2026-09-24T09:56:09.906443Z  INFO ipu_codegen.package.build{tile_count=64 operations=3}:compile_runtime: close time.busy=316µs time.idle=1.29µs
2026-09-24T09:56:10.033041Z  INFO ipu_codegen.package.build{tile_count=64 operations=3}:construct_mid:gemm_candidates: close time.busy=126ms time.idle=2.56µs
2026-09-24T09:56:10.162184Z  INFO ipu_codegen.package.build{tile_count=64 operations=3}:construct_mid:gemm_candidates: close time.busy=129ms time.idle=4.76µs
2026-09-24T09:56:10.253755Z DEBUG ipu_codegen.package.build{tile_count=64 operations=3}:construct_mid: selected local reduction seeds swapped=20 copies=24
2026-09-24T09:56:10.253842Z  INFO ipu_codegen.package.build{tile_count=64 operations=3}:construct_mid:group_exchanges: close time.busy=4.09µs time.idle=2.15µs
2026-09-24T09:56:10.253860Z  INFO ipu_codegen.package.build{tile_count=64 operations=3}:construct_mid:merge_copies: close time.busy=9.52µs time.idle=343ns
2026-09-24T09:56:10.254021Z  INFO ipu_codegen.package.build{tile_count=64 operations=3}:construct_mid:eliminate: close time.busy=156µs time.idle=371ns
2026-09-24T09:56:10.254672Z  INFO ipu_codegen.package.build{tile_count=64 operations=3}:construct_mid: costed planning finalist rank=0 mid_cycles=12476 low_cycles=9712 elapsed_ms=4
2026-09-24T09:56:10.258632Z DEBUG ipu_codegen.package.build{tile_count=64 operations=3}:construct_mid: selected local reduction seeds swapped=20 copies=24
2026-09-24T09:56:10.258708Z  INFO ipu_codegen.package.build{tile_count=64 operations=3}:construct_mid:group_exchanges: close time.busy=3.52µs time.idle=1.27µs
2026-09-24T09:56:10.258722Z  INFO ipu_codegen.package.build{tile_count=64 operations=3}:construct_mid:merge_copies: close time.busy=5.77µs time.idle=341ns
2026-09-24T09:56:10.258869Z  INFO ipu_codegen.package.build{tile_count=64 operations=3}:construct_mid:eliminate: close time.busy=143µs time.idle=384ns
2026-09-24T09:56:10.259370Z  INFO ipu_codegen.package.build{tile_count=64 operations=3}:construct_mid: costed planning finalist rank=1 mid_cycles=12508 low_cycles=9726 elapsed_ms=3
2026-09-24T09:56:10.262683Z DEBUG ipu_codegen.package.build{tile_count=64 operations=3}:construct_mid: selected local reduction seeds swapped=0 copies=0
2026-09-24T09:56:10.262760Z  INFO ipu_codegen.package.build{tile_count=64 operations=3}:construct_mid:group_exchanges: close time.busy=4.19µs time.idle=1.02µs
2026-09-24T09:56:10.262776Z  INFO ipu_codegen.package.build{tile_count=64 operations=3}:construct_mid:merge_copies: close time.busy=8.79µs time.idle=325ns
2026-09-24T09:56:10.262942Z  INFO ipu_codegen.package.build{tile_count=64 operations=3}:construct_mid:eliminate: close time.busy=162µs time.idle=281ns
2026-09-24T09:56:10.263365Z  INFO ipu_codegen.package.build{tile_count=64 operations=3}:construct_mid: costed planning finalist rank=2 mid_cycles=12654 low_cycles=10718 elapsed_ms=3
2026-09-24T09:56:10.267425Z DEBUG ipu_codegen.package.build{tile_count=64 operations=3}:construct_mid: selected local reduction seeds swapped=0 copies=0
2026-09-24T09:56:10.267529Z  INFO ipu_codegen.package.build{tile_count=64 operations=3}:construct_mid:group_exchanges: close time.busy=6.41µs time.idle=1.23µs
2026-09-24T09:56:10.267551Z  INFO ipu_codegen.package.build{tile_count=64 operations=3}:construct_mid:merge_copies: close time.busy=12.0µs time.idle=469ns
2026-09-24T09:56:10.267784Z  INFO ipu_codegen.package.build{tile_count=64 operations=3}:construct_mid:eliminate: close time.busy=227µs time.idle=465ns
2026-09-24T09:56:10.268382Z  INFO ipu_codegen.package.build{tile_count=64 operations=3}:construct_mid: costed planning finalist rank=3 mid_cycles=12670 low_cycles=10734 elapsed_ms=4
2026-09-24T09:56:10.272513Z DEBUG ipu_codegen.package.build{tile_count=64 operations=3}:construct_mid: selected local reduction seeds swapped=20 copies=24
2026-09-24T09:56:10.272586Z  INFO ipu_codegen.package.build{tile_count=64 operations=3}:construct_mid:group_exchanges: close time.busy=3.73µs time.idle=1.15µs
2026-09-24T09:56:10.272602Z  INFO ipu_codegen.package.build{tile_count=64 operations=3}:construct_mid:merge_copies: close time.busy=8.74µs time.idle=388ns
2026-09-24T09:56:10.272781Z  INFO ipu_codegen.package.build{tile_count=64 operations=3}:construct_mid:eliminate: close time.busy=165µs time.idle=397ns
2026-09-24T09:56:10.273340Z  INFO ipu_codegen.package.build{tile_count=64 operations=3}:construct_mid: costed planning finalist rank=4 mid_cycles=12688 low_cycles=9924 elapsed_ms=3
2026-09-24T09:56:10.277127Z DEBUG ipu_codegen.package.build{tile_count=64 operations=3}:construct_mid: selected local reduction seeds swapped=20 copies=24
2026-09-24T09:56:10.277197Z  INFO ipu_codegen.package.build{tile_count=64 operations=3}:construct_mid:group_exchanges: close time.busy=3.21µs time.idle=1.13µs
2026-09-24T09:56:10.277228Z  INFO ipu_codegen.package.build{tile_count=64 operations=3}:construct_mid:merge_copies: close time.busy=6.01µs time.idle=414ns
2026-09-24T09:56:10.277382Z  INFO ipu_codegen.package.build{tile_count=64 operations=3}:construct_mid:eliminate: close time.busy=150µs time.idle=361ns
2026-09-24T09:56:10.277869Z  INFO ipu_codegen.package.build{tile_count=64 operations=3}:construct_mid: costed planning finalist rank=5 mid_cycles=12720 low_cycles=9938 elapsed_ms=3
2026-09-24T09:56:10.283916Z DEBUG ipu_codegen.package.build{tile_count=64 operations=3}:construct_mid: selected local reduction seeds swapped=44 copies=46
2026-09-24T09:56:10.284010Z  INFO ipu_codegen.package.build{tile_count=64 operations=3}:construct_mid:group_exchanges: close time.busy=3.88µs time.idle=1.16µs
2026-09-24T09:56:10.284033Z  INFO ipu_codegen.package.build{tile_count=64 operations=3}:construct_mid:merge_copies: close time.busy=13.4µs time.idle=402ns
2026-09-24T09:56:10.284254Z  INFO ipu_codegen.package.build{tile_count=64 operations=3}:construct_mid:eliminate: close time.busy=215µs time.idle=371ns
2026-09-24T09:56:10.284894Z  INFO ipu_codegen.package.build{tile_count=64 operations=3}:construct_mid: costed planning finalist rank=6 mid_cycles=12810 low_cycles=9364 elapsed_ms=6
2026-09-24T09:56:10.290789Z DEBUG ipu_codegen.package.build{tile_count=64 operations=3}:construct_mid: selected local reduction seeds swapped=44 copies=46
2026-09-24T09:56:10.290879Z  INFO ipu_codegen.package.build{tile_count=64 operations=3}:construct_mid:group_exchanges: close time.busy=3.82µs time.idle=1.11µs
2026-09-24T09:56:10.290894Z  INFO ipu_codegen.package.build{tile_count=64 operations=3}:construct_mid:merge_copies: close time.busy=7.69µs time.idle=297ns
2026-09-24T09:56:10.291095Z  INFO ipu_codegen.package.build{tile_count=64 operations=3}:construct_mid:eliminate: close time.busy=196µs time.idle=359ns
2026-09-24T09:56:10.291690Z  INFO ipu_codegen.package.build{tile_count=64 operations=3}:construct_mid: costed planning finalist rank=7 mid_cycles=12842 low_cycles=9378 elapsed_ms=5
2026-09-24T09:56:10.291983Z  INFO ipu_codegen.package.build{tile_count=64 operations=3}:construct_mid: close time.busy=386ms time.idle=576ns
2026-09-24T09:56:10.296864Z DEBUG ipu_codegen.package.build{tile_count=64 operations=3}: selected local reduction seeds swapped=44 copies=46
2026-09-24T09:56:10.296952Z  INFO ipu_codegen.package.build{tile_count=64 operations=3}:group_exchanges: close time.busy=3.99µs time.idle=795ns
2026-09-24T09:56:10.296968Z  INFO ipu_codegen.package.build{tile_count=64 operations=3}:merge_copies: close time.busy=9.54µs time.idle=334ns
2026-09-24T09:56:10.297175Z  INFO ipu_codegen.package.build{tile_count=64 operations=3}:eliminate: close time.busy=203µs time.idle=312ns
2026-09-24T09:56:10.299543Z  INFO ipu_codegen.package.build{tile_count=64 operations=3}:place_provisional_storage: close time.busy=2.35ms time.idle=3.78µs
2026-09-24T09:56:10.304324Z  INFO ipu_codegen.package.build{tile_count=64 operations=3}:schedule_provisional_exchanges: retrying exchange schedule with incremental instruction-alignment validation transfers=258
2026-09-24T09:56:10.309435Z  INFO ipu_codegen.package.build{tile_count=64 operations=3}:schedule_provisional_exchanges: compared ordinary and paired exchange schedules phase=1 candidates=22 ordinary_horizon=590 paired_horizon=719 use_paired=false
2026-09-24T09:56:10.310090Z  INFO ipu_codegen.package.build{tile_count=64 operations=3}:schedule_provisional_exchanges: finished incremental instruction-alignment validation transfers=258 elapsed_ms=5 success=true
2026-09-24T09:56:10.315782Z  INFO ipu_codegen.package.build{tile_count=64 operations=3}:schedule_provisional_exchanges: retrying exchange schedule with incremental instruction-alignment validation transfers=258
2026-09-24T09:56:10.321775Z  INFO ipu_codegen.package.build{tile_count=64 operations=3}:schedule_provisional_exchanges: finished incremental instruction-alignment validation transfers=258 elapsed_ms=5 success=true
2026-09-24T09:56:10.326259Z  INFO ipu_codegen.package.build{tile_count=64 operations=3}:schedule_provisional_exchanges: compared ordinary and paired exchange schedules phase=3 candidates=60 ordinary_horizon=580 paired_horizon=732 use_paired=false
2026-09-24T09:56:10.328295Z  INFO ipu_codegen.package.build{tile_count=64 operations=3}:schedule_provisional_exchanges: close time.busy=28.7ms time.idle=2.85µs
2026-09-24T09:56:10.329103Z  INFO ipu_codegen.package.build{tile_count=64 operations=3}:plan_kernels: close time.busy=783µs time.idle=1.30µs
2026-09-24T09:56:10.329326Z  INFO ipu_codegen.package.build{tile_count=64 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_r4_p4_c32_p32_codelet" target="ipu21"
2026-09-24T09:56:10.551996Z  INFO ipu_codegen.package.build{tile_count=64 operations=3}:compile_kernels: kernel artifact created object=/srv/home/.cache/ipu-stack/kernels/970669ccd422ec43173f01577a83e08ea6bb68e75a2948da5be00808556d6ea2/kernel.o metadata=/srv/home/.cache/ipu-stack/kernels/970669ccd422ec43173f01577a83e08ea6bb68e75a2948da5be00808556d6ea2/kernel.json
2026-09-24T09:56:10.552117Z  INFO ipu_codegen.package.build{tile_count=64 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_r4_p4_c32_p32_wrapper" target="ipu21"
2026-09-24T09:56:10.597948Z  INFO ipu_codegen.package.build{tile_count=64 operations=3}:compile_kernels: kernel artifact created object=/srv/home/.cache/ipu-stack/kernels/094e2acc00b0767eadeb32b83b065aedef2729fcc234419eaeb42c0be5174ed5/kernel.o metadata=/srv/home/.cache/ipu-stack/kernels/094e2acc00b0767eadeb32b83b065aedef2729fcc234419eaeb42c0be5174ed5/kernel.json
2026-09-24T09:56:10.598476Z  INFO ipu_codegen.package.build{tile_count=64 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_k64_c16_r7" target="ipu21"
2026-09-24T09:56:10.640289Z  INFO ipu_codegen.package.build{tile_count=64 operations=3}:compile_kernels: kernel artifact created object=/srv/home/.cache/ipu-stack/kernels/c5bbe76250d293cd020077d4c123e2d74a172c43f4a77ca44458d89aae539079/kernel.o metadata=/srv/home/.cache/ipu-stack/kernels/c5bbe76250d293cd020077d4c123e2d74a172c43f4a77ca44458d89aae539079/kernel.json
2026-09-24T09:56:10.640584Z  INFO ipu_codegen.package.build{tile_count=64 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_k64_c16_r6" target="ipu21"
2026-09-24T09:56:10.686895Z  INFO ipu_codegen.package.build{tile_count=64 operations=3}:compile_kernels: kernel artifact created object=/srv/home/.cache/ipu-stack/kernels/03108233e1b68154da5b2ad299ddebc88b534c78013301818f064bad3799db96/kernel.o metadata=/srv/home/.cache/ipu-stack/kernels/03108233e1b68154da5b2ad299ddebc88b534c78013301818f064bad3799db96/kernel.json
2026-09-24T09:56:10.687905Z  INFO ipu_codegen.package.build{tile_count=64 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_k32_c16_r14" target="ipu21"
2026-09-24T09:56:10.727249Z  INFO ipu_codegen.package.build{tile_count=64 operations=3}:compile_kernels: kernel artifact created object=/srv/home/.cache/ipu-stack/kernels/497e3912c9b7e015a9f559cc1726670556a8ca16a4d80a228e9d593792df1dc0/kernel.o metadata=/srv/home/.cache/ipu-stack/kernels/497e3912c9b7e015a9f559cc1726670556a8ca16a4d80a228e9d593792df1dc0/kernel.json
2026-09-24T09:56:10.727533Z  INFO ipu_codegen.package.build{tile_count=64 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_k32_c16_r12" target="ipu21"
2026-09-24T09:56:10.769578Z  INFO ipu_codegen.package.build{tile_count=64 operations=3}:compile_kernels: kernel artifact created object=/srv/home/.cache/ipu-stack/kernels/49346a21deca6926074eb8a76c58f55bd4741f7f69e16614f552c510e696ce05/kernel.o metadata=/srv/home/.cache/ipu-stack/kernels/49346a21deca6926074eb8a76c58f55bd4741f7f69e16614f552c510e696ce05/kernel.json
2026-09-24T09:56:10.769805Z  INFO ipu_codegen.package.build{tile_count=64 operations=3}:compile_kernels: close time.busy=441ms time.idle=3.67µs
2026-09-24T09:56:10.770023Z  INFO ipu_codegen.package.build{tile_count=64 operations=3}:link_runtime: close time.busy=186µs time.idle=1.78µs
2026-09-24T09:56:10.772342Z  INFO ipu_codegen.package.build{tile_count=64 operations=3}: exchange table and repeat patch storage exchange_table_bytes=576 maximum_uncompressed_repeat_patch_bytes=0 maximum_elided_arithmetic_patch_bytes=0
2026-09-24T09:56:10.801545Z  INFO ipu_codegen.package.build{tile_count=64 operations=3}:size_tile_code: close time.busy=2.18ms time.idle=4.09µs
2026-09-24T09:56:10.801567Z  INFO ipu_codegen.package.build{tile_count=64 operations=3}: placing generated tile programs linked_end=317752 host_code_base=376832 host_code_bytes=412 generated_code_bytes=756 exchange_table_bytes=576 executable_ranges=[(377312, 524288)]
2026-09-24T09:56:10.801591Z  INFO ipu_codegen.package.build{tile_count=64 operations=3}: allocated package support memory linked_end=317752 profile_bytes=60 exchange_table_bytes=576 host_code_bytes=412 host_data_bytes=228 generated_code_bytes=756 code_address=377312 available_ranges=[(327680, 360448), (362436, 376832), (409600, 949168)]
2026-09-24T09:56:10.804957Z  INFO ipu_codegen.package.build{tile_count=64 operations=3}:place_storage: close time.busy=2.25ms time.idle=1.57µs
2026-09-24T09:56:10.806624Z  INFO ipu_codegen.package.build{tile_count=64 operations=3}:lower_exchanges: reused exchange optimization after validating relocated rows phase=0
2026-09-24T09:56:10.807468Z  INFO ipu_codegen.package.build{tile_count=64 operations=3}:lower_exchanges: reused exchange optimization after validating relocated rows phase=1
2026-09-24T09:56:10.807911Z  INFO ipu_codegen.package.build{tile_count=64 operations=3}:lower_exchanges: reused exchange optimization after validating relocated rows phase=2
2026-09-24T09:56:10.808483Z  INFO ipu_codegen.package.build{tile_count=64 operations=3}:lower_exchanges: reused exchange optimization after validating relocated rows phase=4
2026-09-24T09:56:10.815035Z  INFO ipu_codegen.package.build{tile_count=64 operations=3}:lower_exchanges: reused exchange optimization after validating relocated rows phase=3
2026-09-24T09:56:10.815861Z  INFO ipu_codegen.package.build{tile_count=64 operations=3}:lower_exchanges: close time.busy=10.9ms time.idle=1.07µs
2026-09-24T09:56:10.815938Z  INFO ipu_codegen.package.build{tile_count=64 operations=3}: costed final placed program final_cycles=10080 final_exchange=4652
2026-09-24T09:56:10.824725Z  INFO ipu_codegen.package.build{tile_count=64 operations=3}:prepare_tile_code: close time.busy=2.01ms time.idle=2.79µs
2026-09-24T09:56:10.829766Z  INFO ipu_codegen.package.build{tile_count=64 operations=3}:emit_tile_code: close time.busy=5.01ms time.idle=1.91µs
2026-09-24T09:56:11.009011Z  INFO ipu_codegen.package.build{tile_count=64 operations=3}:build_tile_images: close time.busy=178ms time.idle=3.20µs
2026-09-24T09:56:11.014531Z  INFO ipu_codegen.package.build{tile_count=64 operations=3}: close time.busy=1.11s time.idle=2.93µs
2026-09-24T09:56:11.016062Z  INFO writing application package tiles=1472
2026-09-24T09:56:11.033601Z  INFO application package written
2026-09-24T09:56:11.035036Z  INFO reading application package
2026-09-24T09:56:11.060921Z  INFO application package read tiles=1472 compiler=0.1.0
2026-09-24T09:56:11.061566Z  INFO waiting for hardware ownership path=/tmp/ipu-stack-device.lock
2026-09-24T09:56:11.061576Z  INFO acquired hardware ownership path=/tmp/ipu-stack-device.lock
2026-09-24T09:56:11.061581Z  INFO opening IPU device path="/dev/ipu0"
2026-09-24T09:56:11.061610Z  INFO IPU device attached path=/dev/ipu0
2026-09-24T09:56:11.061613Z  INFO initializing IPU device
2026-09-24T09:56:11.069599Z  INFO IPU device initialized
2026-09-24T09:56:11.069622Z  INFO replaying device configuration register_writes=2003
2026-09-24T09:56:11.073913Z  INFO loading application tile_count=1472 final_mark=23
2026-09-24T09:56:11.251878Z  INFO application loaded and initial synchronization completed tile_count=1472
2026-09-24T09:56:11.259391Z  INFO prepared reference inputs elapsed_ms=4
2026-09-24T09:56:11.260303Z  INFO configured host reference BLAS threads=4
referencePrecision=device
2026-09-24T09:56:11.261483Z  INFO evaluated host reference elapsed_ms=2
2026-09-24T09:56:11.262871Z  INFO starting host exchange session startup_mark=23
2026-09-24T09:56:11.262879Z  INFO attaching host exchange pages pages=2
2026-09-24T09:56:11.263542Z  INFO host exchange session started
2026-09-24T09:56:11.263547Z  INFO invoking host exchange call call="initialize" command=0 phases=4
2026-09-24T09:56:11.264055Z  INFO invoking host exchange call call="run" command=0 phases=9
profile=artifacts/reduction-selection-20260924/small-f8.capnp tiles=1472 samples=7709
referenceInference=1 numericalTest=PASS
referenceMaximumAbsoluteError=0.005859 numericalTest=PASS
package=artifacts/reduction-selection-20260924/small-f8.ipuexe tiles=1472 hardwareTest=PASS