Skip to content

Zephyr — heap mode

Raw cross-binding benchmark results measured on STM32F746G-DISCOVERY (Cortex-M7 @ 216 MHz) running Zephyr RTOS in default heap-allocation mode (_create() / _destroy() API), with CONFIG_OVE_BENCHMARK_WORST_CASE_TIMING=y — caches and flash accelerators disabled to approximate cacheless ARM-MCU timing. See benchmarks overview for the worst-case-timing methodology.

Methodology and reproduction steps: benchmarks overview. Same numbers under CONFIG_OVE_ZERO_HEAP=y: Zephyr zero-heap. Interpretation: heap vs zero-heap, per-binding analysis, wrapper-vs-native notes.

Generated by scripts/bench_compare.py. Trimmed-mean (top 1% dropped) when available, else avg. Delta column is (binding − C) / C — values within ±10.0% are within typical measurement noise.

native_* rows. bench_native_zephyr.c is C code calling raw Zephyr APIs (k_mutex_*, k_sem_*, k_msgq_*), compiled identically into every binary; the CPP/RUST/ZIG columns for those rows are the same C code measured in three different processes.

Zephyr

Suite Case C CPP Δ CPP RUST Δ RUST ZIG Δ ZIG
time time_get_us_overhead 2.5 µs 2.4 µs -4.7% 2.6 µs +6.4% 2.5 µs +2.6%
time delay_1ms 1.09 ms 1.09 ms -0.0% 1.09 ms -0.2% 1.09 ms -0.1%
thread create_destroy 116.1 µs 113.5 µs -2.2% 115.0 µs -0.9% 114.0 µs -1.8%
thread yield 10.6 µs 10.6 µs -0.3% 10.6 µs -0.6% 10.4 µs -2.3%
thread get_self 971 ns 975 ns +0.4% 1.1 µs +12.8% 842 ns -13.3%
thread sleep_1ms 1.09 ms 1.09 ms -0.0% 1.09 ms -0.2% 1.09 ms -0.1%
thread context_switch 56.5 µs 54.6 µs -3.4% — — 54.2 µs -4.0%
sync mutex_memory 32 B 32 B — 32 B — 32 B —
sync sem_memory 24 B 24 B — 24 B — 24 B —
sync event_memory 24 B 24 B — 24 B — 24 B —
sync condvar_memory 16 B 16 B — 16 B — 16 B —
sync mutex_lock_unlock 3.0 µs 3.1 µs +3.6% 3.4 µs +15.7% 2.8 µs -6.0%
sync mutex_create_destroy 11.6 µs 12.0 µs +4.0% 8.7 µs -25.0% 7.9 µs -31.4%
sync mutex_contention_2t 3.1 µs 47.7 µs +1448.5% 38.9 µs +1163.2% 33.9 µs +1000.3%
sync sem_take_give 2.0 µs 2.2 µs +5.2% 2.1 µs +4.1% 2.1 µs +2.2%
sync sem_create_destroy 12.0 µs 12.3 µs +2.7% 12.6 µs +5.4% 11.9 µs -0.7%
sync event_signal_wait 56.6 µs 54.7 µs -3.3% 55.6 µs -1.8% 54.5 µs -3.6%
sync condvar_signal_wait 65.1 µs 63.9 µs -1.8% 64.6 µs -0.7% 63.3 µs -2.8%
sync recursive_mutex_lock_unlock 3.1 µs 3.1 µs +2.4% 3.4 µs +9.9% 2.9 µs -4.3%
queue memory 96 B 96 B — 96 B — 96 B —
queue send_receive 4.7 µs 4.3 µs -9.0% 6.0 µs +26.6% 4.4 µs -6.9%
queue create_destroy 20.0 µs 20.7 µs +3.7% 21.4 µs +7.1% 21.3 µs +6.7%
queue throughput_2t 3.7 µs 3.2 µs -14.1% 4.1 µs +11.4% 3.1 µs -14.6%
timer memory 96 B 96 B — 96 B — 96 B —
timer create_destroy 14.1 µs 14.5 µs +2.7% 15.4 µs +8.6% 14.0 µs -1.1%
timer start_stop 9.0 µs 8.7 µs -3.7% 8.8 µs -2.5% 8.7 µs -3.5%
eventgroup memory 24 B 24 B — 24 B — 24 B —
eventgroup set_get_bits 9.0 µs 9.2 µs +1.3% 9.1 µs +0.9% 8.9 µs -1.0%
eventgroup create_destroy 11.7 µs 11.9 µs +2.3% 12.3 µs +5.4% 11.6 µs -0.8%
workqueue memory 2520 B 2520 B — 2520 B — 2520 B —
workqueue create_destroy 62.2 µs 61.8 µs -0.6% 62.5 µs +0.6% 63.2 µs +1.6%
workqueue submit_execute 58.4 µs 57.6 µs -1.4% 57.6 µs -1.3% 57.2 µs -2.0%
stream memory 320 B 328 B — 320 B — 320 B —
stream send_recv_64B 14.3 µs 14.3 µs +0.3% 17.3 µs +21.2% 14.2 µs -0.9%
stream create_destroy 11.7 µs 12.2 µs +4.0% 12.5 µs +7.2% 11.7 µs +0.1%
stream throughput 24.3 µs 24.0 µs -1.4% 41.7 µs +71.3% 23.8 µs -2.0%
native_zephyr native_mutex_lock_unlock 2.8 µs 2.7 µs -3.3% 2.8 µs +2.5% 2.6 µs -6.2%
native_zephyr native_mutex_create_destroy 813 ns 911 ns +12.1% 799 ns -1.7% 776 ns -4.6%
native_zephyr native_mutex_contention_2t 2.9 µs 2.7 µs -6.2% 2.9 µs -0.6% 2.7 µs -7.5%
native_zephyr native_recursive_mutex_lock_unlock 2.8 µs 2.7 µs -1.6% 2.9 µs +4.7% 2.7 µs -4.2%
native_zephyr native_sem_take_give 1.7 µs 1.6 µs -10.1% 1.6 µs -8.1% 1.7 µs -4.5%
native_zephyr native_sem_create_destroy 749 ns 753 ns +0.5% 824 ns +10.0% 648 ns -13.5%
native_zephyr native_condvar_signal_wait 66.9 µs 66.1 µs -1.2% 66.1 µs -1.2% 65.6 µs -1.8%
native_zephyr native_event_signal_wait 63.2 µs 62.6 µs -1.0% 62.4 µs -1.3% 61.9 µs -2.0%
native_zephyr native_thread_yield 10.5 µs 10.5 µs -0.1% 10.4 µs -0.7% 10.3 µs -2.0%
native_zephyr native_thread_sleep_1ms 1.09 ms 1.09 ms -0.0% 1.09 ms -0.2% 1.09 ms -0.1%
native_zephyr native_thread_create_destroy 80.0 µs 77.9 µs -2.5% 78.2 µs -2.2% 77.5 µs -3.1%
native_zephyr native_thread_context_switch 55.9 µs 53.9 µs -3.6% 54.4 µs -2.7% 53.4 µs -4.5%
native_zephyr native_queue_send_receive 4.8 µs 4.0 µs -15.8% 5.9 µs +24.5% 4.1 µs -13.1%
native_zephyr native_queue_create_destroy 1.3 µs 1.1 µs -18.2% 1.3 µs -4.2% 1.1 µs -13.4%
native_zephyr native_stream_send_recv_64B 12.8 µs 12.8 µs -0.1% 11.0 µs -13.8% 12.7 µs -0.5%
thread ctx_switch — — — 55.6 µs — — —

Cases with |Δ| > 10.0% vs C:

  • CPP sync/mutex_contention_2t 47726 vs 3082 (+1448.5%)
  • RUST sync/mutex_contention_2t 38932 vs 3082 (+1163.2%)
  • ZIG sync/mutex_contention_2t 33912 vs 3082 (+1000.3%)
  • RUST stream/throughput 41669 vs 24320 (+71.3%)
  • ZIG sync/mutex_create_destroy 7924 vs 11554 (-31.4%)
  • RUST queue/send_receive 5993 vs 4733 (+26.6%)
  • RUST sync/mutex_create_destroy 8661 vs 11554 (-25.0%)
  • RUST native_zephyr/native_queue_send_receive 5932 vs 4766 (+24.5%)
  • RUST stream/send_recv_64B 17324 vs 14295 (+21.2%)
  • CPP native_zephyr/native_queue_create_destroy 1074 vs 1313 (-18.2%)
  • CPP native_zephyr/native_queue_send_receive 4013 vs 4766 (-15.8%)
  • RUST sync/mutex_lock_unlock 3415 vs 2952 (+15.7%)
  • ZIG queue/throughput_2t 3133 vs 3669 (-14.6%)
  • CPP queue/throughput_2t 3152 vs 3669 (-14.1%)
  • RUST native_zephyr/native_stream_send_recv_64B 11018 vs 12775 (-13.8%)
  • ZIG native_zephyr/native_sem_create_destroy 648 vs 749 (-13.5%)
  • ZIG native_zephyr/native_queue_create_destroy 1137 vs 1313 (-13.4%)
  • ZIG thread/get_self 842 vs 971 (-13.3%)
  • ZIG native_zephyr/native_queue_send_receive 4142 vs 4766 (-13.1%)
  • RUST thread/get_self 1095 vs 971 (+12.8%)
  • CPP native_zephyr/native_mutex_create_destroy 911 vs 813 (+12.1%)
  • RUST queue/throughput_2t 4089 vs 3669 (+11.4%)
  • CPP native_zephyr/native_sem_take_give 1564 vs 1739 (-10.1%)
  • RUST native_zephyr/native_sem_create_destroy 824 vs 749 (+10.0%)

Wrapper vs native Zephyr API (within-run delta)

Each row pairs one binding's wrapper measurement against the raw Zephyr API baseline measured in the same process. See wrapper-vs-native notes for IPC caveats, lifecycle/intrinsic-cost interpretation, and notes on cross-process baseline variance.

Operation Binding Wrapper called Wrapper ns Native ns Δ
Thread yield C ove_thread_yield 10625 ns 10507 ns +118 ns
Thread yield CPP ove::Thread::yield 10590 ns 10493 ns +97 ns
Thread yield RUST ove::Thread::yield 10562 ns 10438 ns +124 ns
Thread yield ZIG ove.Thread.yield 10383 ns 10300 ns +83 ns
Thread sleep 1ms C ove_thread_sleep_ms(1) 1087786 ns 1087767 ns +19 ns
Thread sleep 1ms CPP ove::Thread::sleep_ms(1) 1087405 ns 1087465 ns -60 ns
Thread sleep 1ms RUST ove::Thread::sleep_ms(1) 1085182 ns 1085182 ns +0 ns
Thread sleep 1ms ZIG ove.Thread.sleepMs(1) 1087211 ns 1087120 ns +91 ns
Thread create+destroy C ove_thread_create+destroy 116073 ns 79956 ns +36117 ns
Thread create+destroy CPP ove::Thread (ctor+dtor) 113501 ns 77933 ns +35568 ns
Thread create+destroy RUST ove::Thread::spawn+join 115000 ns 78193 ns +36807 ns
Thread create+destroy ZIG ove.Thread.spawn+join 113986 ns 77475 ns +36511 ns
Thread context_switch (2t) C ove ping-pong (2t) 56502 ns 55859 ns +643 ns
Thread context_switch (2t) CPP ove ping-pong (2t) 54585 ns 53868 ns +717 ns
Thread context_switch (2t) ZIG ove ping-pong (2t) 54220 ns 53354 ns +866 ns
Mutex lock+unlock C ove_mutex_lock+unlock 2952 ns 2753 ns +199 ns
Mutex lock+unlock CPP ove::Mutex::lock+unlock 3059 ns 2661 ns +398 ns
Mutex lock+unlock RUST ove::Mutex::lock+unlock 3415 ns 2823 ns +592 ns
Mutex lock+unlock ZIG ove.Mutex.lock+unlock 2774 ns 2582 ns +192 ns
Mutex create+destroy C ove_mutex_create+destroy 11554 ns 813 ns +10741 ns
Mutex create+destroy CPP ove::Mutex (ctor+dtor) 12017 ns 911 ns +11106 ns
Mutex create+destroy RUST ove::Mutex (new+drop) 8661 ns 799 ns +7862 ns
Mutex create+destroy ZIG ove.Mutex.create+destroy 7924 ns 776 ns +7148 ns
Mutex contention (2t) C ove_mutex_lock+unlock (×2t) 3082 ns 2897 ns +185 ns
Mutex contention (2t) CPP ove::Mutex::lock+unlock (×2t) 47726 ns 2716 ns +45010 ns
Mutex contention (2t) RUST ove::Mutex::lock+unlock (×2t) 38932 ns 2881 ns +36051 ns
Mutex contention (2t) ZIG ove.Mutex.lock+unlock (×2t) 33912 ns 2679 ns +31233 ns
Recursive mutex lock+unlock C ove_rmtx_lock+unlock 3051 ns 2790 ns +261 ns
Recursive mutex lock+unlock CPP ove::RMutex::lock+unlock 3124 ns 2744 ns +380 ns
Recursive mutex lock+unlock RUST ove::RMutex::lock+unlock 3353 ns 2922 ns +431 ns
Recursive mutex lock+unlock ZIG ove.RMutex.lock+unlock 2920 ns 2674 ns +246 ns
Sem take+give C ove_sem_take+give 2045 ns 1739 ns +306 ns
Sem take+give CPP ove::Sem::take+give 2151 ns 1564 ns +587 ns
Sem take+give RUST ove::Sem::take+give 2128 ns 1599 ns +529 ns
Sem take+give ZIG ove.Sem.take+give 2091 ns 1661 ns +430 ns
Sem create+destroy C ove_sem_create+destroy 11966 ns 749 ns +11217 ns
Sem create+destroy CPP ove::Sem (ctor+dtor) 12290 ns 753 ns +11537 ns
Sem create+destroy RUST ove::Sem (new+drop) 12607 ns 824 ns +11783 ns
Sem create+destroy ZIG ove.Sem.create+destroy 11883 ns 648 ns +11235 ns
Condvar signal+wait C ove_condvar_signal+wait 65072 ns 66868 ns -1796 ns
Condvar signal+wait CPP ove::Condvar::signal+wait 63871 ns 66068 ns -2197 ns
Condvar signal+wait RUST ove::Condvar::signal+wait 64594 ns 66094 ns -1500 ns
Condvar signal+wait ZIG ove.Condvar.signal+wait 63280 ns 65637 ns -2357 ns
Event signal+wait C ove_event_signal+wait 56605 ns 63236 ns -6631 ns
Event signal+wait CPP ove::Event::signal+wait 54739 ns 62632 ns -7893 ns
Event signal+wait RUST ove::Event::signal+wait 55586 ns 62407 ns -6821 ns
Event signal+wait ZIG ove.Event.signal+wait 54540 ns 61944 ns -7404 ns
Queue send+receive C ove_queue_send+receive 4733 ns 4766 ns -33 ns
Queue send+receive CPP ove::Queue::send+recv 4308 ns 4013 ns +295 ns
Queue send+receive RUST ove::Queue::send+recv 5993 ns 5932 ns +61 ns
Queue send+receive ZIG ove.Queue.send+recv 4407 ns 4142 ns +265 ns
Queue create+destroy C ove_queue_create+destroy 19975 ns 1313 ns +18662 ns
Queue create+destroy CPP ove::Queue (ctor+dtor) 20712 ns 1074 ns +19638 ns
Queue create+destroy RUST ove::Queue (new+drop) 21392 ns 1258 ns +20134 ns
Queue create+destroy ZIG ove.Queue.create+destroy 21309 ns 1137 ns +20172 ns
Stream send+recv 64B C ove_stream_send+recv 64B 14295 ns 12775 ns +1520 ns
Stream send+recv 64B CPP ove::Stream::send+recv 64B 14345 ns 12762 ns +1583 ns
Stream send+recv 64B RUST ove::Stream::send+recv 64B 17324 ns 11018 ns +6306 ns
Stream send+recv 64B ZIG ove.Stream.send+recv 64B 14163 ns 12712 ns +1451 ns