0
2026-09-22T15:42:49.740002Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:lower_exchanges: reused exchange optimization after validating relocated rows phase=5
2026-09-22T15:42:49.937381Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:lower_exchanges: prepared large physical exchange phase phase=5 transfers=454896 maximum_identical_destinations=1
2026-09-22T15:42:49.937423Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:lower_exchanges: optimized physical exchange schedule phase=5 initial_horizon=9273 selected_horizon=9273 endpoint_lower_bound=7488 lower_bound_gap=1785 selected_kind="reused" neighborhood_improvements=0
2026-09-22T15:42:50.823075Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:lower_exchanges: close time.busy=12.6s time.idle=4.25µs
2026-09-22T15:42:50.842622Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}: costed final placed program final_cycles=205503 final_exchange=55519
2026-09-22T15:42:51.151624Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:prepare_tile_code: close time.busy=37.7ms time.idle=4.53µs
2026-09-22T15:42:51.172135Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:emit_tile_code: close time.busy=20.5ms time.idle=2.13µs
2026-09-22T15:42:51.343388Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:build_tile_images: close time.busy=165ms time.idle=3.05µs
2026-09-22T15:42:51.541857Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}: wrote exact tile placement profile path=artifacts/kernel-cost-mlp-20260922/fp8-b1-n1-default/memory/placement-1176494.html
2026-09-22T15:42:51.644757Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}: close time.busy=94.6s time.idle=6.11µs
2026-09-22T15:42:51.648627Z  INFO writing application package tiles=1472
2026-09-22T15:42:51.799956Z  INFO application package written
2026-09-22T15:42:51.804153Z  INFO reading application package
2026-09-22T15:42:51.984472Z  INFO application package read tiles=1472 compiler=0.1.0
2026-09-22T15:42:51.990011Z  INFO waiting for hardware ownership path=/tmp/ipu-stack-device.lock
2026-09-22T15:42:51.990024Z  INFO acquired hardware ownership path=/tmp/ipu-stack-device.lock
2026-09-22T15:42:51.990035Z  INFO opening IPU device path="/dev/ipu0"
2026-09-22T15:42:51.990065Z  INFO IPU device attached path=/dev/ipu0
2026-09-22T15:42:51.990068Z  INFO initializing IPU device
2026-09-22T15:42:51.997819Z  INFO IPU device initialized
2026-09-22T15:42:51.997824Z  INFO replaying device configuration register_writes=2003
2026-09-22T15:42:52.003370Z  INFO loading application tile_count=1472 final_mark=23
2026-09-22T15:42:52.189008Z  INFO application loaded and initial synchronization completed tile_count=1472
2026-09-22T15:42:52.931378Z  INFO prepared reference inputs elapsed_ms=738
2026-09-22T15:42:53.284806Z  INFO configured host reference BLAS threads=4
referencePrecision=device
2026-09-22T15:42:53.837906Z  INFO evaluated host reference elapsed_ms=906
2026-09-22T15:42:53.840335Z  INFO starting host exchange session startup_mark=23
2026-09-22T15:42:53.840349Z  INFO attaching host exchange pages pages=2
2026-09-22T15:42:53.841120Z  INFO host exchange session started
2026-09-22T15:42:53.841164Z  INFO invoking host exchange call call="initialize" command=0 phases=4
2026-09-22T15:42:53.845599Z  INFO invoking host exchange call call="run" command=0 phases=13
referenceInference=1 numericalTest=PASS
profile=artifacts/kernel-cost-mlp-20260922/fp8-b1-n1-default/profile.capnp tiles=1472 samples=26676
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=1469 executionTiles=1472 inputBytes=839808 weightBytes=9934848 cycles=204006 minimumTileCycles=199986 deviceMicroseconds=136.004 effectiveGemmTflops=106.307 peakTflops=282.624 efficiencyPercent=37.61 maximumAbsoluteError=0.005554
package=artifacts/kernel-cost-mlp-20260922/fp8-b1-n1-default/model.ipuexe tiles=1472 hardwareTest=PASS