{"builds":{"c":[{"binary_sha256":"9389a14426ccea80a393d00217db2d19397fbc11b43ace515f4eae5c32e06216","compiler":"gcc","openblas_compiled":true,"telemetry_instrumented":false,"version":"gcc (Debian 15.2.0-17) 15.2.0"},{"binary_sha256":"59e11997def34b3540996efc8e07ceca8b8df2a751040fd438f88ec145949d0f","compiler":"clang","openblas_compiled":true,"telemetry_instrumented":false,"version":"Debian clang version 21.1.8 (10)"}],"onednn":{"binary_sha256":"345db95a29d431426946b50f8690a0681eb281814eaf8aa5aabc9056f8c8f9ce","compiler_version":"g++ (Debian 15.2.0-17) 15.2.0","effective_cpu_isa":"cpu_isa_avx","implementation":"oneDNN 3.12.1 CPU matmul","openmp_control_build_enabled":true,"threading_runtime":"openmp"}},"configuration":{"backend":"auto","cache_scrub_mib":32,"cold_samples":5,"compilers":["gcc","clang"],"extent":384,"thread_counts":[1,4],"tile":32,"timeout_seconds":900,"warm_samples":15,"workspace_mib":64},"coverage":{"all_c_reports_exact":true,"all_engine_measurements_exact":true,"all_onednn_reports_exact":true,"c_engine_measurement_count":56,"c_execution_count":4,"external_oracle_observation_count":14,"onednn_1_to_4_scaling_count":2,"onednn_engine_measurement_count":4,"onednn_execution_count":2,"openblas_1_to_4_scaling_count":4,"published_direct_comparison_count":28,"published_direct_ratio_value_count":56,"published_direct_ratio_values_below_one_count":40,"qqfenx_1_to_4_scaling_count":12},"direct_comparisons":[{"baseline":"dense-c-blocked","baseline_over_candidate_end_to_end":0.457043,"baseline_over_candidate_prepared":0.475175,"candidate":"qqfenx-c11","case":"synthetic-q32","compiler":"gcc","interpretation":"greater than 1 means QQfenx was faster in this retained host run","matched_configured_lanes":true,"ratio_published":true,"same_observed_output":true,"source":"c-run","threads":1},{"baseline":"dense-c-blocked","baseline_over_candidate_end_to_end":1.197603,"baseline_over_candidate_prepared":1.265249,"candidate":"qqfenx-c11","case":"synthetic-q16","compiler":"gcc","interpretation":"greater than 1 means QQfenx was faster in this retained host run","matched_configured_lanes":true,"ratio_published":true,"same_observed_output":true,"source":"c-run","threads":1},{"baseline":"dense-c-blocked","baseline_over_candidate_end_to_end":1.251559,"baseline_over_candidate_prepared":1.322559,"candidate":"qqfenx-c11","case":"synthetic-q8-safe-u6","compiler":"gcc","interpretation":"greater than 1 means QQfenx was faster in this retained host run","matched_configured_lanes":true,"ratio_published":true,"same_observed_output":true,"source":"c-run","threads":1},{"baseline":"openblas-sgemm","baseline_over_candidate_end_to_end":0.531287,"baseline_over_candidate_prepared":0.497064,"candidate":"qqfenx-c11","case":"synthetic-q8-safe-u6","compiler":"gcc","interpretation":"greater than 1 means QQfenx was faster in this retained host run","matched_configured_lanes":true,"ratio_published":true,"same_observed_output":true,"source":"c-run","threads":1},{"baseline":"dense-c-blocked","baseline_over_candidate_end_to_end":8.937884,"baseline_over_candidate_prepared":13.644358,"candidate":"qqfenx-c11","case":"synthetic-zero-ideal","compiler":"gcc","interpretation":"greater than 1 means QQfenx was faster in this retained host run","matched_configured_lanes":true,"ratio_published":true,"same_observed_output":true,"source":"c-run","threads":1},{"baseline":"dense-c-blocked","baseline_over_candidate_end_to_end":2.526254,"baseline_over_candidate_prepared":2.541017,"candidate":"qqfenx-c11","case":"synthetic-cancellation-cascade","compiler":"gcc","interpretation":"greater than 1 means QQfenx was faster in this retained host run","matched_configured_lanes":true,"ratio_published":true,"same_observed_output":true,"source":"c-run","threads":1},{"baseline":"dense-c-blocked","baseline_over_candidate_end_to_end":0.044432,"baseline_over_candidate_prepared":0.033792,"candidate":"qqfenx-c11","case":"uci-iris-feature-gram-i32-v1","compiler":"gcc","interpretation":"greater than 1 means QQfenx was faster in this retained host run","matched_configured_lanes":true,"ratio_published":true,"same_observed_output":true,"source":"c-run","threads":1},{"baseline":"openblas-sgemm","baseline_over_candidate_end_to_end":0.043483,"baseline_over_candidate_prepared":0.029326,"candidate":"qqfenx-c11","case":"uci-iris-feature-gram-i32-v1","compiler":"gcc","interpretation":"greater than 1 means QQfenx was faster in this retained host run","matched_configured_lanes":true,"ratio_published":true,"same_observed_output":true,"source":"c-run","threads":1},{"baseline":"openblas-sgemm","baseline_over_candidate_end_to_end":0.958663,"baseline_over_candidate_prepared":0.477651,"candidate":"qqfenx-c11","case":"synthetic-q8-safe-u6","compiler":"gcc","interpretation":"greater than 1 means QQfenx was faster in this retained host run","matched_configured_lanes":true,"ratio_published":true,"same_observed_output":true,"source":"c-run","threads":4},{"baseline":"openblas-sgemm","baseline_over_candidate_end_to_end":0.015689,"baseline_over_candidate_prepared":0.008929,"candidate":"qqfenx-c11","case":"uci-iris-feature-gram-i32-v1","compiler":"gcc","interpretation":"greater than 1 means QQfenx was faster in this retained host run","matched_configured_lanes":true,"ratio_published":true,"same_observed_output":true,"source":"c-run","threads":4},{"baseline":"dense-c-blocked","baseline_over_candidate_end_to_end":0.624562,"baseline_over_candidate_prepared":0.549084,"candidate":"qqfenx-c11","case":"synthetic-q32","compiler":"clang","interpretation":"greater than 1 means QQfenx was faster in this retained host run","matched_configured_lanes":true,"ratio_published":true,"same_observed_output":true,"source":"c-run","threads":1},{"baseline":"dense-c-blocked","baseline_over_candidate_end_to_end":1.243312,"baseline_over_candidate_prepared":1.237048,"candidate":"qqfenx-c11","case":"synthetic-q16","compiler":"clang","interpretation":"greater than 1 means QQfenx was faster in this retained host run","matched_configured_lanes":true,"ratio_published":true,"same_observed_output":true,"source":"c-run","threads":1},{"baseline":"dense-c-blocked","baseline_over_candidate_end_to_end":1.274154,"baseline_over_candidate_prepared":1.258486,"candidate":"qqfenx-c11","case":"synthetic-q8-safe-u6","compiler":"clang","interpretation":"greater than 1 means QQfenx was faster in this retained host run","matched_configured_lanes":true,"ratio_published":true,"same_observed_output":true,"source":"c-run","threads":1},{"baseline":"openblas-sgemm","baseline_over_candidate_end_to_end":0.470314,"baseline_over_candidate_prepared":0.393089,"candidate":"qqfenx-c11","case":"synthetic-q8-safe-u6","compiler":"clang","interpretation":"greater than 1 means QQfenx was faster in this retained host run","matched_configured_lanes":true,"ratio_published":true,"same_observed_output":true,"source":"c-run","threads":1},{"baseline":"dense-c-blocked","baseline_over_candidate_end_to_end":8.016763,"baseline_over_candidate_prepared":12.600807,"candidate":"qqfenx-c11","case":"synthetic-zero-ideal","compiler":"clang","interpretation":"greater than 1 means QQfenx was faster in this retained host run","matched_configured_lanes":true,"ratio_published":true,"same_observed_output":true,"source":"c-run","threads":1},{"baseline":"dense-c-blocked","baseline_over_candidate_end_to_end":2.4205,"baseline_over_candidate_prepared":2.416981,"candidate":"qqfenx-c11","case":"synthetic-cancellation-cascade","compiler":"clang","interpretation":"greater than 1 means QQfenx was faster in this retained host run","matched_configured_lanes":true,"ratio_published":true,"same_observed_output":true,"source":"c-run","threads":1},{"baseline":"dense-c-blocked","baseline_over_candidate_end_to_end":0.056362,"baseline_over_candidate_prepared":0.102033,"candidate":"qqfenx-c11","case":"uci-iris-feature-gram-i32-v1","compiler":"clang","interpretation":"greater than 1 means QQfenx was faster in this retained host run","matched_configured_lanes":true,"ratio_published":true,"same_observed_output":true,"source":"c-run","threads":1},{"baseline":"openblas-sgemm","baseline_over_candidate_end_to_end":0.045815,"baseline_over_candidate_prepared":0.042477,"candidate":"qqfenx-c11","case":"uci-iris-feature-gram-i32-v1","compiler":"clang","interpretation":"greater than 1 means QQfenx was faster in this retained host run","matched_configured_lanes":true,"ratio_published":true,"same_observed_output":true,"source":"c-run","threads":1},{"baseline":"openblas-sgemm","baseline_over_candidate_end_to_end":0.711742,"baseline_over_candidate_prepared":0.500387,"candidate":"qqfenx-c11","case":"synthetic-q8-safe-u6","compiler":"clang","interpretation":"greater than 1 means QQfenx was faster in this retained host run","matched_configured_lanes":true,"ratio_published":true,"same_observed_output":true,"source":"c-run","threads":4},{"baseline":"openblas-sgemm","baseline_over_candidate_end_to_end":0.017262,"baseline_over_candidate_prepared":0.0075,"candidate":"qqfenx-c11","case":"uci-iris-feature-gram-i32-v1","compiler":"clang","interpretation":"greater than 1 means QQfenx was faster in this retained host run","matched_configured_lanes":true,"ratio_published":true,"same_observed_output":true,"source":"c-run","threads":4},{"baseline":"onednn-matmul-u8s8s32","baseline_over_candidate_end_to_end":0.661541,"baseline_over_candidate_prepared":0.681771,"baseline_threading_runtime":"openmp","candidate":"qqfenx-c11","case":"synthetic-q8-safe-u6","compiler":"gcc","interpretation":"greater than 1 means QQfenx was faster in this retained host run","ratio_published":true,"requested_thread_counts_equal":true,"same_observed_output":true,"source":"onednn-comparison","thread_control_verified":true,"threads":1},{"baseline":"onednn-matmul-u8s8s32","baseline_over_candidate_end_to_end":0.31536,"baseline_over_candidate_prepared":0.107271,"baseline_threading_runtime":"openmp","candidate":"qqfenx-c11","case":"uci-iris-feature-gram-i32-v1","compiler":"gcc","interpretation":"greater than 1 means QQfenx was faster in this retained host run","ratio_published":true,"requested_thread_counts_equal":true,"same_observed_output":true,"source":"onednn-comparison","thread_control_verified":true,"threads":1},{"baseline":"onednn-matmul-u8s8s32","baseline_over_candidate_end_to_end":0.527812,"baseline_over_candidate_prepared":0.474527,"baseline_threading_runtime":"openmp","candidate":"qqfenx-c11","case":"synthetic-q8-safe-u6","compiler":"gcc","interpretation":"greater than 1 means QQfenx was faster in this retained host run","ratio_published":true,"requested_thread_counts_equal":true,"same_observed_output":true,"source":"onednn-comparison","thread_control_verified":true,"threads":4},{"baseline":"onednn-matmul-u8s8s32","baseline_over_candidate_end_to_end":0.161096,"baseline_over_candidate_prepared":0.077317,"baseline_threading_runtime":"openmp","candidate":"qqfenx-c11","case":"uci-iris-feature-gram-i32-v1","compiler":"gcc","interpretation":"greater than 1 means QQfenx was faster in this retained host run","ratio_published":true,"requested_thread_counts_equal":true,"same_observed_output":true,"source":"onednn-comparison","thread_control_verified":true,"threads":4},{"baseline":"onednn-matmul-u8s8s32","baseline_over_candidate_end_to_end":0.609871,"baseline_over_candidate_prepared":0.583699,"baseline_threading_runtime":"openmp","candidate":"qqfenx-c11","case":"synthetic-q8-safe-u6","compiler":"clang","interpretation":"greater than 1 means QQfenx was faster in this retained host run","ratio_published":true,"requested_thread_counts_equal":true,"same_observed_output":true,"source":"onednn-comparison","thread_control_verified":true,"threads":1},{"baseline":"onednn-matmul-u8s8s32","baseline_over_candidate_end_to_end":0.283637,"baseline_over_candidate_prepared":0.161424,"baseline_threading_runtime":"openmp","candidate":"qqfenx-c11","case":"uci-iris-feature-gram-i32-v1","compiler":"clang","interpretation":"greater than 1 means QQfenx was faster in this retained host run","ratio_published":true,"requested_thread_counts_equal":true,"same_observed_output":true,"source":"onednn-comparison","thread_control_verified":true,"threads":1},{"baseline":"onednn-matmul-u8s8s32","baseline_over_candidate_end_to_end":0.517959,"baseline_over_candidate_prepared":0.540049,"baseline_threading_runtime":"openmp","candidate":"qqfenx-c11","case":"synthetic-q8-safe-u6","compiler":"clang","interpretation":"greater than 1 means QQfenx was faster in this retained host run","ratio_published":true,"requested_thread_counts_equal":true,"same_observed_output":true,"source":"onednn-comparison","thread_control_verified":true,"threads":4},{"baseline":"onednn-matmul-u8s8s32","baseline_over_candidate_end_to_end":0.1695,"baseline_over_candidate_prepared":0.055735,"baseline_threading_runtime":"openmp","candidate":"qqfenx-c11","case":"uci-iris-feature-gram-i32-v1","compiler":"clang","interpretation":"greater than 1 means QQfenx was faster in this retained host run","ratio_published":true,"requested_thread_counts_equal":true,"same_observed_output":true,"source":"onednn-comparison","thread_control_verified":true,"threads":4}],"external_workload":{"all_observations_exact":true,"dataset_id":53,"doi":"10.24432/C56C76","expected_output_u32":[522385,267343,348376,112814,267343,143040,167430,53189,348376,167430,258271,86911,112814,53189,86911,30233],"left_sha256":"ba39b677d07ee0258c56d9dd0df40d1b42927255450399ec9b76c0858f4fc2ba","observations":[{"compiler":"gcc","engine":"qqfenx-c11","exact_match":true,"output_fnv1a64":"21255eaa70abb088","retained_output_u32":[522385,267343,348376,112814,267343,143040,167430,53189,348376,167430,258271,86911,112814,53189,86911,30233],"source":"c-run","threads":1},{"compiler":"gcc","engine":"dense-c-blocked","exact_match":true,"output_fnv1a64":"21255eaa70abb088","retained_output_u32":[522385,267343,348376,112814,267343,143040,167430,53189,348376,167430,258271,86911,112814,53189,86911,30233],"source":"c-run","threads":1},{"compiler":"gcc","engine":"openblas-sgemm","exact_match":true,"output_fnv1a64":"21255eaa70abb088","retained_output_u32":[522385,267343,348376,112814,267343,143040,167430,53189,348376,167430,258271,86911,112814,53189,86911,30233],"source":"c-run","threads":1},{"compiler":"gcc","engine":"qqfenx-c11","exact_match":true,"output_fnv1a64":"21255eaa70abb088","retained_output_u32":[522385,267343,348376,112814,267343,143040,167430,53189,348376,167430,258271,86911,112814,53189,86911,30233],"source":"c-run","threads":4},{"compiler":"gcc","engine":"dense-c-blocked","exact_match":true,"output_fnv1a64":"21255eaa70abb088","retained_output_u32":[522385,267343,348376,112814,267343,143040,167430,53189,348376,167430,258271,86911,112814,53189,86911,30233],"source":"c-run","threads":4},{"compiler":"gcc","engine":"openblas-sgemm","exact_match":true,"output_fnv1a64":"21255eaa70abb088","retained_output_u32":[522385,267343,348376,112814,267343,143040,167430,53189,348376,167430,258271,86911,112814,53189,86911,30233],"source":"c-run","threads":4},{"compiler":"clang","engine":"qqfenx-c11","exact_match":true,"output_fnv1a64":"21255eaa70abb088","retained_output_u32":[522385,267343,348376,112814,267343,143040,167430,53189,348376,167430,258271,86911,112814,53189,86911,30233],"source":"c-run","threads":1},{"compiler":"clang","engine":"dense-c-blocked","exact_match":true,"output_fnv1a64":"21255eaa70abb088","retained_output_u32":[522385,267343,348376,112814,267343,143040,167430,53189,348376,167430,258271,86911,112814,53189,86911,30233],"source":"c-run","threads":1},{"compiler":"clang","engine":"openblas-sgemm","exact_match":true,"output_fnv1a64":"21255eaa70abb088","retained_output_u32":[522385,267343,348376,112814,267343,143040,167430,53189,348376,167430,258271,86911,112814,53189,86911,30233],"source":"c-run","threads":1},{"compiler":"clang","engine":"qqfenx-c11","exact_match":true,"output_fnv1a64":"21255eaa70abb088","retained_output_u32":[522385,267343,348376,112814,267343,143040,167430,53189,348376,167430,258271,86911,112814,53189,86911,30233],"source":"c-run","threads":4},{"compiler":"clang","engine":"dense-c-blocked","exact_match":true,"output_fnv1a64":"21255eaa70abb088","retained_output_u32":[522385,267343,348376,112814,267343,143040,167430,53189,348376,167430,258271,86911,112814,53189,86911,30233],"source":"c-run","threads":4},{"compiler":"clang","engine":"openblas-sgemm","exact_match":true,"output_fnv1a64":"21255eaa70abb088","retained_output_u32":[522385,267343,348376,112814,267343,143040,167430,53189,348376,167430,258271,86911,112814,53189,86911,30233],"source":"c-run","threads":4},{"compiler":"oneDNN","engine":"onednn-matmul-u8s8s32","exact_match":true,"output_fnv1a64":"21255eaa70abb088","retained_output_u32":[522385,267343,348376,112814,267343,143040,167430,53189,348376,167430,258271,86911,112814,53189,86911,30233],"source":"onednn-run","threads":1},{"compiler":"oneDNN","engine":"onednn-matmul-u8s8s32","exact_match":true,"output_fnv1a64":"21255eaa70abb088","retained_output_u32":[522385,267343,348376,112814,267343,143040,167430,53189,348376,167430,258271,86911,112814,53189,86911,30233],"source":"onednn-run","threads":4}],"operation":"transpose(X10) * X10","oracle_sha256":"90f359758f26cc028c4ea52ed99e30c629512b3efb6e53dfd18bb6b669c36f35","publisher":"UCI Machine Learning Repository","right_sha256":"658164630a3c710c59a6cddaa35b2e011c6b77746851ecb6369f1c2f80c3b417","shape":{"k":150,"m":4,"n":4},"source_member":"bezdekIris.data","source_sha256":"0fed2a99db77ec533a62dc66894d3ec6df3b58b6a8f3cf4a6b47e4086b7f97dc","workload_id":"uci-iris-feature-gram-i32-v1"},"host":{"architecture":"x86_64","cores_per_socket":4,"cpu_max_mhz":"3700.0000","cpu_model":"Intel(R) Core(TM) i5-2500S CPU @ 2.70GHz","kernel":{"machine":"x86_64","release":"7.0.12+kali-amd64","system":"Linux"},"l1d_cache":"128 KiB (4 instances)","l1i_cache":"128 KiB (4 instances)","l2_cache":"1 MiB (4 instances)","l3_cache":"6 MiB (1 instance)","logical_cpus":4,"measurement_window":{"end":{"captured_at_utc":"2026-08-26T08:26:43.301098+00:00","load_average":{"fifteen_minutes":4.93896484375,"five_minutes":4.05859375,"one_minute":2.77197265625},"logical_cpu_count":4},"start":{"captured_at_utc":"2026-08-26T08:26:17.754391+00:00","load_average":{"fifteen_minutes":4.95361328125,"five_minutes":4.02783203125,"one_minute":1.912109375},"logical_cpu_count":4}},"memory_total_kib":8110868,"numa_nodes":1,"os":{"id":"kali","pretty_name":"Kali GNU/Linux Rolling","version_id":"2026.2"},"process_affinity":[0,1,2,3],"sockets":1,"threads_per_core":1},"license":"LicenseRef-MFENX-Commercial","measurements":{"c_engines":[{"active_lanes":1,"active_lanes_observed":true,"case":"synthetic-q32","compiler":"gcc","configured_lanes":1,"end_to_end_scope":"dense inputs; canonical import, public GEMM execution, dense export, and input/intermediate cleanup; final dense output cleanup excluded","engine":"qqfenx-c11","exact_match":true,"general_wrapping_u32":true,"implementation":"SSE4.1","medians_ns":{"end_to_end_cold":19993369,"end_to_end_warm":19406053,"prepared_cold":20427243,"prepared_warm":19357487},"output_fnv1a64":"8697e608b10aed1e","reserved_workspace_bytes":4312,"sample_counts":{"end_to_end_cold":5,"end_to_end_warm":15,"prepared_cold":5,"prepared_warm":15},"semantic_scope":"complete wrapping-u32 modulo 2^32","source":"c-run","status":"measured","threads":1,"timing_scope":"prepared canonical inputs; public GEMM output allocation, execution, and cascade intermediate cleanup; dense export and final cleanup excluded","work":{"annihilated_macs":0,"executed_macs":56623104,"logical_macs":56623104,"promoted_tiles":0,"promotion_bits":0,"q16_macs":0,"q24_macs":0,"q32_macs":56623104,"q8_macs":0}},{"active_lanes":1,"active_lanes_observed":true,"case":"synthetic-q32","compiler":"gcc","configured_lanes":1,"end_to_end_scope":"caller-dense inputs; input copies, output allocation, zeroing, execution, and input/intermediate cleanup; final output cleanup excluded","engine":"dense-c-blocked","exact_match":true,"general_wrapping_u32":true,"implementation":"compiler-optimized row-major 32x32 blocked C11","medians_ns":{"end_to_end_cold":9483463,"end_to_end_warm":8869407,"prepared_cold":9160413,"prepared_warm":9198201},"output_fnv1a64":"8697e608b10aed1e","reserved_workspace_bytes":0,"sample_counts":{"end_to_end_cold":5,"end_to_end_warm":15,"prepared_cold":5,"prepared_warm":15},"semantic_scope":"complete wrapping-u32 modulo 2^32","source":"c-run","status":"measured","threads":1,"timing_scope":"prepared dense inputs; output allocation, zeroing, execution, and cascade intermediate cleanup; final output cleanup excluded","work":{"annihilated_macs":0,"executed_macs":56623104,"logical_macs":56623104,"promoted_tiles":0,"promotion_bits":0,"q16_macs":0,"q24_macs":0,"q32_macs":0,"q8_macs":0}},{"active_lanes":1,"active_lanes_observed":true,"case":"synthetic-q16","compiler":"gcc","configured_lanes":1,"end_to_end_scope":"dense inputs; canonical import, public GEMM execution, dense export, and input/intermediate cleanup; final dense output cleanup excluded","engine":"qqfenx-c11","exact_match":true,"general_wrapping_u32":true,"implementation":"SSE4.1","medians_ns":{"end_to_end_cold":9606405,"end_to_end_warm":8348733,"prepared_cold":7378896,"prepared_warm":7376446},"output_fnv1a64":"3c3c99e972219fea","reserved_workspace_bytes":4312,"sample_counts":{"end_to_end_cold":5,"end_to_end_warm":15,"prepared_cold":5,"prepared_warm":15},"semantic_scope":"complete wrapping-u32 modulo 2^32","source":"c-run","status":"measured","threads":1,"timing_scope":"prepared canonical inputs; public GEMM output allocation, execution, and cascade intermediate cleanup; dense export and final cleanup excluded","work":{"annihilated_macs":0,"executed_macs":56623104,"logical_macs":56623104,"promoted_tiles":0,"promotion_bits":0,"q16_macs":56623104,"q24_macs":0,"q32_macs":0,"q8_macs":0}},{"active_lanes":1,"active_lanes_observed":true,"case":"synthetic-q16","compiler":"gcc","configured_lanes":1,"end_to_end_scope":"caller-dense inputs; input copies, output allocation, zeroing, execution, and input/intermediate cleanup; final output cleanup excluded","engine":"dense-c-blocked","exact_match":true,"general_wrapping_u32":true,"implementation":"compiler-optimized row-major 32x32 blocked C11","medians_ns":{"end_to_end_cold":9937638,"end_to_end_warm":9998470,"prepared_cold":9070467,"prepared_warm":9333043},"output_fnv1a64":"3c3c99e972219fea","reserved_workspace_bytes":0,"sample_counts":{"end_to_end_cold":5,"end_to_end_warm":15,"prepared_cold":5,"prepared_warm":15},"semantic_scope":"complete wrapping-u32 modulo 2^32","source":"c-run","status":"measured","threads":1,"timing_scope":"prepared dense inputs; output allocation, zeroing, execution, and cascade intermediate cleanup; final output cleanup excluded","work":{"annihilated_macs":0,"executed_macs":56623104,"logical_macs":56623104,"promoted_tiles":0,"promotion_bits":0,"q16_macs":0,"q24_macs":0,"q32_macs":0,"q8_macs":0}},{"active_lanes":1,"active_lanes_observed":true,"case":"synthetic-q8-safe-u6","compiler":"gcc","configured_lanes":1,"end_to_end_scope":"dense inputs; canonical import, public GEMM execution, dense export, and input/intermediate cleanup; final dense output cleanup excluded","engine":"qqfenx-c11","exact_match":true,"general_wrapping_u32":true,"implementation":"SSE4.1","medians_ns":{"end_to_end_cold":8763203,"end_to_end_warm":8004578,"prepared_cold":7133608,"prepared_warm":6921010},"output_fnv1a64":"cc47066bd39dcf86","reserved_workspace_bytes":4312,"sample_counts":{"end_to_end_cold":5,"end_to_end_warm":15,"prepared_cold":5,"prepared_warm":15},"semantic_scope":"complete wrapping-u32 modulo 2^32","source":"c-run","status":"measured","threads":1,"timing_scope":"prepared canonical inputs; public GEMM output allocation, execution, and cascade intermediate cleanup; dense export and final cleanup excluded","work":{"annihilated_macs":0,"executed_macs":56623104,"logical_macs":56623104,"promoted_tiles":0,"promotion_bits":0,"q16_macs":0,"q24_macs":0,"q32_macs":0,"q8_macs":56623104}},{"active_lanes":1,"active_lanes_observed":true,"case":"synthetic-q8-safe-u6","compiler":"gcc","configured_lanes":1,"end_to_end_scope":"caller-dense inputs; input copies, output allocation, zeroing, execution, and input/intermediate cleanup; final output cleanup excluded","engine":"dense-c-blocked","exact_match":true,"general_wrapping_u32":true,"implementation":"compiler-optimized row-major 32x32 blocked C11","medians_ns":{"end_to_end_cold":10119871,"end_to_end_warm":10018205,"prepared_cold":9024609,"prepared_warm":9153441},"output_fnv1a64":"cc47066bd39dcf86","reserved_workspace_bytes":0,"sample_counts":{"end_to_end_cold":5,"end_to_end_warm":15,"prepared_cold":5,"prepared_warm":15},"semantic_scope":"complete wrapping-u32 modulo 2^32","source":"c-run","status":"measured","threads":1,"timing_scope":"prepared dense inputs; output allocation, zeroing, execution, and cascade intermediate cleanup; final output cleanup excluded","work":{"annihilated_macs":0,"executed_macs":56623104,"logical_macs":56623104,"promoted_tiles":0,"promotion_bits":0,"q16_macs":0,"q24_macs":0,"q32_macs":0,"q8_macs":0}},{"active_lanes":0,"active_lanes_observed":false,"case":"synthetic-q8-safe-u6","compiler":"gcc","configured_lanes":1,"end_to_end_scope":"caller wrapping-u32 inputs; float conversion, input/output allocation, SGEMM, exact-integral validation, wrapping-u32 reconstruction, and float-input cleanup; final output cleanup excluded","engine":"openblas-sgemm","exact_match":true,"general_wrapping_u32":false,"implementation":"external CBLAS SGEMM","medians_ns":{"end_to_end_cold":4373739,"end_to_end_warm":4252727,"prepared_cold":3871925,"prepared_warm":3440184},"output_fnv1a64":"cc47066bd39dcf86","reserved_workspace_bytes":0,"sample_counts":{"end_to_end_cold":5,"end_to_end_warm":15,"prepared_cold":5,"prepared_warm":15},"semantic_scope":"workload-constrained exact integer subset: nonnegative integral dot products remain below 2^24; independently reconstructed and compared to wrapping-u32 oracle","source":"c-run","status":"measured","threads":1,"timing_scope":"prepared float inputs; output allocation, SGEMM, exact-integral validation, and wrapping-u32 reconstruction; cleanup excluded","work":{"annihilated_macs":0,"executed_macs":56623104,"logical_macs":56623104,"promoted_tiles":0,"promotion_bits":0,"q16_macs":0,"q24_macs":0,"q32_macs":0,"q8_macs":0}},{"active_lanes":1,"active_lanes_observed":true,"case":"synthetic-zero-ideal","compiler":"gcc","configured_lanes":1,"end_to_end_scope":"dense inputs; canonical import, public GEMM execution, dense export, and input/intermediate cleanup; final dense output cleanup excluded","engine":"qqfenx-c11","exact_match":true,"general_wrapping_u32":true,"implementation":"SSE4.1","medians_ns":{"end_to_end_cold":1446387,"end_to_end_warm":1116801,"prepared_cold":476319,"prepared_warm":661915},"output_fnv1a64":"47e2fabfb2562325","reserved_workspace_bytes":4312,"sample_counts":{"end_to_end_cold":5,"end_to_end_warm":15,"prepared_cold":5,"prepared_warm":15},"semantic_scope":"complete wrapping-u32 modulo 2^32","source":"c-run","status":"measured","threads":1,"timing_scope":"prepared canonical inputs; public GEMM output allocation, execution, and cascade intermediate cleanup; dense export and final cleanup excluded","work":{"annihilated_macs":56623104,"executed_macs":0,"logical_macs":56623104,"promoted_tiles":0,"promotion_bits":0,"q16_macs":0,"q24_macs":0,"q32_macs":0,"q8_macs":0}},{"active_lanes":1,"active_lanes_observed":true,"case":"synthetic-zero-ideal","compiler":"gcc","configured_lanes":1,"end_to_end_scope":"caller-dense inputs; input copies, output allocation, zeroing, execution, and input/intermediate cleanup; final output cleanup excluded","engine":"dense-c-blocked","exact_match":true,"general_wrapping_u32":true,"implementation":"compiler-optimized row-major 32x32 blocked C11","medians_ns":{"end_to_end_cold":10033425,"end_to_end_warm":9981838,"prepared_cold":9879049,"prepared_warm":9031405},"output_fnv1a64":"47e2fabfb2562325","reserved_workspace_bytes":0,"sample_counts":{"end_to_end_cold":5,"end_to_end_warm":15,"prepared_cold":5,"prepared_warm":15},"semantic_scope":"complete wrapping-u32 modulo 2^32","source":"c-run","status":"measured","threads":1,"timing_scope":"prepared dense inputs; output allocation, zeroing, execution, and cascade intermediate cleanup; final output cleanup excluded","work":{"annihilated_macs":0,"executed_macs":56623104,"logical_macs":56623104,"promoted_tiles":0,"promotion_bits":0,"q16_macs":0,"q24_macs":0,"q32_macs":0,"q8_macs":0}},{"active_lanes":1,"active_lanes_observed":true,"case":"synthetic-cancellation-cascade","compiler":"gcc","configured_lanes":1,"end_to_end_scope":"dense inputs; canonical import, public GEMM execution, dense export, and input/intermediate cleanup; final dense output cleanup excluded","engine":"qqfenx-c11","exact_match":true,"general_wrapping_u32":true,"implementation":"SSE4.1","medians_ns":{"end_to_end_cold":8486057,"end_to_end_warm":8305739,"prepared_cold":7776927,"prepared_warm":7203382},"output_fnv1a64":"47e2fabfb2562325","reserved_workspace_bytes":4312,"sample_counts":{"end_to_end_cold":5,"end_to_end_warm":15,"prepared_cold":5,"prepared_warm":15},"semantic_scope":"complete wrapping-u32 modulo 2^32","source":"c-run","status":"measured","threads":1,"timing_scope":"prepared canonical inputs; public GEMM output allocation, execution, and cascade intermediate cleanup; dense export and final cleanup excluded","work":{"annihilated_macs":56623104,"executed_macs":56623104,"logical_macs":113246208,"promoted_tiles":144,"promotion_bits":1008,"q16_macs":0,"q24_macs":0,"q32_macs":0,"q8_macs":56623104}},{"active_lanes":1,"active_lanes_observed":true,"case":"synthetic-cancellation-cascade","compiler":"gcc","configured_lanes":1,"end_to_end_scope":"caller-dense inputs; input copies, output allocation, zeroing, execution, and input/intermediate cleanup; final output cleanup excluded","engine":"dense-c-blocked","exact_match":true,"general_wrapping_u32":true,"implementation":"compiler-optimized row-major 32x32 blocked C11","medians_ns":{"end_to_end_cold":19648787,"end_to_end_warm":20982410,"prepared_cold":19480077,"prepared_warm":18303917},"output_fnv1a64":"47e2fabfb2562325","reserved_workspace_bytes":0,"sample_counts":{"end_to_end_cold":5,"end_to_end_warm":15,"prepared_cold":5,"prepared_warm":15},"semantic_scope":"complete wrapping-u32 modulo 2^32","source":"c-run","status":"measured","threads":1,"timing_scope":"prepared dense inputs; output allocation, zeroing, execution, and cascade intermediate cleanup; final output cleanup excluded","work":{"annihilated_macs":0,"executed_macs":113246208,"logical_macs":113246208,"promoted_tiles":0,"promotion_bits":0,"q16_macs":0,"q24_macs":0,"q32_macs":0,"q8_macs":0}},{"active_lanes":1,"active_lanes_observed":true,"case":"uci-iris-feature-gram-i32-v1","compiler":"gcc","configured_lanes":1,"end_to_end_scope":"dense inputs; canonical import, public GEMM execution, dense export, and input/intermediate cleanup; final dense output cleanup excluded","engine":"qqfenx-c11","exact_match":true,"general_wrapping_u32":true,"implementation":"SSE4.1","medians_ns":{"end_to_end_cold":36673,"end_to_end_warm":29506,"prepared_cold":29864,"prepared_warm":27314},"output_fnv1a64":"21255eaa70abb088","reserved_workspace_bytes":232,"retained_output_u32":[522385,267343,348376,112814,267343,143040,167430,53189,348376,167430,258271,86911,112814,53189,86911,30233],"sample_counts":{"end_to_end_cold":5,"end_to_end_warm":15,"prepared_cold":5,"prepared_warm":15},"semantic_scope":"complete wrapping-u32 modulo 2^32","source":"c-run","status":"measured","threads":1,"timing_scope":"prepared canonical inputs; public GEMM output allocation, execution, and cascade intermediate cleanup; dense export and final cleanup excluded","work":{"annihilated_macs":0,"executed_macs":2400,"logical_macs":2400,"promoted_tiles":0,"promotion_bits":0,"q16_macs":0,"q24_macs":0,"q32_macs":2400,"q8_macs":0}},{"active_lanes":1,"active_lanes_observed":true,"case":"uci-iris-feature-gram-i32-v1","compiler":"gcc","configured_lanes":1,"end_to_end_scope":"caller-dense inputs; input copies, output allocation, zeroing, execution, and input/intermediate cleanup; final output cleanup excluded","engine":"dense-c-blocked","exact_match":true,"general_wrapping_u32":true,"implementation":"compiler-optimized row-major 32x32 blocked C11","medians_ns":{"end_to_end_cold":5616,"end_to_end_warm":1311,"prepared_cold":2146,"prepared_warm":923},"output_fnv1a64":"21255eaa70abb088","reserved_workspace_bytes":0,"retained_output_u32":[522385,267343,348376,112814,267343,143040,167430,53189,348376,167430,258271,86911,112814,53189,86911,30233],"sample_counts":{"end_to_end_cold":5,"end_to_end_warm":15,"prepared_cold":5,"prepared_warm":15},"semantic_scope":"complete wrapping-u32 modulo 2^32","source":"c-run","status":"measured","threads":1,"timing_scope":"prepared dense inputs; output allocation, zeroing, execution, and cascade intermediate cleanup; final output cleanup excluded","work":{"annihilated_macs":0,"executed_macs":2400,"logical_macs":2400,"promoted_tiles":0,"promotion_bits":0,"q16_macs":0,"q24_macs":0,"q32_macs":0,"q8_macs":0}},{"active_lanes":0,"active_lanes_observed":false,"case":"uci-iris-feature-gram-i32-v1","compiler":"gcc","configured_lanes":1,"end_to_end_scope":"caller wrapping-u32 inputs; float conversion, input/output allocation, SGEMM, exact-integral validation, wrapping-u32 reconstruction, and float-input cleanup; final output cleanup excluded","engine":"openblas-sgemm","exact_match":true,"general_wrapping_u32":false,"implementation":"external CBLAS SGEMM","medians_ns":{"end_to_end_cold":10200,"end_to_end_warm":1283,"prepared_cold":9076,"prepared_warm":801},"output_fnv1a64":"21255eaa70abb088","reserved_workspace_bytes":0,"retained_output_u32":[522385,267343,348376,112814,267343,143040,167430,53189,348376,167430,258271,86911,112814,53189,86911,30233],"sample_counts":{"end_to_end_cold":5,"end_to_end_warm":15,"prepared_cold":5,"prepared_warm":15},"semantic_scope":"workload-constrained exact integer subset: nonnegative integral dot products remain below 2^24; independently reconstructed and compared to wrapping-u32 oracle","source":"c-run","status":"measured","threads":1,"timing_scope":"prepared float inputs; output allocation, SGEMM, exact-integral validation, and wrapping-u32 reconstruction; cleanup excluded","work":{"annihilated_macs":0,"executed_macs":2400,"logical_macs":2400,"promoted_tiles":0,"promotion_bits":0,"q16_macs":0,"q24_macs":0,"q32_macs":0,"q8_macs":0}},{"active_lanes":4,"active_lanes_observed":true,"case":"synthetic-q32","compiler":"gcc","configured_lanes":4,"end_to_end_scope":"dense inputs; canonical import, public GEMM execution, dense export, and input/intermediate cleanup; final dense output cleanup excluded","engine":"qqfenx-c11","exact_match":true,"general_wrapping_u32":true,"implementation":"SSE4.1","medians_ns":{"end_to_end_cold":7203211,"end_to_end_warm":8168781,"prepared_cold":12073749,"prepared_warm":9682944},"output_fnv1a64":"8697e608b10aed1e","reserved_workspace_bytes":17248,"sample_counts":{"end_to_end_cold":5,"end_to_end_warm":15,"prepared_cold":5,"prepared_warm":15},"semantic_scope":"complete wrapping-u32 modulo 2^32","source":"c-run","status":"measured","threads":4,"timing_scope":"prepared canonical inputs; public GEMM output allocation, execution, and cascade intermediate cleanup; dense export and final cleanup excluded","work":{"annihilated_macs":0,"executed_macs":56623104,"logical_macs":56623104,"promoted_tiles":0,"promotion_bits":0,"q16_macs":0,"q24_macs":0,"q32_macs":56623104,"q8_macs":0}},{"active_lanes":1,"active_lanes_observed":true,"case":"synthetic-q32","compiler":"gcc","configured_lanes":1,"end_to_end_scope":"caller-dense inputs; input copies, output allocation, zeroing, execution, and input/intermediate cleanup; final output cleanup excluded","engine":"dense-c-blocked","exact_match":true,"general_wrapping_u32":true,"implementation":"compiler-optimized row-major 32x32 blocked C11","medians_ns":{"end_to_end_cold":10554241,"end_to_end_warm":10382912,"prepared_cold":9590447,"prepared_warm":9202152},"output_fnv1a64":"8697e608b10aed1e","reserved_workspace_bytes":0,"sample_counts":{"end_to_end_cold":5,"end_to_end_warm":15,"prepared_cold":5,"prepared_warm":15},"semantic_scope":"complete wrapping-u32 modulo 2^32","source":"c-run","status":"measured","threads":4,"timing_scope":"prepared dense inputs; output allocation, zeroing, execution, and cascade intermediate cleanup; final output cleanup excluded","work":{"annihilated_macs":0,"executed_macs":56623104,"logical_macs":56623104,"promoted_tiles":0,"promotion_bits":0,"q16_macs":0,"q24_macs":0,"q32_macs":0,"q8_macs":0}},{"active_lanes":4,"active_lanes_observed":true,"case":"synthetic-q16","compiler":"gcc","configured_lanes":4,"end_to_end_scope":"dense inputs; canonical import, public GEMM execution, dense export, and input/intermediate cleanup; final dense output cleanup excluded","engine":"qqfenx-c11","exact_match":true,"general_wrapping_u32":true,"implementation":"SSE4.1","medians_ns":{"end_to_end_cold":6007652,"end_to_end_warm":4246865,"prepared_cold":2872716,"prepared_warm":2559337},"output_fnv1a64":"3c3c99e972219fea","reserved_workspace_bytes":17248,"sample_counts":{"end_to_end_cold":5,"end_to_end_warm":15,"prepared_cold":5,"prepared_warm":15},"semantic_scope":"complete wrapping-u32 modulo 2^32","source":"c-run","status":"measured","threads":4,"timing_scope":"prepared canonical inputs; public GEMM output allocation, execution, and cascade intermediate cleanup; dense export and final cleanup excluded","work":{"annihilated_macs":0,"executed_macs":56623104,"logical_macs":56623104,"promoted_tiles":0,"promotion_bits":0,"q16_macs":56623104,"q24_macs":0,"q32_macs":0,"q8_macs":0}},{"active_lanes":1,"active_lanes_observed":true,"case":"synthetic-q16","compiler":"gcc","configured_lanes":1,"end_to_end_scope":"caller-dense inputs; input copies, output allocation, zeroing, execution, and input/intermediate cleanup; final output cleanup excluded","engine":"dense-c-blocked","exact_match":true,"general_wrapping_u32":true,"implementation":"compiler-optimized row-major 32x32 blocked C11","medians_ns":{"end_to_end_cold":9935969,"end_to_end_warm":9784937,"prepared_cold":9064091,"prepared_warm":9408821},"output_fnv1a64":"3c3c99e972219fea","reserved_workspace_bytes":0,"sample_counts":{"end_to_end_cold":5,"end_to_end_warm":15,"prepared_cold":5,"prepared_warm":15},"semantic_scope":"complete wrapping-u32 modulo 2^32","source":"c-run","status":"measured","threads":4,"timing_scope":"prepared dense inputs; output allocation, zeroing, execution, and cascade intermediate cleanup; final output cleanup excluded","work":{"annihilated_macs":0,"executed_macs":56623104,"logical_macs":56623104,"promoted_tiles":0,"promotion_bits":0,"q16_macs":0,"q24_macs":0,"q32_macs":0,"q8_macs":0}},{"active_lanes":4,"active_lanes_observed":true,"case":"synthetic-q8-safe-u6","compiler":"gcc","configured_lanes":4,"end_to_end_scope":"dense inputs; canonical import, public GEMM execution, dense export, and input/intermediate cleanup; final dense output cleanup excluded","engine":"qqfenx-c11","exact_match":true,"general_wrapping_u32":true,"implementation":"SSE4.1","medians_ns":{"end_to_end_cold":3459371,"end_to_end_warm":3715542,"prepared_cold":2581637,"prepared_warm":3314000},"output_fnv1a64":"cc47066bd39dcf86","reserved_workspace_bytes":17248,"sample_counts":{"end_to_end_cold":5,"end_to_end_warm":15,"prepared_cold":5,"prepared_warm":15},"semantic_scope":"complete wrapping-u32 modulo 2^32","source":"c-run","status":"measured","threads":4,"timing_scope":"prepared canonical inputs; public GEMM output allocation, execution, and cascade intermediate cleanup; dense export and final cleanup excluded","work":{"annihilated_macs":0,"executed_macs":56623104,"logical_macs":56623104,"promoted_tiles":0,"promotion_bits":0,"q16_macs":0,"q24_macs":0,"q32_macs":0,"q8_macs":56623104}},{"active_lanes":1,"active_lanes_observed":true,"case":"synthetic-q8-safe-u6","compiler":"gcc","configured_lanes":1,"end_to_end_scope":"caller-dense inputs; input copies, output allocation, zeroing, execution, and input/intermediate cleanup; final output cleanup excluded","engine":"dense-c-blocked","exact_match":true,"general_wrapping_u32":true,"implementation":"compiler-optimized row-major 32x32 blocked C11","medians_ns":{"end_to_end_cold":11182129,"end_to_end_warm":10542350,"prepared_cold":9695927,"prepared_warm":9128778},"output_fnv1a64":"cc47066bd39dcf86","reserved_workspace_bytes":0,"sample_counts":{"end_to_end_cold":5,"end_to_end_warm":15,"prepared_cold":5,"prepared_warm":15},"semantic_scope":"complete wrapping-u32 modulo 2^32","source":"c-run","status":"measured","threads":4,"timing_scope":"prepared dense inputs; output allocation, zeroing, execution, and cascade intermediate cleanup; final output cleanup excluded","work":{"annihilated_macs":0,"executed_macs":56623104,"logical_macs":56623104,"promoted_tiles":0,"promotion_bits":0,"q16_macs":0,"q24_macs":0,"q32_macs":0,"q8_macs":0}},{"active_lanes":0,"active_lanes_observed":false,"case":"synthetic-q8-safe-u6","compiler":"gcc","configured_lanes":4,"end_to_end_scope":"caller wrapping-u32 inputs; float conversion, input/output allocation, SGEMM, exact-integral validation, wrapping-u32 reconstruction, and float-input cleanup; final output cleanup excluded","engine":"openblas-sgemm","exact_match":true,"general_wrapping_u32":false,"implementation":"external CBLAS SGEMM","medians_ns":{"end_to_end_cold":5958872,"end_to_end_warm":3561951,"prepared_cold":1879297,"prepared_warm":1582937},"output_fnv1a64":"cc47066bd39dcf86","reserved_workspace_bytes":0,"sample_counts":{"end_to_end_cold":5,"end_to_end_warm":15,"prepared_cold":5,"prepared_warm":15},"semantic_scope":"workload-constrained exact integer subset: nonnegative integral dot products remain below 2^24; independently reconstructed and compared to wrapping-u32 oracle","source":"c-run","status":"measured","threads":4,"timing_scope":"prepared float inputs; output allocation, SGEMM, exact-integral validation, and wrapping-u32 reconstruction; cleanup excluded","work":{"annihilated_macs":0,"executed_macs":56623104,"logical_macs":56623104,"promoted_tiles":0,"promotion_bits":0,"q16_macs":0,"q24_macs":0,"q32_macs":0,"q8_macs":0}},{"active_lanes":4,"active_lanes_observed":true,"case":"synthetic-zero-ideal","compiler":"gcc","configured_lanes":4,"end_to_end_scope":"dense inputs; canonical import, public GEMM execution, dense export, and input/intermediate cleanup; final dense output cleanup excluded","engine":"qqfenx-c11","exact_match":true,"general_wrapping_u32":true,"implementation":"SSE4.1","medians_ns":{"end_to_end_cold":1435740,"end_to_end_warm":951811,"prepared_cold":1190678,"prepared_warm":899477},"output_fnv1a64":"47e2fabfb2562325","reserved_workspace_bytes":17248,"sample_counts":{"end_to_end_cold":5,"end_to_end_warm":15,"prepared_cold":5,"prepared_warm":15},"semantic_scope":"complete wrapping-u32 modulo 2^32","source":"c-run","status":"measured","threads":4,"timing_scope":"prepared canonical inputs; public GEMM output allocation, execution, and cascade intermediate cleanup; dense export and final cleanup excluded","work":{"annihilated_macs":56623104,"executed_macs":0,"logical_macs":56623104,"promoted_tiles":0,"promotion_bits":0,"q16_macs":0,"q24_macs":0,"q32_macs":0,"q8_macs":0}},{"active_lanes":1,"active_lanes_observed":true,"case":"synthetic-zero-ideal","compiler":"gcc","configured_lanes":1,"end_to_end_scope":"caller-dense inputs; input copies, output allocation, zeroing, execution, and input/intermediate cleanup; final output cleanup excluded","engine":"dense-c-blocked","exact_match":true,"general_wrapping_u32":true,"implementation":"compiler-optimized row-major 32x32 blocked C11","medians_ns":{"end_to_end_cold":9924362,"end_to_end_warm":9205559,"prepared_cold":9574662,"prepared_warm":9571637},"output_fnv1a64":"47e2fabfb2562325","reserved_workspace_bytes":0,"sample_counts":{"end_to_end_cold":5,"end_to_end_warm":15,"prepared_cold":5,"prepared_warm":15},"semantic_scope":"complete wrapping-u32 modulo 2^32","source":"c-run","status":"measured","threads":4,"timing_scope":"prepared dense inputs; output allocation, zeroing, execution, and cascade intermediate cleanup; final output cleanup excluded","work":{"annihilated_macs":0,"executed_macs":56623104,"logical_macs":56623104,"promoted_tiles":0,"promotion_bits":0,"q16_macs":0,"q24_macs":0,"q32_macs":0,"q8_macs":0}},{"active_lanes":4,"active_lanes_observed":true,"case":"synthetic-cancellation-cascade","compiler":"gcc","configured_lanes":4,"end_to_end_scope":"dense inputs; canonical import, public GEMM execution, dense export, and input/intermediate cleanup; final dense output cleanup excluded","engine":"qqfenx-c11","exact_match":true,"general_wrapping_u32":true,"implementation":"SSE4.1","medians_ns":{"end_to_end_cold":4366553,"end_to_end_warm":3431578,"prepared_cold":3915876,"prepared_warm":2748542},"output_fnv1a64":"47e2fabfb2562325","reserved_workspace_bytes":17248,"sample_counts":{"end_to_end_cold":5,"end_to_end_warm":15,"prepared_cold":5,"prepared_warm":15},"semantic_scope":"complete wrapping-u32 modulo 2^32","source":"c-run","status":"measured","threads":4,"timing_scope":"prepared canonical inputs; public GEMM output allocation, execution, and cascade intermediate cleanup; dense export and final cleanup excluded","work":{"annihilated_macs":56623104,"executed_macs":56623104,"logical_macs":113246208,"promoted_tiles":144,"promotion_bits":1008,"q16_macs":0,"q24_macs":0,"q32_macs":0,"q8_macs":56623104}},{"active_lanes":1,"active_lanes_observed":true,"case":"synthetic-cancellation-cascade","compiler":"gcc","configured_lanes":1,"end_to_end_scope":"caller-dense inputs; input copies, output allocation, zeroing, execution, and input/intermediate cleanup; final output cleanup excluded","engine":"dense-c-blocked","exact_match":true,"general_wrapping_u32":true,"implementation":"compiler-optimized row-major 32x32 blocked C11","medians_ns":{"end_to_end_cold":18902446,"end_to_end_warm":18839365,"prepared_cold":18993421,"prepared_warm":18746602},"output_fnv1a64":"47e2fabfb2562325","reserved_workspace_bytes":0,"sample_counts":{"end_to_end_cold":5,"end_to_end_warm":15,"prepared_cold":5,"prepared_warm":15},"semantic_scope":"complete wrapping-u32 modulo 2^32","source":"c-run","status":"measured","threads":4,"timing_scope":"prepared dense inputs; output allocation, zeroing, execution, and cascade intermediate cleanup; final output cleanup excluded","work":{"annihilated_macs":0,"executed_macs":113246208,"logical_macs":113246208,"promoted_tiles":0,"promotion_bits":0,"q16_macs":0,"q24_macs":0,"q32_macs":0,"q8_macs":0}},{"active_lanes":4,"active_lanes_observed":true,"case":"uci-iris-feature-gram-i32-v1","compiler":"gcc","configured_lanes":4,"end_to_end_scope":"dense inputs; canonical import, public GEMM execution, dense export, and input/intermediate cleanup; final dense output cleanup excluded","engine":"qqfenx-c11","exact_match":true,"general_wrapping_u32":true,"implementation":"SSE4.1","medians_ns":{"end_to_end_cold":489547,"end_to_end_warm":91914,"prepared_cold":180161,"prepared_warm":76607},"output_fnv1a64":"21255eaa70abb088","reserved_workspace_bytes":928,"retained_output_u32":[522385,267343,348376,112814,267343,143040,167430,53189,348376,167430,258271,86911,112814,53189,86911,30233],"sample_counts":{"end_to_end_cold":5,"end_to_end_warm":15,"prepared_cold":5,"prepared_warm":15},"semantic_scope":"complete wrapping-u32 modulo 2^32","source":"c-run","status":"measured","threads":4,"timing_scope":"prepared canonical inputs; public GEMM output allocation, execution, and cascade intermediate cleanup; dense export and final cleanup excluded","work":{"annihilated_macs":0,"executed_macs":2400,"logical_macs":2400,"promoted_tiles":0,"promotion_bits":0,"q16_macs":0,"q24_macs":0,"q32_macs":2400,"q8_macs":0}},{"active_lanes":1,"active_lanes_observed":true,"case":"uci-iris-feature-gram-i32-v1","compiler":"gcc","configured_lanes":1,"end_to_end_scope":"caller-dense inputs; input copies, output allocation, zeroing, execution, and input/intermediate cleanup; final output cleanup excluded","engine":"dense-c-blocked","exact_match":true,"general_wrapping_u32":true,"implementation":"compiler-optimized row-major 32x32 blocked C11","medians_ns":{"end_to_end_cold":5154,"end_to_end_warm":1398,"prepared_cold":2459,"prepared_warm":1067},"output_fnv1a64":"21255eaa70abb088","reserved_workspace_bytes":0,"retained_output_u32":[522385,267343,348376,112814,267343,143040,167430,53189,348376,167430,258271,86911,112814,53189,86911,30233],"sample_counts":{"end_to_end_cold":5,"end_to_end_warm":15,"prepared_cold":5,"prepared_warm":15},"semantic_scope":"complete wrapping-u32 modulo 2^32","source":"c-run","status":"measured","threads":4,"timing_scope":"prepared dense inputs; output allocation, zeroing, execution, and cascade intermediate cleanup; final output cleanup excluded","work":{"annihilated_macs":0,"executed_macs":2400,"logical_macs":2400,"promoted_tiles":0,"promotion_bits":0,"q16_macs":0,"q24_macs":0,"q32_macs":0,"q8_macs":0}},{"active_lanes":0,"active_lanes_observed":false,"case":"uci-iris-feature-gram-i32-v1","compiler":"gcc","configured_lanes":4,"end_to_end_scope":"caller wrapping-u32 inputs; float conversion, input/output allocation, SGEMM, exact-integral validation, wrapping-u32 reconstruction, and float-input cleanup; final output cleanup excluded","engine":"openblas-sgemm","exact_match":true,"general_wrapping_u32":false,"implementation":"external CBLAS SGEMM","medians_ns":{"end_to_end_cold":16397,"end_to_end_warm":1442,"prepared_cold":8483,"prepared_warm":684},"output_fnv1a64":"21255eaa70abb088","reserved_workspace_bytes":0,"retained_output_u32":[522385,267343,348376,112814,267343,143040,167430,53189,348376,167430,258271,86911,112814,53189,86911,30233],"sample_counts":{"end_to_end_cold":5,"end_to_end_warm":15,"prepared_cold":5,"prepared_warm":15},"semantic_scope":"workload-constrained exact integer subset: nonnegative integral dot products remain below 2^24; independently reconstructed and compared to wrapping-u32 oracle","source":"c-run","status":"measured","threads":4,"timing_scope":"prepared float inputs; output allocation, SGEMM, exact-integral validation, and wrapping-u32 reconstruction; cleanup excluded","work":{"annihilated_macs":0,"executed_macs":2400,"logical_macs":2400,"promoted_tiles":0,"promotion_bits":0,"q16_macs":0,"q24_macs":0,"q32_macs":0,"q8_macs":0}},{"active_lanes":1,"active_lanes_observed":true,"case":"synthetic-q32","compiler":"clang","configured_lanes":1,"end_to_end_scope":"dense inputs; canonical import, public GEMM execution, dense export, and input/intermediate cleanup; final dense output cleanup excluded","engine":"qqfenx-c11","exact_match":true,"general_wrapping_u32":true,"implementation":"SSE4.1","medians_ns":{"end_to_end_cold":17969592,"end_to_end_warm":17761731,"prepared_cold":18490633,"prepared_warm":17584908},"output_fnv1a64":"8697e608b10aed1e","reserved_workspace_bytes":4312,"sample_counts":{"end_to_end_cold":5,"end_to_end_warm":15,"prepared_cold":5,"prepared_warm":15},"semantic_scope":"complete wrapping-u32 modulo 2^32","source":"c-run","status":"measured","threads":1,"timing_scope":"prepared canonical inputs; public GEMM output allocation, execution, and cascade intermediate cleanup; dense export and final cleanup excluded","work":{"annihilated_macs":0,"executed_macs":56623104,"logical_macs":56623104,"promoted_tiles":0,"promotion_bits":0,"q16_macs":0,"q24_macs":0,"q32_macs":56623104,"q8_macs":0}},{"active_lanes":1,"active_lanes_observed":true,"case":"synthetic-q32","compiler":"clang","configured_lanes":1,"end_to_end_scope":"caller-dense inputs; input copies, output allocation, zeroing, execution, and input/intermediate cleanup; final output cleanup excluded","engine":"dense-c-blocked","exact_match":true,"general_wrapping_u32":true,"implementation":"compiler-optimized row-major 32x32 blocked C11","medians_ns":{"end_to_end_cold":10581736,"end_to_end_warm":11093301,"prepared_cold":10605372,"prepared_warm":9655588},"output_fnv1a64":"8697e608b10aed1e","reserved_workspace_bytes":0,"sample_counts":{"end_to_end_cold":5,"end_to_end_warm":15,"prepared_cold":5,"prepared_warm":15},"semantic_scope":"complete wrapping-u32 modulo 2^32","source":"c-run","status":"measured","threads":1,"timing_scope":"prepared dense inputs; output allocation, zeroing, execution, and cascade intermediate cleanup; final output cleanup excluded","work":{"annihilated_macs":0,"executed_macs":56623104,"logical_macs":56623104,"promoted_tiles":0,"promotion_bits":0,"q16_macs":0,"q24_macs":0,"q32_macs":0,"q8_macs":0}},{"active_lanes":1,"active_lanes_observed":true,"case":"synthetic-q16","compiler":"clang","configured_lanes":1,"end_to_end_scope":"dense inputs; canonical import, public GEMM execution, dense export, and input/intermediate cleanup; final dense output cleanup excluded","engine":"qqfenx-c11","exact_match":true,"general_wrapping_u32":true,"implementation":"SSE4.1","medians_ns":{"end_to_end_cold":8749889,"end_to_end_warm":9058738,"prepared_cold":7328136,"prepared_warm":8063382},"output_fnv1a64":"3c3c99e972219fea","reserved_workspace_bytes":4312,"sample_counts":{"end_to_end_cold":5,"end_to_end_warm":15,"prepared_cold":5,"prepared_warm":15},"semantic_scope":"complete wrapping-u32 modulo 2^32","source":"c-run","status":"measured","threads":1,"timing_scope":"prepared canonical inputs; public GEMM output allocation, execution, and cascade intermediate cleanup; dense export and final cleanup excluded","work":{"annihilated_macs":0,"executed_macs":56623104,"logical_macs":56623104,"promoted_tiles":0,"promotion_bits":0,"q16_macs":56623104,"q24_macs":0,"q32_macs":0,"q8_macs":0}},{"active_lanes":1,"active_lanes_observed":true,"case":"synthetic-q16","compiler":"clang","configured_lanes":1,"end_to_end_scope":"caller-dense inputs; input copies, output allocation, zeroing, execution, and input/intermediate cleanup; final output cleanup excluded","engine":"dense-c-blocked","exact_match":true,"general_wrapping_u32":true,"implementation":"compiler-optimized row-major 32x32 blocked C11","medians_ns":{"end_to_end_cold":11089191,"end_to_end_warm":11262840,"prepared_cold":10105884,"prepared_warm":9974793},"output_fnv1a64":"3c3c99e972219fea","reserved_workspace_bytes":0,"sample_counts":{"end_to_end_cold":5,"end_to_end_warm":15,"prepared_cold":5,"prepared_warm":15},"semantic_scope":"complete wrapping-u32 modulo 2^32","source":"c-run","status":"measured","threads":1,"timing_scope":"prepared dense inputs; output allocation, zeroing, execution, and cascade intermediate cleanup; final output cleanup excluded","work":{"annihilated_macs":0,"executed_macs":56623104,"logical_macs":56623104,"promoted_tiles":0,"promotion_bits":0,"q16_macs":0,"q24_macs":0,"q32_macs":0,"q8_macs":0}},{"active_lanes":1,"active_lanes_observed":true,"case":"synthetic-q8-safe-u6","compiler":"clang","configured_lanes":1,"end_to_end_scope":"dense inputs; canonical import, public GEMM execution, dense export, and input/intermediate cleanup; final dense output cleanup excluded","engine":"qqfenx-c11","exact_match":true,"general_wrapping_u32":true,"implementation":"SSE4.1","medians_ns":{"end_to_end_cold":8890527,"end_to_end_warm":8682740,"prepared_cold":7713873,"prepared_warm":8083858},"output_fnv1a64":"cc47066bd39dcf86","reserved_workspace_bytes":4312,"sample_counts":{"end_to_end_cold":5,"end_to_end_warm":15,"prepared_cold":5,"prepared_warm":15},"semantic_scope":"complete wrapping-u32 modulo 2^32","source":"c-run","status":"measured","threads":1,"timing_scope":"prepared canonical inputs; public GEMM output allocation, execution, and cascade intermediate cleanup; dense export and final cleanup excluded","work":{"annihilated_macs":0,"executed_macs":56623104,"logical_macs":56623104,"promoted_tiles":0,"promotion_bits":0,"q16_macs":0,"q24_macs":0,"q32_macs":0,"q8_macs":56623104}},{"active_lanes":1,"active_lanes_observed":true,"case":"synthetic-q8-safe-u6","compiler":"clang","configured_lanes":1,"end_to_end_scope":"caller-dense inputs; input copies, output allocation, zeroing, execution, and input/intermediate cleanup; final output cleanup excluded","engine":"dense-c-blocked","exact_match":true,"general_wrapping_u32":true,"implementation":"compiler-optimized row-major 32x32 blocked C11","medians_ns":{"end_to_end_cold":10693614,"end_to_end_warm":11063148,"prepared_cold":10401999,"prepared_warm":10173421},"output_fnv1a64":"cc47066bd39dcf86","reserved_workspace_bytes":0,"sample_counts":{"end_to_end_cold":5,"end_to_end_warm":15,"prepared_cold":5,"prepared_warm":15},"semantic_scope":"complete wrapping-u32 modulo 2^32","source":"c-run","status":"measured","threads":1,"timing_scope":"prepared dense inputs; output allocation, zeroing, execution, and cascade intermediate cleanup; final output cleanup excluded","work":{"annihilated_macs":0,"executed_macs":56623104,"logical_macs":56623104,"promoted_tiles":0,"promotion_bits":0,"q16_macs":0,"q24_macs":0,"q32_macs":0,"q8_macs":0}},{"active_lanes":0,"active_lanes_observed":false,"case":"synthetic-q8-safe-u6","compiler":"clang","configured_lanes":1,"end_to_end_scope":"caller wrapping-u32 inputs; float conversion, input/output allocation, SGEMM, exact-integral validation, wrapping-u32 reconstruction, and float-input cleanup; final output cleanup excluded","engine":"openblas-sgemm","exact_match":true,"general_wrapping_u32":false,"implementation":"external CBLAS SGEMM","medians_ns":{"end_to_end_cold":5433832,"end_to_end_warm":4083611,"prepared_cold":4633705,"prepared_warm":3177676},"output_fnv1a64":"cc47066bd39dcf86","reserved_workspace_bytes":0,"sample_counts":{"end_to_end_cold":5,"end_to_end_warm":15,"prepared_cold":5,"prepared_warm":15},"semantic_scope":"workload-constrained exact integer subset: nonnegative integral dot products remain below 2^24; independently reconstructed and compared to wrapping-u32 oracle","source":"c-run","status":"measured","threads":1,"timing_scope":"prepared float inputs; output allocation, SGEMM, exact-integral validation, and wrapping-u32 reconstruction; cleanup excluded","work":{"annihilated_macs":0,"executed_macs":56623104,"logical_macs":56623104,"promoted_tiles":0,"promotion_bits":0,"q16_macs":0,"q24_macs":0,"q32_macs":0,"q8_macs":0}},{"active_lanes":1,"active_lanes_observed":true,"case":"synthetic-zero-ideal","compiler":"clang","configured_lanes":1,"end_to_end_scope":"dense inputs; canonical import, public GEMM execution, dense export, and input/intermediate cleanup; final dense output cleanup excluded","engine":"qqfenx-c11","exact_match":true,"general_wrapping_u32":true,"implementation":"SSE4.1","medians_ns":{"end_to_end_cold":1817451,"end_to_end_warm":1427712,"prepared_cold":737230,"prepared_warm":817560},"output_fnv1a64":"47e2fabfb2562325","reserved_workspace_bytes":4312,"sample_counts":{"end_to_end_cold":5,"end_to_end_warm":15,"prepared_cold":5,"prepared_warm":15},"semantic_scope":"complete wrapping-u32 modulo 2^32","source":"c-run","status":"measured","threads":1,"timing_scope":"prepared canonical inputs; public GEMM output allocation, execution, and cascade intermediate cleanup; dense export and final cleanup excluded","work":{"annihilated_macs":56623104,"executed_macs":0,"logical_macs":56623104,"promoted_tiles":0,"promotion_bits":0,"q16_macs":0,"q24_macs":0,"q32_macs":0,"q8_macs":0}},{"active_lanes":1,"active_lanes_observed":true,"case":"synthetic-zero-ideal","compiler":"clang","configured_lanes":1,"end_to_end_scope":"caller-dense inputs; input copies, output allocation, zeroing, execution, and input/intermediate cleanup; final output cleanup excluded","engine":"dense-c-blocked","exact_match":true,"general_wrapping_u32":true,"implementation":"compiler-optimized row-major 32x32 blocked C11","medians_ns":{"end_to_end_cold":11382084,"end_to_end_warm":11445629,"prepared_cold":10042491,"prepared_warm":10301916},"output_fnv1a64":"47e2fabfb2562325","reserved_workspace_bytes":0,"sample_counts":{"end_to_end_cold":5,"end_to_end_warm":15,"prepared_cold":5,"prepared_warm":15},"semantic_scope":"complete wrapping-u32 modulo 2^32","source":"c-run","status":"measured","threads":1,"timing_scope":"prepared dense inputs; output allocation, zeroing, execution, and cascade intermediate cleanup; final output cleanup excluded","work":{"annihilated_macs":0,"executed_macs":56623104,"logical_macs":56623104,"promoted_tiles":0,"promotion_bits":0,"q16_macs":0,"q24_macs":0,"q32_macs":0,"q8_macs":0}},{"active_lanes":1,"active_lanes_observed":true,"case":"synthetic-cancellation-cascade","compiler":"clang","configured_lanes":1,"end_to_end_scope":"dense inputs; canonical import, public GEMM execution, dense export, and input/intermediate cleanup; final dense output cleanup excluded","engine":"qqfenx-c11","exact_match":true,"general_wrapping_u32":true,"implementation":"SSE4.1","medians_ns":{"end_to_end_cold":9687757,"end_to_end_warm":9185793,"prepared_cold":8483271,"prepared_warm":8296693},"output_fnv1a64":"47e2fabfb2562325","reserved_workspace_bytes":4312,"sample_counts":{"end_to_end_cold":5,"end_to_end_warm":15,"prepared_cold":5,"prepared_warm":15},"semantic_scope":"complete wrapping-u32 modulo 2^32","source":"c-run","status":"measured","threads":1,"timing_scope":"prepared canonical inputs; public GEMM output allocation, execution, and cascade intermediate cleanup; dense export and final cleanup excluded","work":{"annihilated_macs":56623104,"executed_macs":56623104,"logical_macs":113246208,"promoted_tiles":144,"promotion_bits":1008,"q16_macs":0,"q24_macs":0,"q32_macs":0,"q8_macs":56623104}},{"active_lanes":1,"active_lanes_observed":true,"case":"synthetic-cancellation-cascade","compiler":"clang","configured_lanes":1,"end_to_end_scope":"caller-dense inputs; input copies, output allocation, zeroing, execution, and input/intermediate cleanup; final output cleanup excluded","engine":"dense-c-blocked","exact_match":true,"general_wrapping_u32":true,"implementation":"compiler-optimized row-major 32x32 blocked C11","medians_ns":{"end_to_end_cold":23476417,"end_to_end_warm":22234211,"prepared_cold":20628121,"prepared_warm":20052952},"output_fnv1a64":"47e2fabfb2562325","reserved_workspace_bytes":0,"sample_counts":{"end_to_end_cold":5,"end_to_end_warm":15,"prepared_cold":5,"prepared_warm":15},"semantic_scope":"complete wrapping-u32 modulo 2^32","source":"c-run","status":"measured","threads":1,"timing_scope":"prepared dense inputs; output allocation, zeroing, execution, and cascade intermediate cleanup; final output cleanup excluded","work":{"annihilated_macs":0,"executed_macs":113246208,"logical_macs":113246208,"promoted_tiles":0,"promotion_bits":0,"q16_macs":0,"q24_macs":0,"q32_macs":0,"q8_macs":0}},{"active_lanes":1,"active_lanes_observed":true,"case":"uci-iris-feature-gram-i32-v1","compiler":"clang","configured_lanes":1,"end_to_end_scope":"dense inputs; canonical import, public GEMM execution, dense export, and input/intermediate cleanup; final dense output cleanup excluded","engine":"qqfenx-c11","exact_match":true,"general_wrapping_u32":true,"implementation":"SSE4.1","medians_ns":{"end_to_end_cold":37962,"end_to_end_warm":32806,"prepared_cold":23508,"prepared_warm":18151},"output_fnv1a64":"21255eaa70abb088","reserved_workspace_bytes":232,"retained_output_u32":[522385,267343,348376,112814,267343,143040,167430,53189,348376,167430,258271,86911,112814,53189,86911,30233],"sample_counts":{"end_to_end_cold":5,"end_to_end_warm":15,"prepared_cold":5,"prepared_warm":15},"semantic_scope":"complete wrapping-u32 modulo 2^32","source":"c-run","status":"measured","threads":1,"timing_scope":"prepared canonical inputs; public GEMM output allocation, execution, and cascade intermediate cleanup; dense export and final cleanup excluded","work":{"annihilated_macs":0,"executed_macs":2400,"logical_macs":2400,"promoted_tiles":0,"promotion_bits":0,"q16_macs":0,"q24_macs":0,"q32_macs":2400,"q8_macs":0}},{"active_lanes":1,"active_lanes_observed":true,"case":"uci-iris-feature-gram-i32-v1","compiler":"clang","configured_lanes":1,"end_to_end_scope":"caller-dense inputs; input copies, output allocation, zeroing, execution, and input/intermediate cleanup; final output cleanup excluded","engine":"dense-c-blocked","exact_match":true,"general_wrapping_u32":true,"implementation":"compiler-optimized row-major 32x32 blocked C11","medians_ns":{"end_to_end_cold":5485,"end_to_end_warm":1849,"prepared_cold":3124,"prepared_warm":1852},"output_fnv1a64":"21255eaa70abb088","reserved_workspace_bytes":0,"retained_output_u32":[522385,267343,348376,112814,267343,143040,167430,53189,348376,167430,258271,86911,112814,53189,86911,30233],"sample_counts":{"end_to_end_cold":5,"end_to_end_warm":15,"prepared_cold":5,"prepared_warm":15},"semantic_scope":"complete wrapping-u32 modulo 2^32","source":"c-run","status":"measured","threads":1,"timing_scope":"prepared dense inputs; output allocation, zeroing, execution, and cascade intermediate cleanup; final output cleanup excluded","work":{"annihilated_macs":0,"executed_macs":2400,"logical_macs":2400,"promoted_tiles":0,"promotion_bits":0,"q16_macs":0,"q24_macs":0,"q32_macs":0,"q8_macs":0}},{"active_lanes":0,"active_lanes_observed":false,"case":"uci-iris-feature-gram-i32-v1","compiler":"clang","configured_lanes":1,"end_to_end_scope":"caller wrapping-u32 inputs; float conversion, input/output allocation, SGEMM, exact-integral validation, wrapping-u32 reconstruction, and float-input cleanup; final output cleanup excluded","engine":"openblas-sgemm","exact_match":true,"general_wrapping_u32":false,"implementation":"external CBLAS SGEMM","medians_ns":{"end_to_end_cold":10047,"end_to_end_warm":1503,"prepared_cold":7924,"prepared_warm":771},"output_fnv1a64":"21255eaa70abb088","reserved_workspace_bytes":0,"retained_output_u32":[522385,267343,348376,112814,267343,143040,167430,53189,348376,167430,258271,86911,112814,53189,86911,30233],"sample_counts":{"end_to_end_cold":5,"end_to_end_warm":15,"prepared_cold":5,"prepared_warm":15},"semantic_scope":"workload-constrained exact integer subset: nonnegative integral dot products remain below 2^24; independently reconstructed and compared to wrapping-u32 oracle","source":"c-run","status":"measured","threads":1,"timing_scope":"prepared float inputs; output allocation, SGEMM, exact-integral validation, and wrapping-u32 reconstruction; cleanup excluded","work":{"annihilated_macs":0,"executed_macs":2400,"logical_macs":2400,"promoted_tiles":0,"promotion_bits":0,"q16_macs":0,"q24_macs":0,"q32_macs":0,"q8_macs":0}},{"active_lanes":4,"active_lanes_observed":true,"case":"synthetic-q32","compiler":"clang","configured_lanes":4,"end_to_end_scope":"dense inputs; canonical import, public GEMM execution, dense export, and input/intermediate cleanup; final dense output cleanup excluded","engine":"qqfenx-c11","exact_match":true,"general_wrapping_u32":true,"implementation":"SSE4.1","medians_ns":{"end_to_end_cold":9558640,"end_to_end_warm":8722135,"prepared_cold":6495980,"prepared_warm":7994858},"output_fnv1a64":"8697e608b10aed1e","reserved_workspace_bytes":17248,"sample_counts":{"end_to_end_cold":5,"end_to_end_warm":15,"prepared_cold":5,"prepared_warm":15},"semantic_scope":"complete wrapping-u32 modulo 2^32","source":"c-run","status":"measured","threads":4,"timing_scope":"prepared canonical inputs; public GEMM output allocation, execution, and cascade intermediate cleanup; dense export and final cleanup excluded","work":{"annihilated_macs":0,"executed_macs":56623104,"logical_macs":56623104,"promoted_tiles":0,"promotion_bits":0,"q16_macs":0,"q24_macs":0,"q32_macs":56623104,"q8_macs":0}},{"active_lanes":1,"active_lanes_observed":true,"case":"synthetic-q32","compiler":"clang","configured_lanes":1,"end_to_end_scope":"caller-dense inputs; input copies, output allocation, zeroing, execution, and input/intermediate cleanup; final output cleanup excluded","engine":"dense-c-blocked","exact_match":true,"general_wrapping_u32":true,"implementation":"compiler-optimized row-major 32x32 blocked C11","medians_ns":{"end_to_end_cold":10671607,"end_to_end_warm":10665620,"prepared_cold":10042849,"prepared_warm":10036631},"output_fnv1a64":"8697e608b10aed1e","reserved_workspace_bytes":0,"sample_counts":{"end_to_end_cold":5,"end_to_end_warm":15,"prepared_cold":5,"prepared_warm":15},"semantic_scope":"complete wrapping-u32 modulo 2^32","source":"c-run","status":"measured","threads":4,"timing_scope":"prepared dense inputs; output allocation, zeroing, execution, and cascade intermediate cleanup; final output cleanup excluded","work":{"annihilated_macs":0,"executed_macs":56623104,"logical_macs":56623104,"promoted_tiles":0,"promotion_bits":0,"q16_macs":0,"q24_macs":0,"q32_macs":0,"q8_macs":0}},{"active_lanes":4,"active_lanes_observed":true,"case":"synthetic-q16","compiler":"clang","configured_lanes":4,"end_to_end_scope":"dense inputs; canonical import, public GEMM execution, dense export, and input/intermediate cleanup; final dense output cleanup excluded","engine":"qqfenx-c11","exact_match":true,"general_wrapping_u32":true,"implementation":"SSE4.1","medians_ns":{"end_to_end_cold":5130242,"end_to_end_warm":3914413,"prepared_cold":2940761,"prepared_warm":4464485},"output_fnv1a64":"3c3c99e972219fea","reserved_workspace_bytes":17248,"sample_counts":{"end_to_end_cold":5,"end_to_end_warm":15,"prepared_cold":5,"prepared_warm":15},"semantic_scope":"complete wrapping-u32 modulo 2^32","source":"c-run","status":"measured","threads":4,"timing_scope":"prepared canonical inputs; public GEMM output allocation, execution, and cascade intermediate cleanup; dense export and final cleanup excluded","work":{"annihilated_macs":0,"executed_macs":56623104,"logical_macs":56623104,"promoted_tiles":0,"promotion_bits":0,"q16_macs":56623104,"q24_macs":0,"q32_macs":0,"q8_macs":0}},{"active_lanes":1,"active_lanes_observed":true,"case":"synthetic-q16","compiler":"clang","configured_lanes":1,"end_to_end_scope":"caller-dense inputs; input copies, output allocation, zeroing, execution, and input/intermediate cleanup; final output cleanup excluded","engine":"dense-c-blocked","exact_match":true,"general_wrapping_u32":true,"implementation":"compiler-optimized row-major 32x32 blocked C11","medians_ns":{"end_to_end_cold":11104753,"end_to_end_warm":11232044,"prepared_cold":10283389,"prepared_warm":10276552},"output_fnv1a64":"3c3c99e972219fea","reserved_workspace_bytes":0,"sample_counts":{"end_to_end_cold":5,"end_to_end_warm":15,"prepared_cold":5,"prepared_warm":15},"semantic_scope":"complete wrapping-u32 modulo 2^32","source":"c-run","status":"measured","threads":4,"timing_scope":"prepared dense inputs; output allocation, zeroing, execution, and cascade intermediate cleanup; final output cleanup excluded","work":{"annihilated_macs":0,"executed_macs":56623104,"logical_macs":56623104,"promoted_tiles":0,"promotion_bits":0,"q16_macs":0,"q24_macs":0,"q32_macs":0,"q8_macs":0}},{"active_lanes":4,"active_lanes_observed":true,"case":"synthetic-q8-safe-u6","compiler":"clang","configured_lanes":4,"end_to_end_scope":"dense inputs; canonical import, public GEMM execution, dense export, and input/intermediate cleanup; final dense output cleanup excluded","engine":"qqfenx-c11","exact_match":true,"general_wrapping_u32":true,"implementation":"SSE4.1","medians_ns":{"end_to_end_cold":5018918,"end_to_end_warm":3786223,"prepared_cold":3389327,"prepared_warm":2911923},"output_fnv1a64":"cc47066bd39dcf86","reserved_workspace_bytes":17248,"sample_counts":{"end_to_end_cold":5,"end_to_end_warm":15,"prepared_cold":5,"prepared_warm":15},"semantic_scope":"complete wrapping-u32 modulo 2^32","source":"c-run","status":"measured","threads":4,"timing_scope":"prepared canonical inputs; public GEMM output allocation, execution, and cascade intermediate cleanup; dense export and final cleanup excluded","work":{"annihilated_macs":0,"executed_macs":56623104,"logical_macs":56623104,"promoted_tiles":0,"promotion_bits":0,"q16_macs":0,"q24_macs":0,"q32_macs":0,"q8_macs":56623104}},{"active_lanes":1,"active_lanes_observed":true,"case":"synthetic-q8-safe-u6","compiler":"clang","configured_lanes":1,"end_to_end_scope":"caller-dense inputs; input copies, output allocation, zeroing, execution, and input/intermediate cleanup; final output cleanup excluded","engine":"dense-c-blocked","exact_match":true,"general_wrapping_u32":true,"implementation":"compiler-optimized row-major 32x32 blocked C11","medians_ns":{"end_to_end_cold":11750923,"end_to_end_warm":10739464,"prepared_cold":10585746,"prepared_warm":10013079},"output_fnv1a64":"cc47066bd39dcf86","reserved_workspace_bytes":0,"sample_counts":{"end_to_end_cold":5,"end_to_end_warm":15,"prepared_cold":5,"prepared_warm":15},"semantic_scope":"complete wrapping-u32 modulo 2^32","source":"c-run","status":"measured","threads":4,"timing_scope":"prepared dense inputs; output allocation, zeroing, execution, and cascade intermediate cleanup; final output cleanup excluded","work":{"annihilated_macs":0,"executed_macs":56623104,"logical_macs":56623104,"promoted_tiles":0,"promotion_bits":0,"q16_macs":0,"q24_macs":0,"q32_macs":0,"q8_macs":0}},{"active_lanes":0,"active_lanes_observed":false,"case":"synthetic-q8-safe-u6","compiler":"clang","configured_lanes":4,"end_to_end_scope":"caller wrapping-u32 inputs; float conversion, input/output allocation, SGEMM, exact-integral validation, wrapping-u32 reconstruction, and float-input cleanup; final output cleanup excluded","engine":"openblas-sgemm","exact_match":true,"general_wrapping_u32":false,"implementation":"external CBLAS SGEMM","medians_ns":{"end_to_end_cold":4677555,"end_to_end_warm":2694815,"prepared_cold":1722659,"prepared_warm":1457087},"output_fnv1a64":"cc47066bd39dcf86","reserved_workspace_bytes":0,"sample_counts":{"end_to_end_cold":5,"end_to_end_warm":15,"prepared_cold":5,"prepared_warm":15},"semantic_scope":"workload-constrained exact integer subset: nonnegative integral dot products remain below 2^24; independently reconstructed and compared to wrapping-u32 oracle","source":"c-run","status":"measured","threads":4,"timing_scope":"prepared float inputs; output allocation, SGEMM, exact-integral validation, and wrapping-u32 reconstruction; cleanup excluded","work":{"annihilated_macs":0,"executed_macs":56623104,"logical_macs":56623104,"promoted_tiles":0,"promotion_bits":0,"q16_macs":0,"q24_macs":0,"q32_macs":0,"q8_macs":0}},{"active_lanes":4,"active_lanes_observed":true,"case":"synthetic-zero-ideal","compiler":"clang","configured_lanes":4,"end_to_end_scope":"dense inputs; canonical import, public GEMM execution, dense export, and input/intermediate cleanup; final dense output cleanup excluded","engine":"qqfenx-c11","exact_match":true,"general_wrapping_u32":true,"implementation":"SSE4.1","medians_ns":{"end_to_end_cold":1391880,"end_to_end_warm":1541776,"prepared_cold":1236048,"prepared_warm":1017463},"output_fnv1a64":"47e2fabfb2562325","reserved_workspace_bytes":17248,"sample_counts":{"end_to_end_cold":5,"end_to_end_warm":15,"prepared_cold":5,"prepared_warm":15},"semantic_scope":"complete wrapping-u32 modulo 2^32","source":"c-run","status":"measured","threads":4,"timing_scope":"prepared canonical inputs; public GEMM output allocation, execution, and cascade intermediate cleanup; dense export and final cleanup excluded","work":{"annihilated_macs":56623104,"executed_macs":0,"logical_macs":56623104,"promoted_tiles":0,"promotion_bits":0,"q16_macs":0,"q24_macs":0,"q32_macs":0,"q8_macs":0}},{"active_lanes":1,"active_lanes_observed":true,"case":"synthetic-zero-ideal","compiler":"clang","configured_lanes":1,"end_to_end_scope":"caller-dense inputs; input copies, output allocation, zeroing, execution, and input/intermediate cleanup; final output cleanup excluded","engine":"dense-c-blocked","exact_match":true,"general_wrapping_u32":true,"implementation":"compiler-optimized row-major 32x32 blocked C11","medians_ns":{"end_to_end_cold":10240797,"end_to_end_warm":10380654,"prepared_cold":9998026,"prepared_warm":10439013},"output_fnv1a64":"47e2fabfb2562325","reserved_workspace_bytes":0,"sample_counts":{"end_to_end_cold":5,"end_to_end_warm":15,"prepared_cold":5,"prepared_warm":15},"semantic_scope":"complete wrapping-u32 modulo 2^32","source":"c-run","status":"measured","threads":4,"timing_scope":"prepared dense inputs; output allocation, zeroing, execution, and cascade intermediate cleanup; final output cleanup excluded","work":{"annihilated_macs":0,"executed_macs":56623104,"logical_macs":56623104,"promoted_tiles":0,"promotion_bits":0,"q16_macs":0,"q24_macs":0,"q32_macs":0,"q8_macs":0}},{"active_lanes":4,"active_lanes_observed":true,"case":"synthetic-cancellation-cascade","compiler":"clang","configured_lanes":4,"end_to_end_scope":"dense inputs; canonical import, public GEMM execution, dense export, and input/intermediate cleanup; final dense output cleanup excluded","engine":"qqfenx-c11","exact_match":true,"general_wrapping_u32":true,"implementation":"SSE4.1","medians_ns":{"end_to_end_cold":4440699,"end_to_end_warm":4043444,"prepared_cold":4724890,"prepared_warm":3352807},"output_fnv1a64":"47e2fabfb2562325","reserved_workspace_bytes":17248,"sample_counts":{"end_to_end_cold":5,"end_to_end_warm":15,"prepared_cold":5,"prepared_warm":15},"semantic_scope":"complete wrapping-u32 modulo 2^32","source":"c-run","status":"measured","threads":4,"timing_scope":"prepared canonical inputs; public GEMM output allocation, execution, and cascade intermediate cleanup; dense export and final cleanup excluded","work":{"annihilated_macs":56623104,"executed_macs":56623104,"logical_macs":113246208,"promoted_tiles":144,"promotion_bits":1008,"q16_macs":0,"q24_macs":0,"q32_macs":0,"q8_macs":56623104}},{"active_lanes":1,"active_lanes_observed":true,"case":"synthetic-cancellation-cascade","compiler":"clang","configured_lanes":1,"end_to_end_scope":"caller-dense inputs; input copies, output allocation, zeroing, execution, and input/intermediate cleanup; final output cleanup excluded","engine":"dense-c-blocked","exact_match":true,"general_wrapping_u32":true,"implementation":"compiler-optimized row-major 32x32 blocked C11","medians_ns":{"end_to_end_cold":20575759,"end_to_end_warm":20977143,"prepared_cold":20663589,"prepared_warm":21809564},"output_fnv1a64":"47e2fabfb2562325","reserved_workspace_bytes":0,"sample_counts":{"end_to_end_cold":5,"end_to_end_warm":15,"prepared_cold":5,"prepared_warm":15},"semantic_scope":"complete wrapping-u32 modulo 2^32","source":"c-run","status":"measured","threads":4,"timing_scope":"prepared dense inputs; output allocation, zeroing, execution, and cascade intermediate cleanup; final output cleanup excluded","work":{"annihilated_macs":0,"executed_macs":113246208,"logical_macs":113246208,"promoted_tiles":0,"promotion_bits":0,"q16_macs":0,"q24_macs":0,"q32_macs":0,"q8_macs":0}},{"active_lanes":4,"active_lanes_observed":true,"case":"uci-iris-feature-gram-i32-v1","compiler":"clang","configured_lanes":4,"end_to_end_scope":"dense inputs; canonical import, public GEMM execution, dense export, and input/intermediate cleanup; final dense output cleanup excluded","engine":"qqfenx-c11","exact_match":true,"general_wrapping_u32":true,"implementation":"SSE4.1","medians_ns":{"end_to_end_cold":206607,"end_to_end_warm":87357,"prepared_cold":605742,"prepared_warm":106271},"output_fnv1a64":"21255eaa70abb088","reserved_workspace_bytes":928,"retained_output_u32":[522385,267343,348376,112814,267343,143040,167430,53189,348376,167430,258271,86911,112814,53189,86911,30233],"sample_counts":{"end_to_end_cold":5,"end_to_end_warm":15,"prepared_cold":5,"prepared_warm":15},"semantic_scope":"complete wrapping-u32 modulo 2^32","source":"c-run","status":"measured","threads":4,"timing_scope":"prepared canonical inputs; public GEMM output allocation, execution, and cascade intermediate cleanup; dense export and final cleanup excluded","work":{"annihilated_macs":0,"executed_macs":2400,"logical_macs":2400,"promoted_tiles":0,"promotion_bits":0,"q16_macs":0,"q24_macs":0,"q32_macs":2400,"q8_macs":0}},{"active_lanes":1,"active_lanes_observed":true,"case":"uci-iris-feature-gram-i32-v1","compiler":"clang","configured_lanes":1,"end_to_end_scope":"caller-dense inputs; input copies, output allocation, zeroing, execution, and input/intermediate cleanup; final output cleanup excluded","engine":"dense-c-blocked","exact_match":true,"general_wrapping_u32":true,"implementation":"compiler-optimized row-major 32x32 blocked C11","medians_ns":{"end_to_end_cold":5641,"end_to_end_warm":2145,"prepared_cold":3193,"prepared_warm":1856},"output_fnv1a64":"21255eaa70abb088","reserved_workspace_bytes":0,"retained_output_u32":[522385,267343,348376,112814,267343,143040,167430,53189,348376,167430,258271,86911,112814,53189,86911,30233],"sample_counts":{"end_to_end_cold":5,"end_to_end_warm":15,"prepared_cold":5,"prepared_warm":15},"semantic_scope":"complete wrapping-u32 modulo 2^32","source":"c-run","status":"measured","threads":4,"timing_scope":"prepared dense inputs; output allocation, zeroing, execution, and cascade intermediate cleanup; final output cleanup excluded","work":{"annihilated_macs":0,"executed_macs":2400,"logical_macs":2400,"promoted_tiles":0,"promotion_bits":0,"q16_macs":0,"q24_macs":0,"q32_macs":0,"q8_macs":0}},{"active_lanes":0,"active_lanes_observed":false,"case":"uci-iris-feature-gram-i32-v1","compiler":"clang","configured_lanes":4,"end_to_end_scope":"caller wrapping-u32 inputs; float conversion, input/output allocation, SGEMM, exact-integral validation, wrapping-u32 reconstruction, and float-input cleanup; final output cleanup excluded","engine":"openblas-sgemm","exact_match":true,"general_wrapping_u32":false,"implementation":"external CBLAS SGEMM","medians_ns":{"end_to_end_cold":11105,"end_to_end_warm":1508,"prepared_cold":17346,"prepared_warm":797},"output_fnv1a64":"21255eaa70abb088","reserved_workspace_bytes":0,"retained_output_u32":[522385,267343,348376,112814,267343,143040,167430,53189,348376,167430,258271,86911,112814,53189,86911,30233],"sample_counts":{"end_to_end_cold":5,"end_to_end_warm":15,"prepared_cold":5,"prepared_warm":15},"semantic_scope":"workload-constrained exact integer subset: nonnegative integral dot products remain below 2^24; independently reconstructed and compared to wrapping-u32 oracle","source":"c-run","status":"measured","threads":4,"timing_scope":"prepared float inputs; output allocation, SGEMM, exact-integral validation, and wrapping-u32 reconstruction; cleanup excluded","work":{"annihilated_macs":0,"executed_macs":2400,"logical_macs":2400,"promoted_tiles":0,"promotion_bits":0,"q16_macs":0,"q24_macs":0,"q32_macs":0,"q8_macs":0}}],"onednn_engines":[{"case":"synthetic-q8-safe-u6","compiler":"oneDNN","end_to_end_scope":"initialized oneDNN engine/stream and caller-owned inputs; input copies, descriptors, primitive, memory objects, output allocation, execute, wait, and quotient reconstruction; oracle comparison and engine/stream initialization excluded","engine":"onednn-matmul-u8s8s32","exact_match":true,"general_wrapping_u32":false,"implementation":null,"medians_ns":{"end_to_end_cold":5813636,"end_to_end_warm":5295354,"prepared_cold":5069976,"prepared_warm":4718543},"output_fnv1a64":"cc47066bd39dcf86","sample_counts":{"end_to_end_cold":5,"end_to_end_warm":15,"prepared_cold":5,"prepared_warm":15},"semantic_scope":"workload-constrained exact nonnegative integer subset with no s16 pair saturation and no s32 dot overflow; output independently reduced and lifted when required","source":"onednn-run","status":"measured","thread_control":{"dynamic_enabled":false,"mechanism":"openmp-api-team-observation","observed_max_threads":1,"observed_team_threads":1,"requested_threads":1,"runtime":"openmp","verified":true},"threading_runtime":"openmp","threads":1,"timing_scope":"prepared primitive, descriptors, input/output memory, and caller-owned inputs; matmul execute, stream wait, and quotient reconstruction; oracle comparison excluded","work":{"annihilated_macs":0,"executed_macs":56623104,"logical_macs":56623104}},{"case":"uci-iris-feature-gram-i32-v1","compiler":"oneDNN","end_to_end_scope":"initialized oneDNN engine/stream and caller-owned inputs; input copies, descriptors, primitive, memory objects, output allocation, execute, wait, and quotient reconstruction; oracle comparison and engine/stream initialization excluded","engine":"onednn-matmul-u8s8s32","exact_match":true,"general_wrapping_u32":false,"implementation":null,"medians_ns":{"end_to_end_cold":52262,"end_to_end_warm":9305,"prepared_cold":25877,"prepared_warm":2930},"output_fnv1a64":"21255eaa70abb088","retained_output_u32":[522385,267343,348376,112814,267343,143040,167430,53189,348376,167430,258271,86911,112814,53189,86911,30233],"sample_counts":{"end_to_end_cold":5,"end_to_end_warm":15,"prepared_cold":5,"prepared_warm":15},"semantic_scope":"workload-constrained exact nonnegative integer subset with no s16 pair saturation and no s32 dot overflow; output independently reduced and lifted when required","source":"onednn-run","status":"measured","thread_control":{"dynamic_enabled":false,"mechanism":"openmp-api-team-observation","observed_max_threads":1,"observed_team_threads":1,"requested_threads":1,"runtime":"openmp","verified":true},"threading_runtime":"openmp","threads":1,"timing_scope":"prepared primitive, descriptors, input/output memory, and caller-owned inputs; matmul execute, stream wait, and quotient reconstruction; oracle comparison excluded","work":{"annihilated_macs":0,"executed_macs":2400,"logical_macs":2400}},{"case":"synthetic-q8-safe-u6","compiler":"oneDNN","end_to_end_scope":"initialized oneDNN engine/stream and caller-owned inputs; input copies, descriptors, primitive, memory objects, output allocation, execute, wait, and quotient reconstruction; oracle comparison and engine/stream initialization excluded","engine":"onednn-matmul-u8s8s32","exact_match":true,"general_wrapping_u32":false,"implementation":null,"medians_ns":{"end_to_end_cold":3847786,"end_to_end_warm":1961108,"prepared_cold":1657406,"prepared_warm":1572582},"output_fnv1a64":"cc47066bd39dcf86","sample_counts":{"end_to_end_cold":5,"end_to_end_warm":15,"prepared_cold":5,"prepared_warm":15},"semantic_scope":"workload-constrained exact nonnegative integer subset with no s16 pair saturation and no s32 dot overflow; output independently reduced and lifted when required","source":"onednn-run","status":"measured","thread_control":{"dynamic_enabled":false,"mechanism":"openmp-api-team-observation","observed_max_threads":4,"observed_team_threads":4,"requested_threads":4,"runtime":"openmp","verified":true},"threading_runtime":"openmp","threads":4,"timing_scope":"prepared primitive, descriptors, input/output memory, and caller-owned inputs; matmul execute, stream wait, and quotient reconstruction; oracle comparison excluded","work":{"annihilated_macs":0,"executed_macs":56623104,"logical_macs":56623104}},{"case":"uci-iris-feature-gram-i32-v1","compiler":"oneDNN","end_to_end_scope":"initialized oneDNN engine/stream and caller-owned inputs; input copies, descriptors, primitive, memory objects, output allocation, execute, wait, and quotient reconstruction; oracle comparison and engine/stream initialization excluded","engine":"onednn-matmul-u8s8s32","exact_match":true,"general_wrapping_u32":false,"implementation":null,"medians_ns":{"end_to_end_cold":1071344,"end_to_end_warm":14807,"prepared_cold":2308430,"prepared_warm":5923},"output_fnv1a64":"21255eaa70abb088","retained_output_u32":[522385,267343,348376,112814,267343,143040,167430,53189,348376,167430,258271,86911,112814,53189,86911,30233],"sample_counts":{"end_to_end_cold":5,"end_to_end_warm":15,"prepared_cold":5,"prepared_warm":15},"semantic_scope":"workload-constrained exact nonnegative integer subset with no s16 pair saturation and no s32 dot overflow; output independently reduced and lifted when required","source":"onednn-run","status":"measured","thread_control":{"dynamic_enabled":false,"mechanism":"openmp-api-team-observation","observed_max_threads":4,"observed_team_threads":4,"requested_threads":4,"runtime":"openmp","verified":true},"threading_runtime":"openmp","threads":4,"timing_scope":"prepared primitive, descriptors, input/output memory, and caller-owned inputs; matmul execute, stream wait, and quotient reconstruction; oracle comparison excluded","work":{"annihilated_macs":0,"executed_macs":2400,"logical_macs":2400}}]},"methodology":{"arithmetic":"wrapping-u32 modulo 2^32","clock":"CLOCK_MONOTONIC_RAW for C; std::chrono::steady_clock for oneDNN","cold_definition":"touch the configured cache-scrub buffer at 64-byte intervals immediately before each retained cold sample","execution_order":"compiler-major then configured-lane order for native/dense/OpenBLAS; configured-lane order for oneDNN; every declared case is retained","host_load_policy":"timestamped load observations bracket the competitive measurement window; results are retained without favorable-sample filtering","iris_external_exactness":"X10 values are nonnegative integers <=79 and every dot sum is below 2^24; SGEMM and u8/s8/s32 outputs are exact integers and are compared with the frozen 4x4 oracle","onednn_saturation_control":"two maximum Q8 pair products are 2*63*63=7938 and two maximum Iris pair products are at most 2*79*79=12482, both below s16 saturation; complete dot sums are below s32 overflow","onednn_thread_policy":"oneDNN scaling and QQfenx/oneDNN ratios are published only when the OpenMP runtime is controlled through its API and a separate OpenMP probe region in the comparator process observes the requested team size with dynamic teams disabled; TBB, threadpool, sequential, unknown, or unobserved OpenMP runs retain timings but withhold ratios","publication_policy":"retain every successful sample and every measured compiler/thread variant; abort rather than omit an exactness failure","q8_openblas_exactness":"payloads are in [0,63], every product is exactly representable in f32, every nonnegative dot sum is below 2^24, and reconstructed wrapping-u32 words are compared element-by-element","ratio_direction":"baseline_over_candidate greater than 1 means QQfenx was faster for that retained host measurement","scope_warning":"OpenBLAS and oneDNN comparisons cover only the proven exact-compatible domains; neither library is represented as a general wrapping-u32 implementation","warm_definition":"one unretained warmup followed by consecutive retained samples without explicit cache scrubbing"},"packed_kernel_proof":{"backend":"SSE4.1","binary_sha256":"b4c86c18e041aeed84ba3a03b98a42b18e476f320c940438593813a0c5e030b6","compiler":"gcc","compiler_version":"gcc (Debian 15.2.0-17) 15.2.0","q16":{"packed_avx2_q16_calls":0,"packed_avx2_q8_calls":0,"packed_sse41_q16_calls":48,"packed_sse41_q8_calls":0},"q8":{"packed_avx2_q16_calls":0,"packed_avx2_q8_calls":0,"packed_sse41_q16_calls":0,"packed_sse41_q8_calls":48},"status":"pass","telemetry_instrumented":true,"timings_used_as_competitive_results":false},"product":{"name":"QQfenx","release":"v0.1.4","vendor":"MFENX"},"report_binding":{"canonical_payload_algorithm":"sha256","canonical_payload_sha256":"29f4f57604621d7599420aa24b4143c226db73a8029340807854dd4b04b37cbe","file_name":"i5-2500s-extent384-20260826.json","file_sha256":"895e903045e071358066194511600f4054c192a3ad004eb65b0f1a145443cf24","generated_at_utc":"2026-08-26T08:26:46.883285+00:00","schema":"mfenx.qqfenx.competitive-evidence.v1","source_commit":"01e2c6e552e96a3e0e67461827c06f7c83da061a","source_dirty":false,"source_file_sha256":{"benchmarks/qqfenx-competitive/benchmark.schema.json":"178bdf01aa5b1e94dd36a177e1e4d2c86768cccf4f6eec7a6bcba041c925c831","benchmarks/qqfenx-competitive/onednn_q8.cpp":"a2afd879ccf97900da31f2f46eb438214375d8f582f591333c3ae6ffcac4a5d7","benchmarks/qqfenx-competitive/qqfenx_competitive.c":"92dd5da4c43ae1fcd41a21d2baa6b3c1c75e4b46bbd581cd744875cf16f058ae","benchmarks/qqfenx-competitive/run.py":"570a98bb1e80351c61dc414547fff23f4b8b52b5bcf46437a221f76361ccf922","native/qqfenx/qqfenx.c":"bc0eebf668e3b67b636f3cf89975046dbf6a5a31458a5e63b74c323f17669897","native/qqfenx/qqfenx.h":"0e92a9601bc7484a027f39b7f153e1e425a1f1b6f0381549bc74d8ee7f00e87e","native/qqfenx/qqfenx_scaling_bench.c":"83b3e622c16e09862f700436381d4d660800e8358c74bc5cdd99d3de6b584272","packaging/external-workloads/uci-iris/adapter.py":"9bef412bdb894bb292cf295566c234889445f890a19a1dad56b24ca4b7805a97","packaging/external-workloads/uci-iris/bezdekIris.data":"0fed2a99db77ec533a62dc66894d3ec6df3b58b6a8f3cf4a6b47e4086b7f97dc","packaging/external-workloads/uci-iris/oracle.json":"bf171968b1883ecc3f54d218d132711cacc902ac288c1da9041a82516159bb13"},"source_tree":"4441b4e2aa7414cf56a09c37ccf080cc4102ba79","status":"pass"},"scaling":{"onednn_1_to_4_threads":[{"baseline_threading_runtime":"openmp","baseline_threads":1,"case":"synthetic-q8-safe-u6","end_to_end_warm_speedup":2.700185,"prepared_warm_speedup":3.000507,"speedup_published":true,"thread_control_verified":true,"threading_runtime":"openmp","threads":4},{"baseline_threading_runtime":"openmp","baseline_threads":1,"case":"uci-iris-feature-gram-i32-v1","end_to_end_warm_speedup":0.628419,"prepared_warm_speedup":0.494682,"speedup_published":true,"thread_control_verified":true,"threading_runtime":"openmp","threads":4}],"openblas_1_to_4_lanes":[{"active_lanes":0,"baseline_configured_lanes":1,"case":"synthetic-q8-safe-u6","compiler":"clang","configured_lanes":4,"end_to_end_warm_speedup":1.515359,"engine":"openblas-sgemm","prepared_warm_speedup":2.180842},{"active_lanes":0,"baseline_configured_lanes":1,"case":"uci-iris-feature-gram-i32-v1","compiler":"clang","configured_lanes":4,"end_to_end_warm_speedup":0.996684,"engine":"openblas-sgemm","prepared_warm_speedup":0.967378},{"active_lanes":0,"baseline_configured_lanes":1,"case":"synthetic-q8-safe-u6","compiler":"gcc","configured_lanes":4,"end_to_end_warm_speedup":1.193932,"engine":"openblas-sgemm","prepared_warm_speedup":2.173292},{"active_lanes":0,"baseline_configured_lanes":1,"case":"uci-iris-feature-gram-i32-v1","compiler":"gcc","configured_lanes":4,"end_to_end_warm_speedup":0.889736,"engine":"openblas-sgemm","prepared_warm_speedup":1.171053}],"qqfenx_1_to_4_lanes":[{"active_lanes":4,"baseline_configured_lanes":1,"case":"synthetic-q32","compiler":"clang","configured_lanes":4,"end_to_end_warm_speedup":2.036397,"engine":"qqfenx-c11","prepared_warm_speedup":2.199527},{"active_lanes":4,"baseline_configured_lanes":1,"case":"synthetic-q16","compiler":"clang","configured_lanes":4,"end_to_end_warm_speedup":2.314201,"engine":"qqfenx-c11","prepared_warm_speedup":1.806117},{"active_lanes":4,"baseline_configured_lanes":1,"case":"synthetic-q8-safe-u6","compiler":"clang","configured_lanes":4,"end_to_end_warm_speedup":2.293246,"engine":"qqfenx-c11","prepared_warm_speedup":2.776124},{"active_lanes":4,"baseline_configured_lanes":1,"case":"synthetic-zero-ideal","compiler":"clang","configured_lanes":4,"end_to_end_warm_speedup":0.926018,"engine":"qqfenx-c11","prepared_warm_speedup":0.803528},{"active_lanes":4,"baseline_configured_lanes":1,"case":"synthetic-cancellation-cascade","compiler":"clang","configured_lanes":4,"end_to_end_warm_speedup":2.271775,"engine":"qqfenx-c11","prepared_warm_speedup":2.474551},{"active_lanes":4,"baseline_configured_lanes":1,"case":"uci-iris-feature-gram-i32-v1","compiler":"clang","configured_lanes":4,"end_to_end_warm_speedup":0.375539,"engine":"qqfenx-c11","prepared_warm_speedup":0.170799},{"active_lanes":4,"baseline_configured_lanes":1,"case":"synthetic-q32","compiler":"gcc","configured_lanes":4,"end_to_end_warm_speedup":2.375636,"engine":"qqfenx-c11","prepared_warm_speedup":1.999132},{"active_lanes":4,"baseline_configured_lanes":1,"case":"synthetic-q16","compiler":"gcc","configured_lanes":4,"end_to_end_warm_speedup":1.965858,"engine":"qqfenx-c11","prepared_warm_speedup":2.882171},{"active_lanes":4,"baseline_configured_lanes":1,"case":"synthetic-q8-safe-u6","compiler":"gcc","configured_lanes":4,"end_to_end_warm_speedup":2.15435,"engine":"qqfenx-c11","prepared_warm_speedup":2.088416},{"active_lanes":4,"baseline_configured_lanes":1,"case":"synthetic-zero-ideal","compiler":"gcc","configured_lanes":4,"end_to_end_warm_speedup":1.173343,"engine":"qqfenx-c11","prepared_warm_speedup":0.735889},{"active_lanes":4,"baseline_configured_lanes":1,"case":"synthetic-cancellation-cascade","compiler":"gcc","configured_lanes":4,"end_to_end_warm_speedup":2.420385,"engine":"qqfenx-c11","prepared_warm_speedup":2.620801},{"active_lanes":4,"baseline_configured_lanes":1,"case":"uci-iris-feature-gram-i32-v1","compiler":"gcc","configured_lanes":4,"end_to_end_warm_speedup":0.321017,"engine":"qqfenx-c11","prepared_warm_speedup":0.356547}]},"schema":"mfenx.qqfenx.performance-certificate.v1","status":"pass"}