2026-09-07T14:01:06.872374Z INFO retained operator-plan finalist finalist=0 values=18 operations=1 estimated_cycles=235972 estimated_exchange_cycles=51920 exchange_row_bytes=4924 peak_standard_bytes=40764 peak_interleaved_bytes=36864 peak_total_bytes=76604 maximum_standard_allocation_bytes=20480 active_tile_counts={64} padding_group_counts={} conversion_sources=[] 2026-09-07T14:01:06.877322Z INFO ipu_codegen.package.build{tile_count=64 operations=1}:ipu_codegen.package.phase{phase="lower_mid"}: package build phase finished phase="lower_mid" elapsed_ms=296 success=true 2026-09-07T14:01:06.935126Z INFO ipu_codegen.package.build{tile_count=64 operations=1}:ipu_codegen.package.phase{phase="select_finalist"}: screened tile mappings by exchange resource load candidates=11 baseline_cycles=20736 candidate_cycles=20736 baseline_pressure=17990 candidate_pressure=17990 2026-09-07T14:01:06.935410Z INFO ipu_codegen.package.build{tile_count=64 operations=1}:ipu_codegen.package.phase{phase="select_finalist"}: modelled expanded operator plan finalist=0 modelled_cycles=155816 estimated_cycles=312152 estimated_exchange_cycles=183072 2026-09-07T14:01:07.217908Z INFO ipu_codegen.package.build{tile_count=64 operations=1}:ipu_codegen.package.phase{phase="select_finalist"}: prepared large physical exchange phase phase=4 transfers=3584 maximum_identical_destinations=1 2026-09-07T14:01:07.217967Z INFO ipu_codegen.package.build{tile_count=64 operations=1}:ipu_codegen.package.phase{phase="select_finalist"}: optimized physical exchange schedule phase=4 initial_horizon=1673 selected_horizon=1673 endpoint_lower_bound=448 lower_bound_gap=1225 selected_kind="full-duplex" neighborhood_improvements=0 2026-09-07T14:01:07.220427Z INFO ipu_codegen.package.build{tile_count=64 operations=1}:ipu_codegen.package.phase{phase="select_finalist"}: finished large physical exchange phase phase=4 tile=5 row_words=266 horizon=1673 2026-09-07T14:01:07.353173Z INFO ipu_codegen.package.build{tile_count=64 operations=1}:ipu_codegen.package.phase{phase="select_finalist"}: prepared large physical exchange phase phase=0 transfers=2064 maximum_identical_destinations=1 2026-09-07T14:01:07.353246Z INFO ipu_codegen.package.build{tile_count=64 operations=1}:ipu_codegen.package.phase{phase="select_finalist"}: optimized physical exchange schedule phase=0 initial_horizon=3823 selected_horizon=3823 endpoint_lower_bound=1472 lower_bound_gap=2351 selected_kind="full-duplex" neighborhood_improvements=0 2026-09-07T14:01:07.355570Z INFO ipu_codegen.package.build{tile_count=64 operations=1}:ipu_codegen.package.phase{phase="select_finalist"}: finished large physical exchange phase phase=0 tile=4 row_words=366 horizon=3823 2026-09-07T14:01:09.298640Z INFO ipu_codegen.package.build{tile_count=64 operations=1}:ipu_codegen.package.phase{phase="select_finalist"}: prepared large physical exchange phase phase=2 transfers=8200 maximum_identical_destinations=1 2026-09-07T14:01:09.298720Z INFO ipu_codegen.package.build{tile_count=64 operations=1}:ipu_codegen.package.phase{phase="select_finalist"}: optimized physical exchange schedule phase=2 initial_horizon=2830 selected_horizon=2830 endpoint_lower_bound=2048 lower_bound_gap=782 selected_kind="full-duplex" neighborhood_improvements=0 2026-09-07T14:01:09.307908Z INFO ipu_codegen.package.build{tile_count=64 operations=1}:ipu_codegen.package.phase{phase="select_finalist"}: finished large physical exchange phase phase=2 tile=50 row_words=925 horizon=2830 2026-09-07T14:01:09.312818Z INFO ipu_codegen.package.build{tile_count=64 operations=1}:ipu_codegen.package.phase{phase="select_finalist"}: scheduled operator-plan finalist finalist=0 mapped=false analytical_cycles=312152 analytical_exchange_cycles=183072 scheduled_exchange_cycles=35364 refined_cycles=164444 2026-09-07T14:01:09.320473Z INFO ipu_codegen.package.build{tile_count=64 operations=1}:ipu_codegen.package.phase{phase="select_finalist"}: selected physically scheduled operator-plan finalist selected=0 mapped=false 2026-09-07T14:01:09.320497Z INFO ipu_codegen.package.build{tile_count=64 operations=1}:ipu_codegen.package.phase{phase="select_finalist"}: package build phase finished phase="select_finalist" elapsed_ms=2443 success=true 2026-09-07T14:01:09.320523Z INFO ipu_codegen.package.build{tile_count=64 operations=1}: lowered graph for package construction logical_shards=1864 exchange_phases=5 2026-09-07T14:01:09.320829Z INFO ipu_codegen.package.build{tile_count=64 operations=1}:ipu_codegen.package.phase{phase="compile_runtime"}: package build phase finished phase="compile_runtime" elapsed_ms=0 success=true 2026-09-07T14:01:09.321010Z INFO ipu_codegen.package.build{tile_count=64 operations=1}:ipu_codegen.package.phase{phase="plan_kernels"}: package build phase finished phase="plan_kernels" elapsed_ms=0 success=true 2026-09-07T14:01:09.321266Z INFO ipu_codegen.package.build{tile_count=64 operations=1}: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_f8_k32_c64_r128_r128" target=ipu21 2026-09-07T14:01:09.394252Z INFO ipu_codegen.package.build{tile_count=64 operations=1}:ipu_codegen.package.phase{phase="compile_kernels"}: kernel artifact created object=/srv/home/.cache/ipu-stack/kernels/4ac4f9850bf0debe07838c5d5c84848be93a83161c2c3e344f00c9c27b3230a4/kernel.o metadata=/srv/home/.cache/ipu-stack/kernels/4ac4f9850bf0debe07838c5d5c84848be93a83161c2c3e344f00c9c27b3230a4/kernel.json 2026-09-07T14:01:09.394537Z INFO ipu_codegen.package.build{tile_count=64 operations=1}: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_f8_k64_c64_r64_r64" target=ipu21 2026-09-07T14:01:09.450101Z INFO ipu_codegen.package.build{tile_count=64 operations=1}:ipu_codegen.package.phase{phase="compile_kernels"}: kernel artifact created object=/srv/home/.cache/ipu-stack/kernels/910667e0ddc7e0b725603d6a8d6294851070a481a3d02279df20b3c3abf08493/kernel.o metadata=/srv/home/.cache/ipu-stack/kernels/910667e0ddc7e0b725603d6a8d6294851070a481a3d02279df20b3c3abf08493/kernel.json 2026-09-07T14:01:09.450356Z INFO ipu_codegen.package.build{tile_count=64 operations=1}:ipu_codegen.package.phase{phase="compile_kernels"}: compiling kernel source=/srv/home/gc-sdk/ipu-stack/crates/ipu-tests/../../device/gelu_f16.S name="gelu_f16" target=ipu21 2026-09-07T14:01:09.500801Z INFO ipu_codegen.package.build{tile_count=64 operations=1}:ipu_codegen.package.phase{phase="compile_kernels"}: kernel artifact created object=/srv/home/.cache/ipu-stack/kernels/ece65f66eca7d2bc63bd2afe2c9fa1ec2985ad46527a99a77da1d3000ce1ab8d/kernel.o metadata=/srv/home/.cache/ipu-stack/kernels/ece65f66eca7d2bc63bd2afe2c9fa1ec2985ad46527a99a77da1d3000ce1ab8d/kernel.json 2026-09-07T14:01:09.501570Z INFO ipu_codegen.package.build{tile_count=64 operations=1}:ipu_codegen.package.phase{phase="compile_kernels"}: package build phase finished phase="compile_kernels" elapsed_ms=180 success=true 2026-09-07T14:01:09.502169Z INFO ipu_codegen.package.build{tile_count=64 operations=1}:ipu_codegen.package.phase{phase="link_runtime"}: package build phase finished phase="link_runtime" elapsed_ms=0 success=true 2026-09-07T14:01:09.533735Z INFO ipu_codegen.package.build{tile_count=64 operations=1}:ipu_codegen.package.phase{phase="size_tile_code"}: package build phase finished phase="size_tile_code" elapsed_ms=2 success=true 2026-09-07T14:01:09.533829Z INFO ipu_codegen.package.build{tile_count=64 operations=1}: allocated package support memory linked_end=316404 profile_bytes=24 exchange_table_bytes=6280 host_code_bytes=488 generated_code_bytes=2864 code_address=316960 standard_ranges=[(409600, 524288)] 2026-09-07T14:01:09.538095Z INFO ipu_codegen.package.build{tile_count=64 operations=1}:ipu_codegen.package.phase{phase="place_storage"}: package build phase finished phase="place_storage" elapsed_ms=4 success=true 2026-09-07T14:01:09.541188Z INFO ipu_codegen.package.build{tile_count=64 operations=1}:ipu_codegen.package.phase{phase="lower_exchanges"}: reused exchange optimization after validating relocated rows phase=3 2026-09-07T14:01:09.541497Z INFO ipu_codegen.package.build{tile_count=64 operations=1}:ipu_codegen.package.phase{phase="lower_exchanges"}: reused exchange optimization after validating relocated rows phase=1 2026-09-07T14:01:09.674887Z INFO ipu_codegen.package.build{tile_count=64 operations=1}:ipu_codegen.package.phase{phase="lower_exchanges"}: reused exchange optimization after validating relocated rows phase=4 2026-09-07T14:01:09.676636Z INFO ipu_codegen.package.build{tile_count=64 operations=1}:ipu_codegen.package.phase{phase="lower_exchanges"}: prepared large physical exchange phase phase=4 transfers=3584 maximum_identical_destinations=1 2026-09-07T14:01:09.676662Z INFO ipu_codegen.package.build{tile_count=64 operations=1}:ipu_codegen.package.phase{phase="lower_exchanges"}: optimized physical exchange schedule phase=4 initial_horizon=1673 selected_horizon=1673 endpoint_lower_bound=448 lower_bound_gap=1225 selected_kind="reused" neighborhood_improvements=0 2026-09-07T14:01:09.679075Z INFO ipu_codegen.package.build{tile_count=64 operations=1}:ipu_codegen.package.phase{phase="lower_exchanges"}: finished large physical exchange phase phase=4 tile=5 row_words=266 horizon=1673 2026-09-07T14:01:10.200859Z INFO ipu_codegen.package.build{tile_count=64 operations=1}:ipu_codegen.package.phase{phase="lower_exchanges"}: prepared large physical exchange phase phase=0 transfers=2064 maximum_identical_destinations=1 2026-09-07T14:01:10.200915Z INFO ipu_codegen.package.build{tile_count=64 operations=1}:ipu_codegen.package.phase{phase="lower_exchanges"}: optimized physical exchange schedule phase=0 initial_horizon=4159 selected_horizon=4159 endpoint_lower_bound=1472 lower_bound_gap=2687 selected_kind="full-duplex" neighborhood_improvements=0 2026-09-07T14:01:10.203244Z INFO ipu_codegen.package.build{tile_count=64 operations=1}:ipu_codegen.package.phase{phase="lower_exchanges"}: finished large physical exchange phase phase=0 tile=9 row_words=365 horizon=4159 2026-09-07T14:01:10.782895Z INFO ipu_codegen.package.build{tile_count=64 operations=1}:ipu_codegen.package.phase{phase="lower_exchanges"}: reused exchange optimization after validating relocated rows phase=2 2026-09-07T14:01:10.793661Z INFO ipu_codegen.package.build{tile_count=64 operations=1}:ipu_codegen.package.phase{phase="lower_exchanges"}: prepared large physical exchange phase phase=2 transfers=8200 maximum_identical_destinations=1 2026-09-07T14:01:10.793709Z INFO ipu_codegen.package.build{tile_count=64 operations=1}:ipu_codegen.package.phase{phase="lower_exchanges"}: optimized physical exchange schedule phase=2 initial_horizon=2830 selected_horizon=2830 endpoint_lower_bound=2048 lower_bound_gap=782 selected_kind="reused" neighborhood_improvements=0 2026-09-07T14:01:10.807645Z INFO ipu_codegen.package.build{tile_count=64 operations=1}:ipu_codegen.package.phase{phase="lower_exchanges"}: finished large physical exchange phase phase=2 tile=50 row_words=925 horizon=2830 2026-09-07T14:01:10.810757Z INFO ipu_codegen.package.build{tile_count=64 operations=1}:ipu_codegen.package.phase{phase="lower_exchanges"}: package build phase finished phase="lower_exchanges" elapsed_ms=1272 success=true 2026-09-07T14:01:10.844556Z INFO ipu_codegen.package.build{tile_count=64 operations=1}:ipu_codegen.package.phase{phase="optimize_exchange_placement"}: package build phase finished phase="optimize_exchange_placement" elapsed_ms=33 success=true 2026-09-07T14:01:10.844717Z INFO ipu_codegen.package.build{tile_count=64 operations=1}: costed final placed program final_cycles=165116 final_exchange=36036 2026-09-07T14:01:10.862446Z INFO ipu_codegen.package.build{tile_count=64 operations=1}:ipu_codegen.package.phase{phase="prepare_tile_code"}: package build phase finished phase="prepare_tile_code" elapsed_ms=3 success=true 2026-09-07T14:01:10.868229Z INFO ipu_codegen.package.build{tile_count=64 operations=1}:ipu_codegen.package.phase{phase="emit_tile_code"}: package build phase finished phase="emit_tile_code" elapsed_ms=5 success=true 2026-09-07T14:01:10.955993Z INFO ipu_codegen.package.build{tile_count=64 operations=1}:ipu_codegen.package.phase{phase="build_tile_images"}: package build phase finished phase="build_tile_images" elapsed_ms=85 success=true 2026-09-07T14:01:10.962918Z INFO writing application package tiles=1472 2026-09-07T14:01:10.972300Z INFO application package written 2026-09-07T14:01:10.972711Z INFO reading application package 2026-09-07T14:01:10.986425Z INFO application package read tiles=1472 compiler=0.1.0 2026-09-07T14:01:10.986698Z INFO waiting for hardware ownership path=artifacts/layout-sweep/device.lock 2026-09-07T14:01:10.986712Z INFO acquired hardware ownership path=artifacts/layout-sweep/device.lock 2026-09-07T14:01:10.986721Z INFO opening IPU device path="/dev/ipu0" 2026-09-07T14:01:10.986760Z INFO IPU device attached path=/dev/ipu0 2026-09-07T14:01:10.986766Z INFO initializing IPU device 2026-09-07T14:01:10.994780Z INFO IPU device initialized 2026-09-07T14:01:10.994825Z INFO replaying device configuration register_writes=2003 2026-09-07T14:01:10.999219Z INFO loading application tile_count=1472 final_mark=23 2026-09-07T14:01:11.190910Z INFO application loaded and initial synchronization completed tile_count=1472 2026-09-07T14:01:11.343104Z INFO starting host exchange session startup_mark=23 2026-09-07T14:01:11.343151Z INFO attaching host exchange pages pages=2 2026-09-07T14:01:11.344013Z INFO host exchange session started 2026-09-07T14:01:11.344031Z INFO invoking streaming host exchange call call="initialize" command=0 phases=4 2026-09-07T14:01:11.344623Z INFO invoking streaming host exchange call call="run" command=0 phases=9 referenceMaximumAbsoluteError=0.023132 numericalTest=PASS package=artifacts/fp8/mlp-small/model.ipuexe tiles=1472 hardwareTest=PASS