tiles=1472 matched=308035/308035 spanCycles=1011180 spanMs=0.674120 name="sync" phases=73 tiles=1457 samples=27405 timelineCycles=699090 workCycles=165571068 range=0..1010454 p95=19332 max=49686 useful=n/a coverage=0.0% name="exchange" phases=91 tiles=1472 samples=106547 timelineCycles=611886 workCycles=664297164 range=0..1011030 p95=17208 max=48276 useful=n/a coverage=0.0% name="copy_u64" phases=289 tiles=1458 samples=24757 timelineCycles=158190 workCycles=22298826 range=44868..1007304 p95=1578 max=13146 useful=n/a coverage=0.0% name="cast_f16_f8" phases=251 tiles=1458 samples=14184 timelineCycles=84396 workCycles=81134334 range=18276..997758 p95=11442 max=11634 useful=22.7% coverage=100.0% name="reduce_sum_f16" phases=200 tiles=1458 samples=8968 timelineCycles=74088 workCycles=39129168 range=45162..1007034 p95=7212 max=7212 useful=56.5% coverage=100.0% name="copy_strided_u64" phases=145 tiles=1117 samples=2360 timelineCycles=57606 workCycles=988668 range=0..1004040 p95=618 max=5430 useful=n/a coverage=0.0% name="gemm_f16_init_large_rows_interleaved_k64_c80_r7_r8" phases=75 tiles=1360 samples=16320 timelineCycles=51648 workCycles=68935680 range=211812..441606 p95=4224 max=4224 useful=12.9% coverage=100.0% name="gemm_f16_init_large_rows_interleaved_k80_c64_r7_r8" phases=75 tiles=1360 samples=16320 timelineCycles=51354 workCycles=68544000 range=204438..434484 p95=4200 max=4200 useful=13.0% coverage=100.0% name="gemm_f16_init_small_rows_interleaved_k64_c80_r7_r8" phases=70 tiles=112 samples=1344 timelineCycles=50898 workCycles=5644800 range=211800..441540 p95=4200 max=4200 useful=11.4% coverage=100.0% name="gemm_f16_init_small_rows_interleaved_k80_c64_r7_r8" phases=70 tiles=112 samples=1344 timelineCycles=50592 workCycles=5612544 range=204450..434442 p95=4176 max=4176 useful=11.5% coverage=100.0% name="fill_zero_u64" phases=140 tiles=1472 samples=5074 timelineCycles=47796 workCycles=1959738 range=34464..998376 p95=618 max=3150 useful=0.0% coverage=32.4% name="attention_merge_v72_p80_k64_out32_f16" phases=80 tiles=1472 samples=16192 timelineCycles=37878 workCycles=54042960 range=216000..424728 p95=3426 max=3426 useful=n/a coverage=0.0% name="attention_softmax_d72_p64_full_f16" phases=80 tiles=1472 samples=16192 timelineCycles=35850 workCycles=51393408 range=208626..417078 p95=3174 max=3174 useful=30.2% coverage=100.0% name="layer_norm_f16" phases=69 tiles=729 samples=2187 timelineCycles=30396 workCycles=20509686 range=66102..770754 p95=9378 max=9378 useful=35.3% coverage=100.0% name="add_f16" phases=276 tiles=1472 samples=15364 timelineCycles=30168 workCycles=20306160 range=53844..1010118 p95=2652 max=2652 useful=15.2% coverage=100.0% name="gemm_f8_init_large_rows_interleaved_k192_c160_r80_r82" phases=28 tiles=780 samples=780 timelineCycles=29856 workCycles=23180040 range=570312..600168 p95=29718 max=29718 useful=66.2% coverage=100.0% name="gemm_f8_init_small_rows_interleaved_k192_c160_r80_r82" phases=18 tiles=624 samples=624 timelineCycles=29274 workCycles=18225792 range=570378..599652 p95=29208 max=29208 useful=65.5% coverage=100.0% name="gemm_f8_init_large_rows_interleaved_k192_c144_r80_r82" phases=13 tiles=30 samples=30 timelineCycles=26898 workCycles=803340 range=569550..596448 p95=26778 max=26778 useful=66.1% coverage=100.0% name="gemm_f8_init_large_rows_interleaved_k160_c128_r120_r122" phases=33 tiles=1215 samples=1215 timelineCycles=26478 workCycles=31995810 range=691098..717576 p95=26334 max=26334 useful=73.9% coverage=100.0% name="gemm_f8_init_small_rows_interleaved_k192_c144_r80_r82" phases=11 tiles=24 samples=24 timelineCycles=26376 workCycles=631728 range=569610..595986 p95=26322 max=26322 useful=65.4% coverage=100.0%