2026-09-24T06:52:37.787945Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:compile_runtime: close time.busy=285µs time.idle=1.70µs
2026-09-24T06:54:44.437622Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:construct_mid:gemm_candidates: close time.busy=127s time.idle=6.97µs
2026-09-24T06:55:23.838418Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:construct_mid:gemm_candidates: close time.busy=39.4s time.idle=4.68µs
2026-09-24T06:56:22.671600Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:construct_mid:group_exchanges: close time.busy=95.7µs time.idle=3.00µs
2026-09-24T06:56:22.672886Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:construct_mid:merge_copies: close time.busy=1.05ms time.idle=1.05µs
2026-09-24T06:56:22.701704Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:construct_mid:eliminate: eliminated padding clears using kernel access contracts removed=0
2026-09-24T06:56:22.703109Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:construct_mid:eliminate: close time.busy=30.2ms time.idle=1.21µs
2026-09-24T06:56:22.837688Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:construct_mid: costed planning finalist rank=0 mid_cycles=143728 low_cycles=135418 elapsed_ms=907
2026-09-24T06:56:23.649870Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:construct_mid:group_exchanges: close time.busy=139µs time.idle=3.71µs
2026-09-24T06:56:23.651446Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:construct_mid:merge_copies: close time.busy=1.33ms time.idle=1.57µs
2026-09-24T06:56:23.679644Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:construct_mid:eliminate: eliminated padding clears using kernel access contracts removed=0
2026-09-24T06:56:23.681081Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:construct_mid:eliminate: close time.busy=29.6ms time.idle=1.03µs
2026-09-24T06:56:23.799906Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:construct_mid: costed planning finalist rank=1 mid_cycles=144894 low_cycles=139884 elapsed_ms=866
2026-09-24T06:56:24.631320Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:construct_mid:group_exchanges: close time.busy=94.6µs time.idle=4.26µs
2026-09-24T06:56:24.632624Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:construct_mid:merge_copies: close time.busy=1.07ms time.idle=1.37µs
2026-09-24T06:56:24.662126Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:construct_mid:eliminate: eliminated padding clears using kernel access contracts removed=0
2026-09-24T06:56:24.663549Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:construct_mid:eliminate: close time.busy=30.9ms time.idle=1.08µs
2026-09-24T06:56:24.794351Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:construct_mid: costed planning finalist rank=2 mid_cycles=145636 low_cycles=137326 elapsed_ms=899
2026-09-24T06:56:25.719432Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:construct_mid:group_exchanges: close time.busy=93.0µs time.idle=3.34µs
2026-09-24T06:56:25.720738Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:construct_mid:merge_copies: close time.busy=1.05ms time.idle=1.20µs
2026-09-24T06:56:25.754786Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:construct_mid:eliminate: eliminated padding clears using kernel access contracts removed=0
2026-09-24T06:56:25.756380Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:construct_mid:eliminate: close time.busy=35.6ms time.idle=1.32µs
2026-09-24T06:56:25.899498Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:construct_mid: costed planning finalist rank=3 mid_cycles=145704 low_cycles=137844 elapsed_ms=1001
2026-09-24T06:56:26.747636Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:construct_mid:group_exchanges: close time.busy=117µs time.idle=3.55µs
2026-09-24T06:56:26.749046Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:construct_mid:merge_copies: close time.busy=1.18ms time.idle=984ns
2026-09-24T06:56:26.777983Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:construct_mid:eliminate: eliminated padding clears using kernel access contracts removed=0
2026-09-24T06:56:26.779400Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:construct_mid:eliminate: close time.busy=30.3ms time.idle=1.04µs
2026-09-24T06:56:26.910262Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:construct_mid: costed planning finalist rank=4 mid_cycles=145828 low_cycles=140654 elapsed_ms=900
2026-09-24T06:56:27.969764Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:construct_mid:group_exchanges: close time.busy=94.9µs time.idle=3.79µs
2026-09-24T06:56:27.971070Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:construct_mid:merge_copies: close time.busy=1.07ms time.idle=1.23µs
2026-09-24T06:56:28.004274Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:construct_mid:eliminate: eliminated padding clears using kernel access contracts removed=0
2026-09-24T06:56:28.005639Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:construct_mid:eliminate: close time.busy=34.6ms time.idle=1.11µs
2026-09-24T06:56:28.154016Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:construct_mid: costed planning finalist rank=5 mid_cycles=146570 low_cycles=138796 elapsed_ms=1146
2026-09-24T06:56:28.994540Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:construct_mid:group_exchanges: close time.busy=98.0µs time.idle=3.76µs
2026-09-24T06:56:28.995858Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:construct_mid:merge_copies: close time.busy=1.07ms time.idle=931ns
2026-09-24T06:56:29.025560Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:construct_mid:eliminate: eliminated padding clears using kernel access contracts removed=0
2026-09-24T06:56:29.026972Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:construct_mid:eliminate: close time.busy=31.1ms time.idle=917ns
2026-09-24T06:56:29.159866Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:construct_mid: costed planning finalist rank=6 mid_cycles=146590 low_cycles=138280 elapsed_ms=901
2026-09-24T06:56:30.216867Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:construct_mid:group_exchanges: close time.busy=96.8µs time.idle=4.04µs
2026-09-24T06:56:30.218165Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:construct_mid:merge_copies: close time.busy=1.05ms time.idle=1.24µs
2026-09-24T06:56:30.250374Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:construct_mid:eliminate: eliminated padding clears using kernel access contracts removed=0
2026-09-24T06:56:30.251757Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:construct_mid:eliminate: close time.busy=33.6ms time.idle=1.02µs
2026-09-24T06:56:30.396965Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:construct_mid: costed planning finalist rank=7 mid_cycles=146620 low_cycles=141594 elapsed_ms=1130
2026-09-24T06:56:30.449102Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:construct_mid: close time.busy=233s time.idle=1.72µs
2026-09-24T06:56:31.199281Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:group_exchanges: close time.busy=127µs time.idle=3.56µs
2026-09-24T06:56:31.200573Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:merge_copies: close time.busy=1.05ms time.idle=824ns
2026-09-24T06:56:31.229776Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:eliminate: eliminated padding clears using kernel access contracts removed=0
2026-09-24T06:56:31.231182Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:eliminate: close time.busy=30.6ms time.idle=1.20µs
2026-09-24T06:56:31.273265Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:place_provisional_storage: close time.busy=41.6ms time.idle=1.70µs
2026-09-24T06:56:31.493774Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:schedule_provisional_exchanges: retrying exchange schedule with incremental instruction-alignment validation transfers=27712
2026-09-24T06:56:31.596870Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:schedule_provisional_exchanges: prepared large physical exchange phase phase=2 transfers=16140 maximum_identical_destinations=1
2026-09-24T06:56:31.596904Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:schedule_provisional_exchanges: optimized physical exchange schedule phase=2 initial_horizon=4127 selected_horizon=4127 endpoint_lower_bound=3384 lower_bound_gap=743 selected_kind="full-duplex" neighborhood_improvements=0
2026-09-24T06:56:31.612776Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:schedule_provisional_exchanges: retrying exchange schedule with incremental instruction-alignment validation transfers=27380
2026-09-24T06:56:31.793066Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:schedule_provisional_exchanges: finished incremental instruction-alignment validation transfers=27712 elapsed_ms=299 success=true
2026-09-24T06:56:32.083996Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:schedule_provisional_exchanges: finished incremental instruction-alignment validation transfers=27380 elapsed_ms=471 success=true
2026-09-24T06:56:32.105723Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:schedule_provisional_exchanges: retrying exchange schedule with incremental instruction-alignment validation transfers=29337
2026-09-24T06:56:32.197890Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:schedule_provisional_exchanges: prepared large physical exchange phase phase=3 transfers=27712 maximum_identical_destinations=1
2026-09-24T06:56:32.197918Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:schedule_provisional_exchanges: optimized physical exchange schedule phase=3 initial_horizon=2071 selected_horizon=2071 endpoint_lower_bound=1104 lower_bound_gap=967 selected_kind="full-duplex" neighborhood_improvements=0
2026-09-24T06:56:32.312122Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:schedule_provisional_exchanges: retrying exchange schedule with incremental instruction-alignment validation transfers=11124
2026-09-24T06:56:32.553015Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:schedule_provisional_exchanges: prepared large physical exchange phase phase=5 transfers=33732 maximum_identical_destinations=1
2026-09-24T06:56:32.553040Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:schedule_provisional_exchanges: optimized physical exchange schedule phase=5 initial_horizon=8462 selected_horizon=7944 endpoint_lower_bound=6912 lower_bound_gap=1032 selected_kind="matching-waves" neighborhood_improvements=0
2026-09-24T06:56:32.625057Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:schedule_provisional_exchanges: prepared large physical exchange phase phase=0 transfers=27380 maximum_identical_destinations=1
2026-09-24T06:56:32.625081Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:schedule_provisional_exchanges: optimized physical exchange schedule phase=0 initial_horizon=341 selected_horizon=341 endpoint_lower_bound=160 lower_bound_gap=181 selected_kind="full-duplex" neighborhood_improvements=0
2026-09-24T06:56:33.465958Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:schedule_provisional_exchanges: finished incremental instruction-alignment validation transfers=29337 elapsed_ms=1360 success=true
2026-09-24T06:56:33.784327Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:schedule_provisional_exchanges: finished incremental instruction-alignment validation transfers=11124 elapsed_ms=1472 success=true
2026-09-24T06:56:34.849909Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:schedule_provisional_exchanges: retrying exchange schedule with incremental instruction-alignment validation transfers=29337
2026-09-24T06:56:36.261134Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:schedule_provisional_exchanges: finished incremental instruction-alignment validation transfers=29337 elapsed_ms=1411 success=true
2026-09-24T06:56:36.551703Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:schedule_provisional_exchanges: compared ordinary and paired exchange schedules phase=1 candidates=1440 ordinary_horizon=18406 paired_horizon=13809 use_paired=true
2026-09-24T06:56:36.624092Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:schedule_provisional_exchanges: prepared large physical exchange phase phase=1 transfers=11124 maximum_identical_destinations=1
2026-09-24T06:56:36.624111Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:schedule_provisional_exchanges: optimized physical exchange schedule phase=1 initial_horizon=13809 selected_horizon=13809 endpoint_lower_bound=10044 lower_bound_gap=3765 selected_kind="full-duplex" neighborhood_improvements=0
2026-09-24T06:56:37.261763Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:schedule_provisional_exchanges: compared ordinary and paired exchange schedules phase=4 candidates=4320 ordinary_horizon=23505 paired_horizon=15659 use_paired=true
2026-09-24T06:56:37.367133Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:schedule_provisional_exchanges: prepared large physical exchange phase phase=4 transfers=29337 maximum_identical_destinations=1
2026-09-24T06:56:37.367163Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:schedule_provisional_exchanges: optimized physical exchange schedule phase=4 initial_horizon=15659 selected_horizon=15659 endpoint_lower_bound=11520 lower_bound_gap=4139 selected_kind="full-duplex" neighborhood_improvements=0
2026-09-24T06:56:37.433617Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:schedule_provisional_exchanges: close time.busy=6.16s time.idle=3.07µs
2026-09-24T06:56:37.464042Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:plan_kernels: close time.busy=30.4ms time.idle=2.11µs
2026-09-24T06:56:37.464396Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:compile_kernels: compiling kernel source=/srv/home/gc-sdk/ipu-stack/crates/ipu-tests/../../device/rearrange_f8.cpp name="rearrange_row_major_to_amp_f8_o0_r20_p20_c32_p32_codelet" target="ipu21"
2026-09-24T06:56:37.854931Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:compile_kernels: kernel artifact created object=/srv/home/.cache/ipu-stack/kernels/21eba6af44230dfe1261c963a9c22e6da4be1fe3d98366a3b17bbc4c51d176c8/kernel.o metadata=/srv/home/.cache/ipu-stack/kernels/21eba6af44230dfe1261c963a9c22e6da4be1fe3d98366a3b17bbc4c51d176c8/kernel.json
2026-09-24T06:56:37.855522Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:compile_kernels: compiling kernel source=/srv/home/gc-sdk/ipu-stack/crates/ipu-tests/../../device/worker_call.S name="rearrange_row_major_to_amp_f8_o0_r20_p20_c32_p32_wrapper" target="ipu21"
2026-09-24T06:56:37.903499Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:compile_kernels: kernel artifact created object=/srv/home/.cache/ipu-stack/kernels/2498d3fe24026a9939ecc4cbc29b422a5aacffc4d0bc72513840bc3a335ba99d/kernel.o metadata=/srv/home/.cache/ipu-stack/kernels/2498d3fe24026a9939ecc4cbc29b422a5aacffc4d0bc72513840bc3a335ba99d/kernel.json
2026-09-24T06:56:37.904009Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:compile_kernels: compiling kernel source=/srv/home/gc-sdk/ipu-stack/crates/ipu-tests/../../device/rearrange_f8.cpp name="rearrange_row_major_to_amp_f8_o0_r16_p16_c32_p32_codelet" target="ipu21"
2026-09-24T06:56:38.206970Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:compile_kernels: kernel artifact created object=/srv/home/.cache/ipu-stack/kernels/552d016a20e2d004aebd1d01d5c81fc525b700640cab1a7f9328144b37d6fe17/kernel.o metadata=/srv/home/.cache/ipu-stack/kernels/552d016a20e2d004aebd1d01d5c81fc525b700640cab1a7f9328144b37d6fe17/kernel.json
2026-09-24T06:56:38.207349Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:compile_kernels: compiling kernel source=/srv/home/gc-sdk/ipu-stack/crates/ipu-tests/../../device/worker_call.S name="rearrange_row_major_to_amp_f8_o0_r16_p16_c32_p32_wrapper" target="ipu21"
2026-09-24T06:56:38.251502Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:compile_kernels: kernel artifact created object=/srv/home/.cache/ipu-stack/kernels/70166a3d62424e0c0b6e8fa2b3cf24c435d01b40cf707d4170964439dc28c8e8/kernel.o metadata=/srv/home/.cache/ipu-stack/kernels/70166a3d62424e0c0b6e8fa2b3cf24c435d01b40cf707d4170964439dc28c8e8/kernel.json
2026-09-24T06:56:38.251711Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:compile_kernels: compiling kernel source=/srv/home/gc-sdk/ipu-stack/crates/ipu-tests/../../device/rearrange_f8.cpp name="rearrange_row_major_to_amp_f8_o0_r13_p16_c32_p32_codelet" target="ipu21"
2026-09-24T06:56:38.559971Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:compile_kernels: kernel artifact created object=/srv/home/.cache/ipu-stack/kernels/00e1f11d0162ada395f11eed645a9425f87fe344073f3672e8ab91184ab9bb56/kernel.o metadata=/srv/home/.cache/ipu-stack/kernels/00e1f11d0162ada395f11eed645a9425f87fe344073f3672e8ab91184ab9bb56/kernel.json
2026-09-24T06:56:38.560126Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:compile_kernels: compiling kernel source=/srv/home/gc-sdk/ipu-stack/crates/ipu-tests/../../device/worker_call.S name="rearrange_row_major_to_amp_f8_o0_r13_p16_c32_p32_wrapper" target="ipu21"
2026-09-24T06:56:38.606282Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:compile_kernels: kernel artifact created object=/srv/home/.cache/ipu-stack/kernels/e14e1144e733b49334745bf5b20b3c4a80a7a543f546661b203ae9013d121e0f/kernel.o metadata=/srv/home/.cache/ipu-stack/kernels/e14e1144e733b49334745bf5b20b3c4a80a7a543f546661b203ae9013d121e0f/kernel.json
2026-09-24T06:56:38.607281Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:compile_kernels: compiling kernel source=/srv/home/gc-sdk/ipu-stack/crates/ipu-tests/../../device/gemm_f16_amp.S name="gemm_f8_interleaved_packed0_k288_c48_r183" target="ipu21"
2026-09-24T06:56:38.654938Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:compile_kernels: kernel artifact created object=/srv/home/.cache/ipu-stack/kernels/1f6bb9b5ddfe8103836c97b36d33b43aa1df70d83899c183f93f83eec024f7ea/kernel.o metadata=/srv/home/.cache/ipu-stack/kernels/1f6bb9b5ddfe8103836c97b36d33b43aa1df70d83899c183f93f83eec024f7ea/kernel.json
2026-09-24T06:56:38.655458Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:compile_kernels: compiling kernel source=/srv/home/gc-sdk/ipu-stack/crates/ipu-tests/../../device/gemm_f16_amp.S name="gemm_f8_interleaved_packed0_k288_c48_r182" target="ipu21"
2026-09-24T06:56:38.702468Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:compile_kernels: kernel artifact created object=/srv/home/.cache/ipu-stack/kernels/e6e19d29e7106d46b9e81b30c627c5533782ba64a25474bd216e1575875e3d49/kernel.o metadata=/srv/home/.cache/ipu-stack/kernels/e6e19d29e7106d46b9e81b30c627c5533782ba64a25474bd216e1575875e3d49/kernel.json
2026-09-24T06:56:38.703812Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:compile_kernels: close time.busy=1.24s time.idle=5.67µs
2026-09-24T06:56:38.704286Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:link_runtime: close time.busy=325µs time.idle=2.14µs
2026-09-24T06:56:38.752724Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}: exchange table and repeat patch storage exchange_table_bytes=4576 maximum_uncompressed_repeat_patch_bytes=0 maximum_elided_arithmetic_patch_bytes=0
2026-09-24T06:56:38.889151Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:size_tile_code: close time.busy=6.81ms time.idle=4.20µs
2026-09-24T06:56:38.889183Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}: placing generated tile programs linked_end=318824 host_code_base=376832 host_code_bytes=532 generated_code_bytes=980 exchange_table_bytes=4576 executable_ranges=[(377432, 524288)]
2026-09-24T06:56:38.889253Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}: allocated package support memory linked_end=318824 profile_bytes=80 exchange_table_bytes=4576 host_code_bytes=532 host_data_bytes=352 generated_code_bytes=980 code_address=377432 available_ranges=[(327680, 360448), (362560, 376832), (409600, 949168)]
2026-09-24T06:56:38.948521Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:place_storage: close time.busy=45.9ms time.idle=2.91µs
2026-09-24T06:56:39.080087Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:lower_exchanges: reused exchange optimization after validating relocated rows phase=2
2026-09-24T06:56:39.085682Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:lower_exchanges: prepared large physical exchange phase phase=2 transfers=16140 maximum_identical_destinations=1
2026-09-24T06:56:39.085698Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:lower_exchanges: optimized physical exchange schedule phase=2 initial_horizon=4127 selected_horizon=4127 endpoint_lower_bound=3384 lower_bound_gap=743 selected_kind="reused" neighborhood_improvements=0
2026-09-24T06:56:39.204394Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:lower_exchanges: reused exchange optimization after validating relocated rows phase=5
2026-09-24T06:56:39.216351Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:lower_exchanges: prepared large physical exchange phase phase=5 transfers=33732 maximum_identical_destinations=1
2026-09-24T06:56:39.216372Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:lower_exchanges: optimized physical exchange schedule phase=5 initial_horizon=7944 selected_horizon=7944 endpoint_lower_bound=6912 lower_bound_gap=1032 selected_kind="reused" neighborhood_improvements=0
2026-09-24T06:56:39.406604Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:lower_exchanges: reused exchange optimization after validating relocated rows phase=3
2026-09-24T06:56:39.416399Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:lower_exchanges: prepared large physical exchange phase phase=3 transfers=27712 maximum_identical_destinations=1
2026-09-24T06:56:39.416415Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:lower_exchanges: optimized physical exchange schedule phase=3 initial_horizon=2071 selected_horizon=2071 endpoint_lower_bound=1104 lower_bound_gap=967 selected_kind="reused" neighborhood_improvements=0
2026-09-24T06:56:39.436425Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:lower_exchanges: reused exchange optimization after validating relocated rows phase=0
2026-09-24T06:56:39.446322Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:lower_exchanges: prepared large physical exchange phase phase=0 transfers=27380 maximum_identical_destinations=1
2026-09-24T06:56:39.446337Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:lower_exchanges: optimized physical exchange schedule phase=0 initial_horizon=341 selected_horizon=341 endpoint_lower_bound=160 lower_bound_gap=181 selected_kind="reused" neighborhood_improvements=0
2026-09-24T06:56:40.879694Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:lower_exchanges: retrying exchange schedule with incremental instruction-alignment validation transfers=11124
2026-09-24T06:56:42.060194Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:lower_exchanges: retrying exchange schedule with incremental instruction-alignment validation transfers=29337
2026-09-24T06:56:42.209757Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:lower_exchanges: finished incremental instruction-alignment validation transfers=11124 elapsed_ms=1330 success=true
2026-09-24T06:56:43.886575Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:lower_exchanges: finished incremental instruction-alignment validation transfers=29337 elapsed_ms=1826 success=true
2026-09-24T06:56:45.658005Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:lower_exchanges: compared ordinary and paired exchange schedules phase=1 candidates=1440 ordinary_horizon=17273 paired_horizon=13150 use_paired=true
2026-09-24T06:56:45.756940Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:lower_exchanges: prepared large physical exchange phase phase=1 transfers=11124 maximum_identical_destinations=1
2026-09-24T06:56:45.756972Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:lower_exchanges: optimized physical exchange schedule phase=1 initial_horizon=13150 selected_horizon=13150 endpoint_lower_bound=10044 lower_bound_gap=3106 selected_kind="full-duplex" neighborhood_improvements=0
2026-09-24T06:56:46.353539Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:lower_exchanges: retrying exchange schedule with incremental instruction-alignment validation transfers=29337
2026-09-24T06:56:48.630955Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:lower_exchanges: finished incremental instruction-alignment validation transfers=29337 elapsed_ms=2277 success=true
2026-09-24T06:56:50.046518Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:lower_exchanges: compared ordinary and paired exchange schedules phase=4 candidates=4320 ordinary_horizon=19382 paired_horizon=15121 use_paired=true
2026-09-24T06:56:50.158837Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:lower_exchanges: prepared large physical exchange phase phase=4 transfers=29337 maximum_identical_destinations=1
2026-09-24T06:56:50.158881Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:lower_exchanges: optimized physical exchange schedule phase=4 initial_horizon=15121 selected_horizon=15121 endpoint_lower_bound=11520 lower_bound_gap=3601 selected_kind="full-duplex" neighborhood_improvements=0
2026-09-24T06:56:50.227137Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:lower_exchanges: close time.busy=11.3s time.idle=6.07µs
2026-09-24T06:56:50.238521Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}: costed final placed program final_cycles=131548 final_exchange=46354
2026-09-24T06:56:50.353896Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:prepare_tile_code: close time.busy=8.95ms time.idle=6.24µs
2026-09-24T06:56:50.369591Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:emit_tile_code: close time.busy=15.6ms time.idle=2.82µs
2026-09-24T06:56:50.517521Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:build_tile_images: close time.busy=141ms time.idle=5.11µs
2026-09-24T06:56:50.619966Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}: close time.busy=253s time.idle=4.52µs
2026-09-24T06:56:50.624994Z  INFO writing application package tiles=1472
2026-09-24T06:56:50.684337Z  INFO application package written
2026-09-24T06:56:50.684380Z  INFO reading application package
2026-09-24T06:56:50.788059Z  INFO application package read tiles=1472 compiler=0.1.0
2026-09-24T06:56:50.788137Z  INFO waiting for hardware ownership path=/tmp/ipu-stack-device.lock
2026-09-24T06:56:50.788146Z  INFO acquired hardware ownership path=/tmp/ipu-stack-device.lock
2026-09-24T06:56:50.788157Z  INFO opening IPU device path="/dev/ipu0"
2026-09-24T06:56:50.788236Z  INFO IPU device attached path=/dev/ipu0
2026-09-24T06:56:50.788241Z  INFO initializing IPU device
2026-09-24T06:56:50.796271Z  INFO IPU device initialized
2026-09-24T06:56:50.796277Z  INFO replaying device configuration register_writes=2003
2026-09-24T06:56:50.802095Z  INFO loading application tile_count=1472 final_mark=23
2026-09-24T06:56:51.001865Z  INFO application loaded and initial synchronization completed tile_count=1472
2026-09-24T06:56:51.821090Z  INFO prepared reference inputs elapsed_ms=815
2026-09-24T06:56:52.153174Z  INFO configured host reference BLAS threads=4
referencePrecision=device
2026-09-24T06:56:52.743176Z  INFO evaluated host reference elapsed_ms=922
2026-09-24T06:56:52.746307Z  INFO starting host exchange session startup_mark=23
2026-09-24T06:56:52.746327Z  INFO attaching host exchange pages pages=2
2026-09-24T06:56:52.747129Z  INFO host exchange session started
2026-09-24T06:56:52.747229Z  INFO invoking host exchange call call="initialize" command=0 phases=4
2026-09-24T06:56:52.753233Z  INFO invoking host exchange call call="run" command=0 phases=11
referenceInference=1 numericalTest=PASS
profile=artifacts/mlp-prepacking-20260924/f8.capnp tiles=1472 samples=21050
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=1472 executionTiles=1472 inputBytes=839808 weightBytes=9934848 cycles=132618 minimumTileCycles=127956 deviceMicroseconds=88.412 effectiveGemmTflops=163.531 peakTflops=282.624 efficiencyPercent=57.86 maximumAbsoluteError=0.005493
package=artifacts/mlp-prepacking-20260924/f8.ipuexe tiles=1472 hardwareTest=PASS