einsum-opt_flops on zen5-cpu

What was measured

Hardware

Every timed cell matches a successful verification census, including operation variants. Each cell is median of process medians ± sample SD of those medians; W is the median within-process sample SD. All values are milliseconds. These are descriptive spreads, not confidence intervals. The between-process spread includes the rotating-order schedule. Each process retains 5 chronological calls; 2 complete process sets are recorded. Spread guide (computed from this run's own rows): per-engine between-start range median 0.45%, worst 4.34% among cells of 1 us or more (a sub-microsecond row is timer and scheduler jitter, not kernel time); 0 of 26 engine cells exceed 5% and carry !; median within-process relative SD 0.37%. A cell whose ±, ! or p is as large as the difference you are reading is not evidence of a regression. no row, so no ratio there is position-balanced). ! marks a cell where some engine's own per-start median range exceeds 5% of its median, and p a balanced cell whose estimated position-effect ratio exceeds 1.05. Flagged cells are retained, not discarded; the between-start variation can exceed the difference being compared. ! between-start range over 5%, p position effect over 5%, dispatch-dominated means the row is under 20 us so 4T/8T pay thread-region start-up). Per-start ratios and every chronological sample are in the run directory named at the top of this page (run*.csv). These describe the recorded revisions; they are not a regression claim.

cpueinsum, float64, 1T (CPU 4)

| case | opt_flops (ms) | notes | |---|---| | bin_batched_matmul_b32_m64_n64_k64/n1/b1 | 0.558895 ± 0.000814587 (W 0.00239534) | | | bin_elementwise_mul_2048x2048/n1/b1 | 4.02009 ± 0.0383966 (W 0.0134363) | | | bin_matmul_256/n1/b1 | 1.07024 ± 0.00289065 (W 0.00275378) | | | bin_outer_product_4096/n1/b1 | 6.19889 ± 0.00403122 (W 0.022405) | | | gm_queen5_5_3.wcsp/n1/b1 | 4568 ± 52.9609 (W 22.5582) | | | lm_batch_likelihood_brackets_4_4d/n1/b1 | 26.9812 ± 0.176017 (W 0.330914) | | | lm_batch_likelihood_sentence_3_12d/n1/b1 | 72.4797 ± 0.340885 (W 0.138893) | | | lm_batch_likelihood_sentence_4_4d/n1/b1 | 27.3383 ± 0.156766 (W 0.173524) | | | str_matrix_chain_multiplication_100/n1/b1 | 12.657 ± 0.0196279 (W 0.0312983) | | | str_nw_mera_closed_120/n1/b1 | 1766.27 ± 0.155851 (W 1.28391) | | | str_nw_mera_open_26/n1/b1 | 1165.41 ± 0.0126049 (W 1.18565) | | | tensornetwork_permutation_focus_step409_316/n1/b1 | 233.829 ± 0.575277 (W 0.450488) | | | tensornetwork_permutation_light_415/n1/b1 | 231.935 ± 0.0904404 (W 0.290097) | |

cpueinsum, float64, 4T (CPU 4-7)

| case | opt_flops (ms) | notes | |---|---| | bin_batched_matmul_b32_m64_n64_k64/n1/b1 | 0.144242 ± 0.000306177 (W 0.00114546) | | | bin_elementwise_mul_2048x2048/n1/b1 | 2.14403 ± 0.0149779 (W 0.0140393) | | | bin_matmul_256/n1/b1 | 1.07535 ± 0.00814941 (W 0.00206279) | | | bin_outer_product_4096/n1/b1 | 6.15475 ± 0.188765 (W 0.0170268) | | | gm_queen5_5_3.wcsp/n1/b1 | 1839.03 ± 4.04695 (W 7.354) | | | lm_batch_likelihood_brackets_4_4d/n1/b1 | 11.5854 ± 0.0838777 (W 0.0379017) | | | lm_batch_likelihood_sentence_3_12d/n1/b1 | 22.6687 ± 0.154074 (W 0.800329) | | | lm_batch_likelihood_sentence_4_4d/n1/b1 | 12.4302 ± 0.0397203 (W 0.12276) | | | str_matrix_chain_multiplication_100/n1/b1 | 12.6621 ± 0.00399374 (W 0.0221998) | | | str_nw_mera_closed_120/n1/b1 | 512.763 ± 6.95692 (W 5.87867) | | | str_nw_mera_open_26/n1/b1 | 327.348 ± 1.5796 (W 4.44016) | | | tensornetwork_permutation_focus_step409_316/n1/b1 | 87.7799 ± 0.0312548 (W 0.880862) | | | tensornetwork_permutation_light_415/n1/b1 | 80.3255 ± 0.254893 (W 4.46232) | |

Repeated process sets describe observed repeatability under the recorded order schedule; they are not a paired baseline/candidate experiment and do not establish equal performance between engines.