2026-09-23T10:18:11.443123Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:compile_runtime: close time.busy=295µs time.idle=6.45µs
2026-09-23T10:18:20.794515Z DEBUG GEMM model frontier assignments=617324 retained=177
2026-09-23T10:18:20.794656Z DEBUG GEMM model frontier assignments=617324 retained=175
2026-09-23T10:18:29.917626Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:construct_mid:gemm_candidates: close time.busy=18.5s time.idle=4.82µs
2026-09-23T10:18:40.299923Z DEBUG GEMM model frontier assignments=649728 retained=144
2026-09-23T10:18:40.300371Z DEBUG GEMM model frontier assignments=649728 retained=144
2026-09-23T10:18:40.322095Z DEBUG GEMM model frontier assignments=649728 retained=144
2026-09-23T10:18:40.322446Z DEBUG GEMM model frontier assignments=649728 retained=144
2026-09-23T10:18:48.278072Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:construct_mid:gemm_candidates: close time.busy=18.4s time.idle=6.33µs
2026-09-23T10:19:15.445079Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:construct_mid:group_exchanges: close time.busy=895µs time.idle=5.76µs
2026-09-23T10:19:15.458973Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:construct_mid:merge_copies: close time.busy=13.4ms time.idle=2.10µs
2026-09-23T10:19:15.554803Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:construct_mid:eliminate: eliminated padding clears using kernel access contracts removed=0
2026-09-23T10:19:15.557911Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:construct_mid:eliminate: close time.busy=98.9ms time.idle=1.47µs
2026-09-23T10:19:16.002573Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:construct_mid: costed planning finalist rank=0 mid_cycles=163242 low_cycles=188565 elapsed_ms=2907
2026-09-23T10:19:18.687054Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:construct_mid:group_exchanges: close time.busy=976µs time.idle=5.86µs
2026-09-23T10:19:18.700452Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:construct_mid:merge_copies: close time.busy=12.9ms time.idle=1.85µs
2026-09-23T10:19:18.805645Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:construct_mid:eliminate: eliminated padding clears using kernel access contracts removed=0
2026-09-23T10:19:18.808630Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:construct_mid:eliminate: close time.busy=108ms time.idle=3.78µs
2026-09-23T10:19:19.230993Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:construct_mid: costed planning finalist rank=1 mid_cycles=163482 low_cycles=188805 elapsed_ms=2911
2026-09-23T10:19:22.160701Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:construct_mid:group_exchanges: close time.busy=656µs time.idle=5.88µs
2026-09-23T10:19:22.172327Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:construct_mid:merge_copies: close time.busy=11.2ms time.idle=1.89µs
2026-09-23T10:19:22.274830Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:construct_mid:eliminate: eliminated padding clears using kernel access contracts removed=0
2026-09-23T10:19:22.277941Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:construct_mid:eliminate: close time.busy=106ms time.idle=2.49µs
2026-09-23T10:19:22.741049Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:construct_mid: costed planning finalist rank=2 mid_cycles=163947 low_cycles=188342 elapsed_ms=3179
2026-09-23T10:19:25.579495Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:construct_mid:group_exchanges: close time.busy=1.16ms time.idle=7.46µs
2026-09-23T10:19:25.594966Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:construct_mid:merge_copies: close time.busy=14.9ms time.idle=3.50µs
2026-09-23T10:19:25.723081Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:construct_mid:eliminate: eliminated padding clears using kernel access contracts removed=0
2026-09-23T10:19:25.726101Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:construct_mid:eliminate: close time.busy=131ms time.idle=3.09µs
2026-09-23T10:19:26.154756Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:construct_mid: costed planning finalist rank=3 mid_cycles=164187 low_cycles=188582 elapsed_ms=3080
2026-09-23T10:19:29.363264Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:construct_mid:group_exchanges: close time.busy=821µs time.idle=5.71µs
2026-09-23T10:19:29.385812Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:construct_mid:merge_copies: close time.busy=22.1ms time.idle=3.13µs
2026-09-23T10:19:29.496403Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:construct_mid:eliminate: eliminated padding clears using kernel access contracts removed=0
2026-09-23T10:19:29.499508Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:construct_mid:eliminate: close time.busy=114ms time.idle=2.49µs
2026-09-23T10:19:29.948997Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:construct_mid: costed planning finalist rank=4 mid_cycles=164567 low_cycles=194779 elapsed_ms=3420
2026-09-23T10:19:33.106283Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:construct_mid:group_exchanges: close time.busy=809µs time.idle=5.56µs
2026-09-23T10:19:33.124859Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:construct_mid:merge_copies: close time.busy=18.2ms time.idle=1.73µs
2026-09-23T10:19:33.233738Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:construct_mid:eliminate: eliminated padding clears using kernel access contracts removed=0
2026-09-23T10:19:33.236494Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:construct_mid:eliminate: close time.busy=112ms time.idle=3.02µs
2026-09-23T10:19:33.650314Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:construct_mid: costed planning finalist rank=5 mid_cycles=164807 low_cycles=195019 elapsed_ms=3346
2026-09-23T10:19:36.556555Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:construct_mid:group_exchanges: close time.busy=677µs time.idle=5.57µs
2026-09-23T10:19:36.569455Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:construct_mid:merge_copies: close time.busy=12.5ms time.idle=1.64µs
2026-09-23T10:19:36.659236Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:construct_mid:eliminate: eliminated padding clears using kernel access contracts removed=0
2026-09-23T10:19:36.661916Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:construct_mid:eliminate: close time.busy=92.4ms time.idle=2.52µs
2026-09-23T10:19:37.062132Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:construct_mid: costed planning finalist rank=6 mid_cycles=165671 low_cycles=187210 elapsed_ms=3099
2026-09-23T10:19:39.911281Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:construct_mid:group_exchanges: close time.busy=842µs time.idle=5.29µs
2026-09-23T10:19:39.921767Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:construct_mid:merge_copies: close time.busy=10.1ms time.idle=2.11µs
2026-09-23T10:19:40.008821Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:construct_mid:eliminate: eliminated padding clears using kernel access contracts removed=0
2026-09-23T10:19:40.011328Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:construct_mid:eliminate: close time.busy=89.5ms time.idle=1.15µs
2026-09-23T10:19:40.390623Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:construct_mid: costed planning finalist rank=7 mid_cycles=165927 low_cycles=187466 elapsed_ms=3041
2026-09-23T10:19:40.810069Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:construct_mid: close time.busy=89.4s time.idle=4.18µs
2026-09-23T10:19:42.960965Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:group_exchanges: close time.busy=949µs time.idle=5.41µs
2026-09-23T10:19:42.972462Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:merge_copies: close time.busy=11.0ms time.idle=1.71µs
2026-09-23T10:19:43.051001Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:eliminate: eliminated padding clears using kernel access contracts removed=0
2026-09-23T10:19:43.053682Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:eliminate: close time.busy=81.2ms time.idle=2.41µs
2026-09-23T10:19:43.127837Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:place_provisional_storage: close time.busy=71.9ms time.idle=3.94µs
2026-09-23T10:19:43.698688Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:schedule_provisional_exchanges: prepared large physical exchange phase phase=1 transfers=16836 maximum_identical_destinations=1
2026-09-23T10:19:43.698751Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:schedule_provisional_exchanges: optimized physical exchange schedule phase=1 initial_horizon=4712 selected_horizon=4712 endpoint_lower_bound=4320 lower_bound_gap=392 selected_kind="full-duplex" neighborhood_improvements=0
2026-09-23T10:19:43.762267Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:schedule_provisional_exchanges: retrying exchange schedule with incremental instruction-alignment validation transfers=40112
2026-09-23T10:19:43.960676Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:schedule_provisional_exchanges: retrying exchange schedule with incremental instruction-alignment validation transfers=6556
2026-09-23T10:19:44.139095Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:schedule_provisional_exchanges: retrying exchange schedule with incremental instruction-alignment validation transfers=44932
2026-09-23T10:19:44.572640Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:schedule_provisional_exchanges: finished incremental instruction-alignment validation transfers=40112 elapsed_ms=810 success=true
2026-09-23T10:19:45.184852Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:schedule_provisional_exchanges: retrying exchange schedule with incremental instruction-alignment validation transfers=68347
2026-09-23T10:19:45.367689Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:schedule_provisional_exchanges: finished incremental instruction-alignment validation transfers=6556 elapsed_ms=1406 success=true
2026-09-23T10:19:45.449937Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:schedule_provisional_exchanges: finished incremental instruction-alignment validation transfers=44932 elapsed_ms=1310 success=true
2026-09-23T10:19:45.623493Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:schedule_provisional_exchanges: prepared large physical exchange phase phase=4 transfers=40112 maximum_identical_destinations=1
2026-09-23T10:19:45.623543Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:schedule_provisional_exchanges: optimized physical exchange schedule phase=4 initial_horizon=5322 selected_horizon=5322 endpoint_lower_bound=4800 lower_bound_gap=522 selected_kind="full-duplex" neighborhood_improvements=0
2026-09-23T10:19:46.765645Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:schedule_provisional_exchanges: retrying exchange schedule with incremental instruction-alignment validation transfers=6556
2026-09-23T10:19:46.778615Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:schedule_provisional_exchanges: prepared large physical exchange phase phase=5 transfers=44932 maximum_identical_destinations=1
2026-09-23T10:19:46.778652Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:schedule_provisional_exchanges: optimized physical exchange schedule phase=5 initial_horizon=604 selected_horizon=604 endpoint_lower_bound=320 lower_bound_gap=284 selected_kind="full-duplex" neighborhood_improvements=0
2026-09-23T10:19:46.786441Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:schedule_provisional_exchanges: prepared large physical exchange phase phase=3 transfers=22094 maximum_identical_destinations=1
2026-09-23T10:19:46.786484Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:schedule_provisional_exchanges: optimized physical exchange schedule phase=3 initial_horizon=35328 selected_horizon=35328 endpoint_lower_bound=25344 lower_bound_gap=9984 selected_kind="full-duplex" neighborhood_improvements=0
2026-09-23T10:19:48.331419Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:schedule_provisional_exchanges: finished incremental instruction-alignment validation transfers=6556 elapsed_ms=1565 success=true
2026-09-23T10:19:48.367106Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:schedule_provisional_exchanges: finished incremental instruction-alignment validation transfers=68347 elapsed_ms=3182 success=true
2026-09-23T10:19:49.056945Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:schedule_provisional_exchanges: compared ordinary and paired exchange schedules phase=0 candidates=2180 ordinary_horizon=24054 paired_horizon=15339 use_paired=true
2026-09-23T10:19:49.140451Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:schedule_provisional_exchanges: prepared large physical exchange phase phase=0 transfers=6556 maximum_identical_destinations=1
2026-09-23T10:19:49.140482Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:schedule_provisional_exchanges: optimized physical exchange schedule phase=0 initial_horizon=15339 selected_horizon=15339 endpoint_lower_bound=10976 lower_bound_gap=4363 selected_kind="full-duplex" neighborhood_improvements=0
2026-09-23T10:19:50.630936Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:schedule_provisional_exchanges: prepared large physical exchange phase phase=2 transfers=68347 maximum_identical_destinations=1
2026-09-23T10:19:50.630993Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:schedule_provisional_exchanges: optimized physical exchange schedule phase=2 initial_horizon=3549 selected_horizon=3549 endpoint_lower_bound=1152 lower_bound_gap=2397 selected_kind="full-duplex" neighborhood_improvements=0
2026-09-23T10:19:50.710428Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:schedule_provisional_exchanges: close time.busy=7.58s time.idle=7.09µs
2026-09-23T10:19:50.754250Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:plan_kernels: close time.busy=43.7ms time.idle=3.07µs
2026-09-23T10:19:50.754854Z  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_r188" target="ipu21"
2026-09-23T10:19:50.813288Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:compile_kernels: kernel artifact created object=/srv/home/.cache/ipu-stack/kernels/65896e8b6dbb2af9e95d8b7237a3a2a092eac6fe990faa083ab3936bbf90798e/kernel.o metadata=/srv/home/.cache/ipu-stack/kernels/65896e8b6dbb2af9e95d8b7237a3a2a092eac6fe990faa083ab3936bbf90798e/kernel.json
2026-09-23T10:19:50.813551Z  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_r186" target="ipu21"
2026-09-23T10:19:50.870542Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:compile_kernels: kernel artifact created object=/srv/home/.cache/ipu-stack/kernels/4bccc40a0d38b0146671fcab59beafb1c51d356b01dc6a3f472f70dd86b56a56/kernel.o metadata=/srv/home/.cache/ipu-stack/kernels/4bccc40a0d38b0146671fcab59beafb1c51d356b01dc6a3f472f70dd86b56a56/kernel.json
2026-09-23T10:19:50.870847Z  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/unpack_transposed_amp_f16.S name="unpack_amp_to_row_major_f16_o1_r48_p48_c48_p48" target="ipu21"
2026-09-23T10:19:50.923527Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:compile_kernels: kernel artifact created object=/srv/home/.cache/ipu-stack/kernels/40f9b172ef7c4eb360c2cfabefa38ed54459c56bd13a695d1f2bb3150d07636d/kernel.o metadata=/srv/home/.cache/ipu-stack/kernels/40f9b172ef7c4eb360c2cfabefa38ed54459c56bd13a695d1f2bb3150d07636d/kernel.json
2026-09-23T10:19:50.923795Z  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/unpack_transposed_amp_f16.S name="unpack_amp_to_row_major_f16_o1_r32_p32_c48_p48" target="ipu21"
2026-09-23T10:19:50.977975Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:compile_kernels: kernel artifact created object=/srv/home/.cache/ipu-stack/kernels/a5724fe897cde1c38ef692ab876931a78c9f9ca727b2b4ff5879de11cbb376c1/kernel.o metadata=/srv/home/.cache/ipu-stack/kernels/a5724fe897cde1c38ef692ab876931a78c9f9ca727b2b4ff5879de11cbb376c1/kernel.json
2026-09-23T10:19:50.978276Z  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/unpack_transposed_amp_f16.S name="unpack_amp_to_row_major_f16_o1_r25_p32_c48_p48" target="ipu21"
2026-09-23T10:19:51.030406Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:compile_kernels: kernel artifact created object=/srv/home/.cache/ipu-stack/kernels/c670e6576165874c005781ffd8abc2975e7a9ea9568c22a3949958b923c1e565/kernel.o metadata=/srv/home/.cache/ipu-stack/kernels/c670e6576165874c005781ffd8abc2975e7a9ea9568c22a3949958b923c1e565/kernel.json
2026-09-23T10:19:51.030653Z  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/unpack_transposed_amp_f16.S name="unpack_amp_to_row_major_f16_o1_r48_p48_c46_p46" target="ipu21"
2026-09-23T10:19:51.088985Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:compile_kernels: kernel artifact created object=/srv/home/.cache/ipu-stack/kernels/52340e70a92c11000b1dfd5696f90571e56d37e48593fbd722a73dc1d582eafb/kernel.o metadata=/srv/home/.cache/ipu-stack/kernels/52340e70a92c11000b1dfd5696f90571e56d37e48593fbd722a73dc1d582eafb/kernel.json
2026-09-23T10:19:51.089195Z  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/unpack_transposed_amp_f16.S name="unpack_amp_to_row_major_f16_o1_r32_p32_c46_p46" target="ipu21"
2026-09-23T10:19:51.140731Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:compile_kernels: kernel artifact created object=/srv/home/.cache/ipu-stack/kernels/c3f87fd823db74d5ad446a145438b0598f1b1a81bf33d675810955945ba9ce90/kernel.o metadata=/srv/home/.cache/ipu-stack/kernels/c3f87fd823db74d5ad446a145438b0598f1b1a81bf33d675810955945ba9ce90/kernel.json
2026-09-23T10:19:51.141390Z  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/unpack_transposed_amp_f16.S name="unpack_amp_to_row_major_f16_o1_r25_p32_c46_p46" target="ipu21"
2026-09-23T10:19:51.191468Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:compile_kernels: kernel artifact created object=/srv/home/.cache/ipu-stack/kernels/61342e7d8a503c6aa995c6757fbf3d7773952b9cbd4f96cac6cba2cef35914f3/kernel.o metadata=/srv/home/.cache/ipu-stack/kernels/61342e7d8a503c6aa995c6757fbf3d7773952b9cbd4f96cac6cba2cef35914f3/kernel.json
2026-09-23T10:19:51.193120Z  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_packed0_k288_c32_r288" target="ipu21"
2026-09-23T10:19:51.245514Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:compile_kernels: kernel artifact created object=/srv/home/.cache/ipu-stack/kernels/4dcc0d56e6d7042f9800beba05c73f3633011c5143c439c22a5944560d768ebe/kernel.o metadata=/srv/home/.cache/ipu-stack/kernels/4dcc0d56e6d7042f9800beba05c73f3633011c5143c439c22a5944560d768ebe/kernel.json
2026-09-23T10:19:51.245897Z  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/unpack_transposed_amp_f16.S name="unpack_amp_to_row_major_f16_o1_r32_p32_c20_p20" target="ipu21"
2026-09-23T10:19:51.298007Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:compile_kernels: kernel artifact created object=/srv/home/.cache/ipu-stack/kernels/06a0b34d1ab92023bd1bbdd58b2143ff2f6e9dda211d15dfd12aa36b45536dc1/kernel.o metadata=/srv/home/.cache/ipu-stack/kernels/06a0b34d1ab92023bd1bbdd58b2143ff2f6e9dda211d15dfd12aa36b45536dc1/kernel.json
2026-09-23T10:19:51.298444Z  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/unpack_transposed_amp_f16.S name="unpack_amp_to_row_major_f16_o1_r25_p32_c20_p20" target="ipu21"
2026-09-23T10:19:51.348678Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:compile_kernels: kernel artifact created object=/srv/home/.cache/ipu-stack/kernels/5b61e4e4b6a51ec426d7f0667d97bb474a49b9931c1360c8f26ea212b6edb046/kernel.o metadata=/srv/home/.cache/ipu-stack/kernels/5b61e4e4b6a51ec426d7f0667d97bb474a49b9931c1360c8f26ea212b6edb046/kernel.json
2026-09-23T10:19:51.349391Z  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/unpack_transposed_amp_f16.S name="unpack_amp_to_row_major_f16_o1_r32_p32_c18_p18" target="ipu21"
2026-09-23T10:19:51.407049Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:compile_kernels: kernel artifact created object=/srv/home/.cache/ipu-stack/kernels/d040b3d94afa2dada5ac387c330612e4b0dca565d081ecb6807e3ae68fc0b41f/kernel.o metadata=/srv/home/.cache/ipu-stack/kernels/d040b3d94afa2dada5ac387c330612e4b0dca565d081ecb6807e3ae68fc0b41f/kernel.json
2026-09-23T10:19:51.407529Z  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/unpack_transposed_amp_f16.S name="unpack_amp_to_row_major_f16_o1_r25_p32_c18_p18" target="ipu21"
2026-09-23T10:19:51.458497Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:compile_kernels: kernel artifact created object=/srv/home/.cache/ipu-stack/kernels/6e8e68a8db0c395fe9b97423f00a769d967acacbcacd5886a7f8d6fbefb72911/kernel.o metadata=/srv/home/.cache/ipu-stack/kernels/6e8e68a8db0c395fe9b97423f00a769d967acacbcacd5886a7f8d6fbefb72911/kernel.json
2026-09-23T10:19:51.458747Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:compile_kernels: close time.busy=704ms time.idle=7.41µs
2026-09-23T10:19:51.459471Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:link_runtime: close time.busy=535µs time.idle=3.13µs
2026-09-23T10:19:51.517751Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}: exchange table and repeat patch storage exchange_table_bytes=5392 maximum_uncompressed_repeat_patch_bytes=0 maximum_elided_arithmetic_patch_bytes=0
2026-09-23T10:19:51.686337Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:size_tile_code: close time.busy=9.53ms time.idle=5.36µs
2026-09-23T10:19:51.686382Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}: placing generated tile programs linked_end=320496 host_code_base=376832 host_code_bytes=620 generated_code_bytes=2884 exchange_table_bytes=5392 executable_ranges=[(377520, 524288)]
2026-09-23T10:19:51.686430Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}: allocated package support memory linked_end=320496 profile_bytes=140 exchange_table_bytes=5392 host_code_bytes=620 host_data_bytes=404 generated_code_bytes=2884 code_address=377520 available_ranges=[(327680, 360448), (362612, 376832), (409600, 949168)]
2026-09-23T10:19:51.808090Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:place_storage: close time.busy=105ms time.idle=5.13µs
2026-09-23T10:19:51.966458Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:lower_exchanges: reused exchange optimization after validating relocated rows phase=1
2026-09-23T10:19:51.972783Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:lower_exchanges: prepared large physical exchange phase phase=1 transfers=16836 maximum_identical_destinations=1
2026-09-23T10:19:51.972809Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:lower_exchanges: optimized physical exchange schedule phase=1 initial_horizon=4712 selected_horizon=4712 endpoint_lower_bound=4320 lower_bound_gap=392 selected_kind="reused" neighborhood_improvements=0
2026-09-23T10:19:52.639713Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:lower_exchanges: reused exchange optimization after validating relocated rows phase=4
2026-09-23T10:19:52.665781Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:lower_exchanges: prepared large physical exchange phase phase=4 transfers=40112 maximum_identical_destinations=1
2026-09-23T10:19:52.665827Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:lower_exchanges: optimized physical exchange schedule phase=4 initial_horizon=5322 selected_horizon=5322 endpoint_lower_bound=4800 lower_bound_gap=522 selected_kind="reused" neighborhood_improvements=0
2026-09-23T10:19:52.689074Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:lower_exchanges: retrying exchange schedule with incremental instruction-alignment validation transfers=6556
2026-09-23T10:19:53.779096Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:lower_exchanges: retrying exchange schedule with incremental instruction-alignment validation transfers=44932
2026-09-23T10:19:53.925554Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:lower_exchanges: reused exchange optimization after validating relocated rows phase=2
2026-09-23T10:19:53.965558Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:lower_exchanges: prepared large physical exchange phase phase=2 transfers=68347 maximum_identical_destinations=1
2026-09-23T10:19:53.965591Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:lower_exchanges: optimized physical exchange schedule phase=2 initial_horizon=3549 selected_horizon=3549 endpoint_lower_bound=1152 lower_bound_gap=2397 selected_kind="reused" neighborhood_improvements=0
2026-09-23T10:19:54.221301Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:lower_exchanges: finished incremental instruction-alignment validation transfers=6556 elapsed_ms=1532 success=true
2026-09-23T10:19:55.131418Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:lower_exchanges: finished incremental instruction-alignment validation transfers=44932 elapsed_ms=1352 success=true
2026-09-23T10:19:55.572177Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:lower_exchanges: retrying exchange schedule with incremental instruction-alignment validation transfers=6556
2026-09-23T10:19:56.541251Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:lower_exchanges: prepared large physical exchange phase phase=5 transfers=44932 maximum_identical_destinations=1
2026-09-23T10:19:56.541301Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:lower_exchanges: optimized physical exchange schedule phase=5 initial_horizon=599 selected_horizon=599 endpoint_lower_bound=320 lower_bound_gap=279 selected_kind="full-duplex" neighborhood_improvements=0
2026-09-23T10:19:57.006413Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:lower_exchanges: finished incremental instruction-alignment validation transfers=6556 elapsed_ms=1434 success=true
2026-09-23T10:19:57.126761Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:lower_exchanges: retrying exchange schedule with incremental instruction-alignment validation transfers=22094
2026-09-23T10:19:57.665658Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:lower_exchanges: compared ordinary and paired exchange schedules phase=0 candidates=43 ordinary_horizon=17737 paired_horizon=17696 use_paired=true
2026-09-23T10:19:57.747508Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:lower_exchanges: prepared large physical exchange phase phase=0 transfers=6556 maximum_identical_destinations=1
2026-09-23T10:19:57.747573Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:lower_exchanges: optimized physical exchange schedule phase=0 initial_horizon=17696 selected_horizon=17696 endpoint_lower_bound=16992 lower_bound_gap=704 selected_kind="full-duplex" neighborhood_improvements=0
2026-09-23T10:20:00.733461Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:lower_exchanges: finished incremental instruction-alignment validation transfers=22094 elapsed_ms=3606 success=true
2026-09-23T10:20:02.297370Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:lower_exchanges: prepared large physical exchange phase phase=3 transfers=22094 maximum_identical_destinations=1
2026-09-23T10:20:02.297423Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:lower_exchanges: optimized physical exchange schedule phase=3 initial_horizon=29866 selected_horizon=29866 endpoint_lower_bound=25344 lower_bound_gap=4522 selected_kind="full-duplex" neighborhood_improvements=0
2026-09-23T10:20:02.359630Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:lower_exchanges: close time.busy=10.6s time.idle=6.56µs
2026-09-23T10:20:02.375992Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}: costed final placed program final_cycles=196026 final_exchange=65344
2026-09-23T10:20:02.501230Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:prepare_tile_code: close time.busy=14.4ms time.idle=5.71µs
2026-09-23T10:20:02.530905Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:emit_tile_code: close time.busy=29.6ms time.idle=2.78µs
2026-09-23T10:20:02.719034Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}:build_tile_images: close time.busy=169ms time.idle=3.12µs
2026-09-23T10:20:02.963161Z  INFO ipu_codegen.package.build{tile_count=1472 operations=3}: close time.busy=112s time.idle=5.01µs
2026-09-23T10:20:02.968697Z  INFO writing application package tiles=1472
2026-09-23T10:20:03.190252Z  INFO application package written
2026-09-23T10:20:03.193522Z  INFO reading application package
2026-09-23T10:20:03.382645Z  INFO application package read tiles=1472 compiler=0.1.0
2026-09-23T10:20:03.388268Z  INFO waiting for hardware ownership path=/tmp/ipu-stack-device.lock
2026-09-23T10:20:03.388287Z  INFO acquired hardware ownership path=/tmp/ipu-stack-device.lock
2026-09-23T10:20:03.388293Z  INFO opening IPU device path="/dev/ipu0"
2026-09-23T10:20:03.388339Z  INFO IPU device attached path=/dev/ipu0
2026-09-23T10:20:03.388343Z  INFO initializing IPU device
2026-09-23T10:20:03.396217Z  INFO IPU device initialized
2026-09-23T10:20:03.396223Z  INFO replaying device configuration register_writes=2003
2026-09-23T10:20:03.402838Z  INFO loading application tile_count=1472 final_mark=23
2026-09-23T10:20:03.625935Z  INFO application loaded and initial synchronization completed tile_count=1472
2026-09-23T10:20:04.472258Z  INFO prepared reference inputs elapsed_ms=842
2026-09-23T10:20:04.807819Z  INFO configured host reference BLAS threads=4
referencePrecision=device
2026-09-23T10:20:05.390557Z  INFO evaluated host reference elapsed_ms=918
2026-09-23T10:20:05.393639Z  INFO starting host exchange session startup_mark=23
2026-09-23T10:20:05.393668Z  INFO attaching host exchange pages pages=2
2026-09-23T10:20:05.394385Z  INFO host exchange session started
2026-09-23T10:20:05.394427Z  INFO invoking host exchange call call="initialize" command=0 phases=4
2026-09-23T10:20:05.398291Z  INFO invoking host exchange call call="run" command=0 phases=13
referenceInference=1 numericalTest=PASS
profile=artifacts/gemm-solver-20260923/f8-b1.capnp tiles=1472 samples=42625
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=189756 minimumTileCycles=186000 deviceMicroseconds=126.504 effectiveGemmTflops=114.290 peakTflops=282.624 efficiencyPercent=40.44 maximumAbsoluteError=0.005493
package=artifacts/gemm-solver-20260923/f8-b1.ipuexe tiles=1472 hardwareTest=PASS