Advance navigation bit clocks across missing measurements - #235
Open
zsoerenm wants to merge 6 commits into
Open
Conversation
…all one The bit-completion test was `count == blocks_per_bit`. A record whose length does not divide the symbol steps the count *past* the boundary, and from then on equality never holds again: that satellite stops producing navigation bits for the rest of its lock while still reporting bit sync as found. Measured on a hardware-correlator receiver (GNSSReceiver.jl#107) folding one 2 ms record per processing chunk: 90 s of unbroken lock, bit sync found at t = 25 s, one soft bit. The same hazard reaches the sample-driven path through `set_preferred_num_code_blocks_to_integrate!` and any producer whose record stream has holes. `_complete_bits!` drains while `>=` and carries the remainder, so a record longer than a whole symbol yields every bit it spans. Identical for every history that did not already stall. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
A producer that knows signal time elapsed but has nothing to report had no way to say so. `append_correlator_output!` means "here is a measurement", so the only way to keep a satellite's navigation bit boundary across a hole in a hardware correlator's record stream was to append an all-zero `CorrelatorOutput` spanning it and hope the consumer saw through the fake. It does not: every discriminator computes 0/0 on a zero correlator, and the NaN leaves the loop filter as a NaN Doppler. That is the actual source of the zero-energy correlators the discriminator guards were added for — nothing in a device produces one. The guards stay as defence for third-party producers, but they were treating a symptom, and a quiet one: with them a gap no longer crashes yet still feeds a zero-*error* measurement to the loop filter, a zero prompt to the prompt filter and a zero-power record to the C/N₀ average. `advance_bit_clock!` sits beside `append_correlator_output!`, is addressed the same way, and moves nothing but the bit buffer. Pre-sync it shifts the missing blocks through the hard-decision search window and advances its length, so both the sliding-window search and the CFAR phase index stay on the true block grid; post-sync it credits them through the same drain `buffer` uses, so a gap longer than a symbol yields every bit it spans. A secondary-coded signal's overlay phase rolls by the gap rather than going stale. Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Codecov Report✅ All modified and coverable lines are covered by tests. Additional details and impacted files@@ Coverage Diff @@
## master #235 +/- ##
==========================================
- Coverage 98.32% 98.18% -0.14%
==========================================
Files 47 47
Lines 3990 4016 +26
==========================================
+ Hits 3923 3943 +20
- Misses 67 73 +6 ☔ View full report in Codecov by Harness. 🚀 New features to boost your workflow:
|
Contributor
Benchmark Results (minimum time) — ubuntu-latestReporting the minimum over all samples (robust to shared-runner contention), not the median. Alternative backends vs Float32 (
|
| Scenario | F32 | I16 | 1b | 2b | ×I16 | ×1b | ×2b |
|---|---|---|---|---|---|---|---|
| GPS L1CA, 1 sat, 3 signals @ 5 MHz | 10.7 μs | 7.41 μs | 6.7 μs | 9.13 μs | 1.44 ✅ | 1.6 ✅ | 1.17 ✅ |
| GPS L1CA, 1 sat, 4 antennas @ 5 MHz | 13.2 μs | 7.28 μs | 6.56 μs | 11.9 μs | 1.81 ✅ | 2.01 ✅ | 1.11 ✅ |
| GPS L1CA, 8 sats @ 40 MHz | 200.0 μs | 80.7 μs | 47.8 μs | 77.0 μs | 2.47 ✅ | 4.17 ✅ | 2.59 ✅ |
| GPS L1CA, 8 sats @ 5 MHz, 100 ms buffer | 3.06 ms | 1.7 ms | 1.24 ms | 1.68 ms | 1.8 ✅ | 2.46 ✅ | 1.82 ✅ |
| GPS L1CA, 8 sats @ 5 MHz | 33.6 μs | 20.4 μs | 15.9 μs | 21.4 μs | 1.65 ✅ | 2.11 ✅ | 1.57 ✅ |
| Galileo E1B, 4 sats @ 25 MHz | 90.0 μs | 42.7 μs | 2.11 ✅ | ||||
| dynamic taps @ 5 MHz (kernel) | 5.79 μs | 3.38 μs | 2.05 μs | 2.79 μs | 1.71 ✅ | 2.83 ✅ | 2.07 ✅ |
Time benchmarks (base vs PR head)
Ratio = 7a55f91… / d71924b…: >1 means the PR is faster. ✅ ≥ 5 % faster,
| 7a55f91… | d71924b… | 7a55f91… / d71924b… | |
|---|---|---|---|
| downconvert and correlate/CPU/Float32 | 2.4 μs | 2.4 μs | 1.0 |
| downconvert and correlate/CPU/Float32 4ant | 4.78 μs | 4.79 μs | 0.997 |
| downconvert and correlate/CPU/Float64 | 2.8 μs | 2.8 μs | 1.0 |
| downconvert and correlate/CPU/Int16 | 2.48 μs | 2.48 μs | 1.0 |
| downconvert and correlate/CPU/Int16 4ant | 5.23 μs | 5.23 μs | 1.0 |
| downconvert and correlate/CPU/Int32 | 2.45 μs | 2.55 μs | 0.961 |
| fused kernel/1-ant dynamic taps | 2.16 μs | 2.16 μs | 1.0 |
| fused kernel/1-ant static taps | 1.88 μs | 1.88 μs | 1.0 |
| fused kernel/4-ant dynamic taps | 5.34 μs | 5.33 μs | 1.0 |
| fused kernel/4-ant static taps | 4.14 μs | 4.16 μs | 0.996 |
| fused tuple kernel/1-ant N=2 | 2.35 μs | 2.37 μs | 0.992 |
| fused tuple kernel/1-ant N=3 | 2.81 μs | 2.82 μs | 1.0 |
| fused tuple kernel/2-ant N=2 | 3.62 μs | 3.6 μs | 1.01 |
| fused tuple kernel/2-ant N=3 | 4.54 μs | 4.56 μs | 0.995 |
| fused tuple kernel/4-ant N=2 | 6.07 μs | 6.03 μs | 1.01 |
| fused tuple kernel/4-ant N=3 | 9.77 μs | 9.75 μs | 1.0 |
| noise estimation/append_noise_observation! – steady state | 30.8 ns | 32.9 ns | 0.934 |
| noise estimation/track! – GPS L1CA, 1 sat, 2K @ 5 MHz with a noise reference | 5.2 μs | 5.2 μs | 0.999 |
| noise estimation/update_noise! – 1 ms @ 20 MHz | 19.3 μs | 19.2 μs | 1.0 |
| noise estimation/update_noise! – 1 ms @ 4 MHz | 4.34 μs | 4.42 μs | 0.982 |
| noise estimation/update_noise! – 1 ms @ 4 MHz, 4 antennas | 8.86 μs | 8.85 μs | 1.0 |
| track/GPS L1CA + Galileo E1B, 8+8 sats, 25K @ 25 MHz – in-place, Float32 | 487.0 μs | 487.0 μs | 0.999 |
| track/GPS L1CA + Galileo E1B, 8+8 sats, 25K @ 25 MHz – in-place, Float32, threaded | 233.0 μs | 247.0 μs | 0.942 |
| track/GPS L1CA + Galileo E1B, 8+8 sats, 25K @ 25 MHz – out-of-place, Float32 | 491.0 μs | 490.0 μs | 1.0 |
| track/GPS L1CA + Galileo E1B, 8+8 sats, 25K @ 25 MHz – out-of-place, Float32, threaded | 251.0 μs | 236.0 μs | 1.06 ✅ |
| track/GPS L1CA, 1 sat, 1 signal, 5K @ 5 MHz – in-place, Float32 | 11.7 μs | 11.8 μs | 0.997 |
| track/GPS L1CA, 1 sat, 1 signal, 5K @ 5 MHz – out-of-place, Float32 | 11.7 μs | 11.7 μs | 1.0 |
| track/GPS L1CA, 1 sat, 2 signals, 5K @ 5 MHz – in-place, Float32 | 14.6 μs | 14.6 μs | 1.0 |
| track/GPS L1CA, 1 sat, 2 signals, 5K @ 5 MHz – out-of-place, Float32 | 14.3 μs | 14.3 μs | 0.996 |
| track/GPS L1CA, 1 sat, 2K @ 5 MHz – in-place, Float32 | 5.21 μs | 5.28 μs | 0.986 |
| track/GPS L1CA, 1 sat, 2K @ 5 MHz – out-of-place, Float32 | 5.2 μs | 5.11 μs | 1.02 |
| track/GPS L1CA, 1 sat, 3 signals, 5K @ 5 MHz – in-place, Float32 | 17.4 μs | 17.3 μs | 1.01 |
| track/GPS L1CA, 1 sat, 3 signals, 5K @ 5 MHz – out-of-place, Float32 | 17.1 μs | 17.0 μs | 1.01 |
| track/GPS L1CA, 1 sat, bit sync found, 2 blk @ 5 MHz – in-place, Float32 | 22.5 μs | 22.5 μs | 0.999 |
| track/GPS L1CA, 1 sat, bit sync found, 20 blk @ 5 MHz – in-place, Float32 | 221.0 μs | 221.0 μs | 1.0 |
| track/GPS L1CA, 1 sat, bit sync pending, 2 blk @ 5 MHz – in-place, Float32 | 22.6 μs | 22.6 μs | 1.0 |
| track/GPS L1CA, 1 sat, bit sync pending, 20 blk @ 5 MHz – in-place, Float32 | 222.0 μs | 222.0 μs | 1.0 |
| track/GPS L1CA, 1 sat, bit sync pending, 200 blk @ 5 MHz – in-place, Float32 | 2.29 ms | 2.29 ms | 0.999 |
| track/GPS L1CA, 8 sats, 500K @ 5 MHz – in-place, Int16 | 2.73 ms | 2.82 ms | 0.968 |
| track/GPS L1CA, 8 sats, 500K @ 5 MHz – in-place, OneBit | 1.87 ms | 1.96 ms | 0.953 |
| track/GPS L1CA, 8 sats, 500K @ 5 MHz – in-place, TwoBit | 2.79 ms | 2.62 ms | 1.06 ✅ |
| track/GPS L1CA, 8 sats, 5K @ 5 MHz – in-place, Float32 | 53.9 μs | 54.0 μs | 0.998 |
| track/GPS L1CA, 8 sats, 5K @ 5 MHz – in-place, Float32, threaded | 31.9 μs | 31.6 μs | 1.01 |
| track/GPS L1CA, 8 sats, 5K @ 5 MHz – in-place, Int16 | 32.4 μs | 32.2 μs | 1.0 |
| track/GPS L1CA, 8 sats, 5K @ 5 MHz – in-place, OneBit | 23.4 μs | 22.9 μs | 1.02 |
| track/GPS L1CA, 8 sats, 5K @ 5 MHz – out-of-place, Float32 | 54.8 μs | 54.9 μs | 0.998 |
| track/GPS L1CA, 8 sats, 5K @ 5 MHz – out-of-place, Float32, threaded | 32.9 μs | 33.3 μs | 0.988 |
| track/Galileo E1B, 4 sats, 25K @ 25 MHz – in-place, Float32 | 145.0 μs | 146.0 μs | 0.997 |
| track/Galileo E1B, 4 sats, 25K @ 25 MHz – in-place, Float32, threaded | 86.7 μs | 86.9 μs | 0.998 |
| track/Galileo E1B, 4 sats, 25K @ 25 MHz – in-place, Int16 | 68.1 μs | 67.7 μs | 1.01 |
| track/Galileo E1B, 4 sats, 25K @ 25 MHz – out-of-place, Float32 | 146.0 μs | 146.0 μs | 0.999 |
| track/Galileo E1B, 4 sats, 25K @ 25 MHz – out-of-place, Float32, threaded | 87.6 μs | 87.7 μs | 0.999 |
| time_to_load | 106.0 μs | 105.0 μs | 1.01 |
Memory benchmarks (base vs PR head)
Ratio = 7a55f91… / d71924b… (bytes allocated): >1 means the PR allocates less. ✅ ≥ 5 % less, ∞/0 mark a benchmark that drops to / picks up allocations, — means both revisions allocate nothing. A blank cell means the benchmark exists on only one revision (🆕 = new on the PR, 🗑 = removed).
| 7a55f91… | d71924b… | 7a55f91… / d71924b… | |
|---|---|---|---|
| downconvert and correlate/CPU/Float32 | 2 allocs: 528 B | 2 allocs: 528 B | 1.0 |
| downconvert and correlate/CPU/Float32 4ant | 2 allocs: 848 B | 2 allocs: 848 B | 1.0 |
| downconvert and correlate/CPU/Float64 | 2 allocs: 528 B | 2 allocs: 528 B | 1.0 |
| downconvert and correlate/CPU/Int16 | 2 allocs: 528 B | 2 allocs: 528 B | 1.0 |
| downconvert and correlate/CPU/Int16 4ant | 2 allocs: 848 B | 2 allocs: 848 B | 1.0 |
| downconvert and correlate/CPU/Int32 | 2 allocs: 528 B | 2 allocs: 528 B | 1.0 |
| fused kernel/1-ant dynamic taps | 0 allocs: 0 B | 0 allocs: 0 B | — |
| fused kernel/1-ant static taps | 0 allocs: 0 B | 0 allocs: 0 B | — |
| fused kernel/4-ant dynamic taps | 0 allocs: 0 B | 0 allocs: 0 B | — |
| fused kernel/4-ant static taps | 0 allocs: 0 B | 0 allocs: 0 B | — |
| fused tuple kernel/1-ant N=2 | 0 allocs: 0 B | 0 allocs: 0 B | — |
| fused tuple kernel/1-ant N=3 | 0 allocs: 0 B | 0 allocs: 0 B | — |
| fused tuple kernel/2-ant N=2 | 0 allocs: 0 B | 0 allocs: 0 B | — |
| fused tuple kernel/2-ant N=3 | 0 allocs: 0 B | 0 allocs: 0 B | — |
| fused tuple kernel/4-ant N=2 | 0 allocs: 0 B | 0 allocs: 0 B | — |
| fused tuple kernel/4-ant N=3 | 0 allocs: 0 B | 0 allocs: 0 B | — |
| noise estimation/append_noise_observation! – steady state | 0 allocs: 0 B | 0 allocs: 0 B | — |
| noise estimation/track! – GPS L1CA, 1 sat, 2K @ 5 MHz with a noise reference | 0 allocs: 0 B | 0 allocs: 0 B | — |
| noise estimation/update_noise! – 1 ms @ 20 MHz | 0 allocs: 0 B | 0 allocs: 0 B | — |
| noise estimation/update_noise! – 1 ms @ 4 MHz | 0 allocs: 0 B | 0 allocs: 0 B | — |
| noise estimation/update_noise! – 1 ms @ 4 MHz, 4 antennas | 0 allocs: 0 B | 0 allocs: 0 B | — |
| track/GPS L1CA + Galileo E1B, 8+8 sats, 25K @ 25 MHz – in-place, Float32 | 0 allocs: 0 B | 0 allocs: 0 B | — |
| track/GPS L1CA + Galileo E1B, 8+8 sats, 25K @ 25 MHz – in-place, Float32, threaded | 4 allocs: 384 B | 4 allocs: 384 B | 1.0 |
| track/GPS L1CA + Galileo E1B, 8+8 sats, 25K @ 25 MHz – out-of-place, Float32 | 26 allocs: 9840 B | 26 allocs: 9840 B | 1.0 |
| track/GPS L1CA + Galileo E1B, 8+8 sats, 25K @ 25 MHz – out-of-place, Float32, threaded | 30 allocs: 10224 B | 30 allocs: 10224 B | 1.0 |
| track/GPS L1CA, 1 sat, 1 signal, 5K @ 5 MHz – in-place, Float32 | 0 allocs: 0 B | 0 allocs: 0 B | — |
| track/GPS L1CA, 1 sat, 1 signal, 5K @ 5 MHz – out-of-place, Float32 | 9 allocs: 896 B | 9 allocs: 896 B | 1.0 |
| track/GPS L1CA, 1 sat, 2 signals, 5K @ 5 MHz – in-place, Float32 | 0 allocs: 0 B | 0 allocs: 0 B | — |
| track/GPS L1CA, 1 sat, 2 signals, 5K @ 5 MHz – out-of-place, Float32 | 9 allocs: 1296 B | 9 allocs: 1296 B | 1.0 |
| track/GPS L1CA, 1 sat, 2K @ 5 MHz – in-place, Float32 | 0 allocs: 0 B | 0 allocs: 0 B | — |
| track/GPS L1CA, 1 sat, 2K @ 5 MHz – out-of-place, Float32 | 9 allocs: 896 B | 9 allocs: 896 B | 1.0 |
| track/GPS L1CA, 1 sat, 3 signals, 5K @ 5 MHz – in-place, Float32 | 0 allocs: 0 B | 0 allocs: 0 B | — |
| track/GPS L1CA, 1 sat, 3 signals, 5K @ 5 MHz – out-of-place, Float32 | 9 allocs: 1648 B | 9 allocs: 1648 B | 1.0 |
| track/GPS L1CA, 1 sat, bit sync found, 2 blk @ 5 MHz – in-place, Float32 | 7 allocs: 384 B | 7 allocs: 384 B | 1.0 |
| track/GPS L1CA, 1 sat, bit sync found, 20 blk @ 5 MHz – in-place, Float32 | 7 allocs: 384 B | 7 allocs: 384 B | 1.0 |
| track/GPS L1CA, 1 sat, bit sync pending, 2 blk @ 5 MHz – in-place, Float32 | 7 allocs: 384 B | 7 allocs: 384 B | 1.0 |
| track/GPS L1CA, 1 sat, bit sync pending, 20 blk @ 5 MHz – in-place, Float32 | 7 allocs: 384 B | 7 allocs: 384 B | 1.0 |
| track/GPS L1CA, 1 sat, bit sync pending, 200 blk @ 5 MHz – in-place, Float32 | 7 allocs: 384 B | 7 allocs: 384 B | 1.0 |
| track/GPS L1CA, 8 sats, 500K @ 5 MHz – in-place, Int16 | 62 allocs: 25920 B | 62 allocs: 25920 B | 1.0 |
| track/GPS L1CA, 8 sats, 500K @ 5 MHz – in-place, OneBit | 62 allocs: 25808 B | 62 allocs: 25808 B | 1.0 |
| track/GPS L1CA, 8 sats, 500K @ 5 MHz – in-place, TwoBit | 62 allocs: 25856 B | 62 allocs: 25856 B | 1.0 |
| track/GPS L1CA, 8 sats, 5K @ 5 MHz – in-place, Float32 | 0 allocs: 0 B | 0 allocs: 0 B | — |
| track/GPS L1CA, 8 sats, 5K @ 5 MHz – in-place, Float32, threaded | 2 allocs: 192 B | 2 allocs: 192 B | 1.0 |
| track/GPS L1CA, 8 sats, 5K @ 5 MHz – in-place, Int16 | 14 allocs: 1280 B | 14 allocs: 1280 B | 1.0 |
| track/GPS L1CA, 8 sats, 5K @ 5 MHz – in-place, OneBit | 14 allocs: 1168 B | 14 allocs: 1168 B | 1.0 |
| track/GPS L1CA, 8 sats, 5K @ 5 MHz – out-of-place, Float32 | 10 allocs: 4280 B | 10 allocs: 4280 B | 1.0 |
| track/GPS L1CA, 8 sats, 5K @ 5 MHz – out-of-place, Float32, threaded | 12 allocs: 4472 B | 12 allocs: 4472 B | 1.0 |
| track/Galileo E1B, 4 sats, 25K @ 25 MHz – in-place, Float32 | 0 allocs: 0 B | 0 allocs: 0 B | — |
| track/Galileo E1B, 4 sats, 25K @ 25 MHz – in-place, Float32, threaded | 2 allocs: 192 B | 2 allocs: 192 B | 1.0 |
| track/Galileo E1B, 4 sats, 25K @ 25 MHz – in-place, Int16 | 6 allocs: 640 B | 6 allocs: 640 B | 1.0 |
| track/Galileo E1B, 4 sats, 25K @ 25 MHz – out-of-place, Float32 | 10 allocs: 2680 B | 10 allocs: 2680 B | 1.0 |
| track/Galileo E1B, 4 sats, 25K @ 25 MHz – out-of-place, Float32, threaded | 12 allocs: 2872 B | 12 allocs: 2872 B | 1.0 |
| time_to_load | 149 allocs: 11408 B | 149 allocs: 11408 B | 1.0 |
Contributor
Benchmark Results (minimum time) — macos-14Reporting the minimum over all samples (robust to shared-runner contention), not the median. Alternative backends vs Float32 (
|
| Scenario | F32 | I16 | 1b | 2b | ×I16 | ×1b | ×2b |
|---|---|---|---|---|---|---|---|
| GPS L1CA, 1 sat, 3 signals @ 5 MHz | 12.1 μs | 6.08 μs | 5.03 μs | 7.79 μs | 1.99 ✅ | 2.41 ✅ | 1.56 ✅ |
| GPS L1CA, 1 sat, 4 antennas @ 5 MHz | 12.9 μs | 6.82 μs | 5.14 μs | 10.5 μs | 1.89 ✅ | 2.51 ✅ | 1.24 ✅ |
| GPS L1CA, 8 sats @ 40 MHz | 151.0 μs | 49.8 μs | 32.3 μs | 55.3 μs | 3.03 ✅ | 4.67 ✅ | 2.72 ✅ |
| GPS L1CA, 8 sats @ 5 MHz, 100 ms buffer | 2.17 ms | 893.0 μs | 662.0 μs | 987.0 μs | 2.43 ✅ | 3.28 ✅ | 2.2 ✅ |
| GPS L1CA, 8 sats @ 5 MHz | 23.4 μs | 10.8 μs | 9.25 μs | 13.5 μs | 2.16 ✅ | 2.53 ✅ | 1.73 ✅ |
| Galileo E1B, 4 sats @ 25 MHz | 109.0 μs | 44.2 μs | 2.47 ✅ | ||||
| dynamic taps @ 5 MHz (kernel) | 7.16 μs | 2.39 μs | 1.59 μs | 2.66 μs | 2.99 ✅ | 4.5 ✅ | 2.69 ✅ |
Time benchmarks (base vs PR head)
Ratio = 7a55f91… / d71924b…: >1 means the PR is faster. ✅ ≥ 5 % faster,
| 7a55f91… | d71924b… | 7a55f91… / d71924b… | |
|---|---|---|---|
| downconvert and correlate/CPU/Float32 | 2.65 μs | 2.63 μs | 1.01 |
| downconvert and correlate/CPU/Float32 4ant | 5.26 μs | 5.23 μs | 1.01 |
| downconvert and correlate/CPU/Float64 | 2.88 μs | 2.84 μs | 1.01 |
| downconvert and correlate/CPU/Int16 | 2.81 μs | 2.78 μs | 1.01 |
| downconvert and correlate/CPU/Int16 4ant | 5.2 μs | 5.19 μs | 1.0 |
| downconvert and correlate/CPU/Int32 | 2.72 μs | 2.69 μs | 1.01 |
| fused kernel/1-ant dynamic taps | 2.74 μs | 2.74 μs | 1.0 |
| fused kernel/1-ant static taps | 2.21 μs | 2.21 μs | 1.0 |
| fused kernel/4-ant dynamic taps | 8.33 μs | 8.33 μs | 1.0 |
| fused kernel/4-ant static taps | 4.76 μs | 4.74 μs | 1.0 |
| fused tuple kernel/1-ant N=2 | 3.51 μs | 3.51 μs | 1.0 |
| fused tuple kernel/1-ant N=3 | 3.65 μs | 3.65 μs | 1.0 |
| fused tuple kernel/2-ant N=2 | 6.77 μs | 6.76 μs | 1.0 |
| fused tuple kernel/2-ant N=3 | 6.02 μs | 6.02 μs | 1.0 |
| fused tuple kernel/4-ant N=2 | 12.4 μs | 12.4 μs | 1.0 |
| fused tuple kernel/4-ant N=3 | 10.8 μs | 10.8 μs | 1.0 |
| noise estimation/append_noise_observation! – steady state | 16.0 ns | 16.0 ns | 1.0 |
| noise estimation/track! – GPS L1CA, 1 sat, 2K @ 5 MHz with a noise reference | 5.6 μs | 5.64 μs | 0.994 |
| noise estimation/update_noise! – 1 ms @ 20 MHz | 22.8 μs | 22.8 μs | 1.0 |
| noise estimation/update_noise! – 1 ms @ 4 MHz | 4.95 μs | 4.96 μs | 0.999 |
| noise estimation/update_noise! – 1 ms @ 4 MHz, 4 antennas | 10.2 μs | 10.2 μs | 1.0 |
| track/GPS L1CA + Galileo E1B, 8+8 sats, 25K @ 25 MHz – in-place, Float32 | 585.0 μs | 586.0 μs | 0.999 |
| track/GPS L1CA + Galileo E1B, 8+8 sats, 25K @ 25 MHz – in-place, Float32, threaded | 228.0 μs | 228.0 μs | 1.0 |
| track/GPS L1CA + Galileo E1B, 8+8 sats, 25K @ 25 MHz – out-of-place, Float32 | 586.0 μs | 587.0 μs | 0.999 |
| track/GPS L1CA + Galileo E1B, 8+8 sats, 25K @ 25 MHz – out-of-place, Float32, threaded | 230.0 μs | 229.0 μs | 1.0 |
| track/GPS L1CA, 1 sat, 1 signal, 5K @ 5 MHz – in-place, Float32 | 13.0 μs | 13.0 μs | 1.0 |
| track/GPS L1CA, 1 sat, 1 signal, 5K @ 5 MHz – out-of-place, Float32 | 13.0 μs | 13.0 μs | 1.0 |
| track/GPS L1CA, 1 sat, 2 signals, 5K @ 5 MHz – in-place, Float32 | 17.7 μs | 17.7 μs | 1.0 |
| track/GPS L1CA, 1 sat, 2 signals, 5K @ 5 MHz – out-of-place, Float32 | 17.7 μs | 17.7 μs | 1.0 |
| track/GPS L1CA, 1 sat, 2K @ 5 MHz – in-place, Float32 | 5.6 μs | 5.62 μs | 0.996 |
| track/GPS L1CA, 1 sat, 2K @ 5 MHz – out-of-place, Float32 | 5.56 μs | 5.56 μs | 0.999 |
| track/GPS L1CA, 1 sat, 3 signals, 5K @ 5 MHz – in-place, Float32 | 19.4 μs | 19.2 μs | 1.01 |
| track/GPS L1CA, 1 sat, 3 signals, 5K @ 5 MHz – out-of-place, Float32 | 19.1 μs | 19.0 μs | 1.0 |
| track/GPS L1CA, 1 sat, bit sync found, 2 blk @ 5 MHz – in-place, Float32 | 25.6 μs | 25.5 μs | 1.0 |
| track/GPS L1CA, 1 sat, bit sync found, 20 blk @ 5 MHz – in-place, Float32 | 250.0 μs | 251.0 μs | 0.998 |
| track/GPS L1CA, 1 sat, bit sync pending, 2 blk @ 5 MHz – in-place, Float32 | 25.7 μs | 25.6 μs | 1.0 |
| track/GPS L1CA, 1 sat, bit sync pending, 20 blk @ 5 MHz – in-place, Float32 | 252.0 μs | 252.0 μs | 1.0 |
| track/GPS L1CA, 1 sat, bit sync pending, 200 blk @ 5 MHz – in-place, Float32 | 2.55 ms | 2.57 ms | 0.994 |
| track/GPS L1CA, 8 sats, 500K @ 5 MHz – in-place, Int16 | 2.27 ms | 2.28 ms | 0.996 |
| track/GPS L1CA, 8 sats, 500K @ 5 MHz – in-place, OneBit | 1.5 ms | 1.51 ms | 0.998 |
| track/GPS L1CA, 8 sats, 500K @ 5 MHz – in-place, TwoBit | 2.43 ms | 2.42 ms | 1.0 |
| track/GPS L1CA, 8 sats, 5K @ 5 MHz – in-place, Float32 | 59.4 μs | 59.8 μs | 0.994 |
| track/GPS L1CA, 8 sats, 5K @ 5 MHz – in-place, Float32, threaded | 22.0 μs | 22.0 μs | 0.998 |
| track/GPS L1CA, 8 sats, 5K @ 5 MHz – in-place, Int16 | 25.8 μs | 25.9 μs | 0.994 |
| track/GPS L1CA, 8 sats, 5K @ 5 MHz – in-place, OneBit | 18.8 μs | 18.9 μs | 0.996 |
| track/GPS L1CA, 8 sats, 5K @ 5 MHz – out-of-place, Float32 | 60.0 μs | 60.2 μs | 0.997 |
| track/GPS L1CA, 8 sats, 5K @ 5 MHz – out-of-place, Float32, threaded | 22.4 μs | 22.5 μs | 0.998 |
| track/Galileo E1B, 4 sats, 25K @ 25 MHz – in-place, Float32 | 176.0 μs | 176.0 μs | 1.0 |
| track/Galileo E1B, 4 sats, 25K @ 25 MHz – in-place, Float32, threaded | 75.4 μs | 75.4 μs | 0.999 |
| track/Galileo E1B, 4 sats, 25K @ 25 MHz – in-place, Int16 | 75.0 μs | 75.1 μs | 0.998 |
| track/Galileo E1B, 4 sats, 25K @ 25 MHz – out-of-place, Float32 | 177.0 μs | 177.0 μs | 1.0 |
| track/Galileo E1B, 4 sats, 25K @ 25 MHz – out-of-place, Float32, threaded | 104.0 μs | 104.0 μs | 1.0 |
| time_to_load | 442.0 μs | 247.0 μs | 1.79 ✅ |
Memory benchmarks (base vs PR head)
Ratio = 7a55f91… / d71924b… (bytes allocated): >1 means the PR allocates less. ✅ ≥ 5 % less, ∞/0 mark a benchmark that drops to / picks up allocations, — means both revisions allocate nothing. A blank cell means the benchmark exists on only one revision (🆕 = new on the PR, 🗑 = removed).
| 7a55f91… | d71924b… | 7a55f91… / d71924b… | |
|---|---|---|---|
| downconvert and correlate/CPU/Float32 | 2 allocs: 528 B | 2 allocs: 528 B | 1.0 |
| downconvert and correlate/CPU/Float32 4ant | 2 allocs: 848 B | 2 allocs: 848 B | 1.0 |
| downconvert and correlate/CPU/Float64 | 2 allocs: 528 B | 2 allocs: 528 B | 1.0 |
| downconvert and correlate/CPU/Int16 | 2 allocs: 528 B | 2 allocs: 528 B | 1.0 |
| downconvert and correlate/CPU/Int16 4ant | 2 allocs: 848 B | 2 allocs: 848 B | 1.0 |
| downconvert and correlate/CPU/Int32 | 2 allocs: 528 B | 2 allocs: 528 B | 1.0 |
| fused kernel/1-ant dynamic taps | 0 allocs: 0 B | 0 allocs: 0 B | — |
| fused kernel/1-ant static taps | 0 allocs: 0 B | 0 allocs: 0 B | — |
| fused kernel/4-ant dynamic taps | 0 allocs: 0 B | 0 allocs: 0 B | — |
| fused kernel/4-ant static taps | 0 allocs: 0 B | 0 allocs: 0 B | — |
| fused tuple kernel/1-ant N=2 | 0 allocs: 0 B | 0 allocs: 0 B | — |
| fused tuple kernel/1-ant N=3 | 0 allocs: 0 B | 0 allocs: 0 B | — |
| fused tuple kernel/2-ant N=2 | 0 allocs: 0 B | 0 allocs: 0 B | — |
| fused tuple kernel/2-ant N=3 | 0 allocs: 0 B | 0 allocs: 0 B | — |
| fused tuple kernel/4-ant N=2 | 0 allocs: 0 B | 0 allocs: 0 B | — |
| fused tuple kernel/4-ant N=3 | 0 allocs: 0 B | 0 allocs: 0 B | — |
| noise estimation/append_noise_observation! – steady state | 0 allocs: 0 B | 0 allocs: 0 B | — |
| noise estimation/track! – GPS L1CA, 1 sat, 2K @ 5 MHz with a noise reference | 0 allocs: 0 B | 0 allocs: 0 B | — |
| noise estimation/update_noise! – 1 ms @ 20 MHz | 0 allocs: 0 B | 0 allocs: 0 B | — |
| noise estimation/update_noise! – 1 ms @ 4 MHz | 0 allocs: 0 B | 0 allocs: 0 B | — |
| noise estimation/update_noise! – 1 ms @ 4 MHz, 4 antennas | 0 allocs: 0 B | 0 allocs: 0 B | — |
| track/GPS L1CA + Galileo E1B, 8+8 sats, 25K @ 25 MHz – in-place, Float32 | 0 allocs: 0 B | 0 allocs: 0 B | — |
| track/GPS L1CA + Galileo E1B, 8+8 sats, 25K @ 25 MHz – in-place, Float32, threaded | 4 allocs: 384 B | 4 allocs: 384 B | 1.0 |
| track/GPS L1CA + Galileo E1B, 8+8 sats, 25K @ 25 MHz – out-of-place, Float32 | 26 allocs: 10464 B | 26 allocs: 10464 B | 1.0 |
| track/GPS L1CA + Galileo E1B, 8+8 sats, 25K @ 25 MHz – out-of-place, Float32, threaded | 30 allocs: 10848 B | 30 allocs: 10848 B | 1.0 |
| track/GPS L1CA, 1 sat, 1 signal, 5K @ 5 MHz – in-place, Float32 | 0 allocs: 0 B | 0 allocs: 0 B | — |
| track/GPS L1CA, 1 sat, 1 signal, 5K @ 5 MHz – out-of-place, Float32 | 9 allocs: 896 B | 9 allocs: 896 B | 1.0 |
| track/GPS L1CA, 1 sat, 2 signals, 5K @ 5 MHz – in-place, Float32 | 0 allocs: 0 B | 0 allocs: 0 B | — |
| track/GPS L1CA, 1 sat, 2 signals, 5K @ 5 MHz – out-of-place, Float32 | 9 allocs: 1296 B | 9 allocs: 1296 B | 1.0 |
| track/GPS L1CA, 1 sat, 2K @ 5 MHz – in-place, Float32 | 0 allocs: 0 B | 0 allocs: 0 B | — |
| track/GPS L1CA, 1 sat, 2K @ 5 MHz – out-of-place, Float32 | 9 allocs: 896 B | 9 allocs: 896 B | 1.0 |
| track/GPS L1CA, 1 sat, 3 signals, 5K @ 5 MHz – in-place, Float32 | 0 allocs: 0 B | 0 allocs: 0 B | — |
| track/GPS L1CA, 1 sat, 3 signals, 5K @ 5 MHz – out-of-place, Float32 | 9 allocs: 1648 B | 9 allocs: 1648 B | 1.0 |
| track/GPS L1CA, 1 sat, bit sync found, 2 blk @ 5 MHz – in-place, Float32 | 7 allocs: 384 B | 7 allocs: 384 B | 1.0 |
| track/GPS L1CA, 1 sat, bit sync found, 20 blk @ 5 MHz – in-place, Float32 | 7 allocs: 384 B | 7 allocs: 384 B | 1.0 |
| track/GPS L1CA, 1 sat, bit sync pending, 2 blk @ 5 MHz – in-place, Float32 | 7 allocs: 384 B | 7 allocs: 384 B | 1.0 |
| track/GPS L1CA, 1 sat, bit sync pending, 20 blk @ 5 MHz – in-place, Float32 | 7 allocs: 384 B | 7 allocs: 384 B | 1.0 |
| track/GPS L1CA, 1 sat, bit sync pending, 200 blk @ 5 MHz – in-place, Float32 | 7 allocs: 384 B | 7 allocs: 384 B | 1.0 |
| track/GPS L1CA, 8 sats, 500K @ 5 MHz – in-place, Int16 | 62 allocs: 28928 B | 62 allocs: 28928 B | 1.0 |
| track/GPS L1CA, 8 sats, 500K @ 5 MHz – in-place, OneBit | 62 allocs: 28816 B | 62 allocs: 28816 B | 1.0 |
| track/GPS L1CA, 8 sats, 500K @ 5 MHz – in-place, TwoBit | 62 allocs: 28864 B | 62 allocs: 28864 B | 1.0 |
| track/GPS L1CA, 8 sats, 5K @ 5 MHz – in-place, Float32 | 0 allocs: 0 B | 0 allocs: 0 B | — |
| track/GPS L1CA, 8 sats, 5K @ 5 MHz – in-place, Float32, threaded | 2 allocs: 192 B | 2 allocs: 192 B | 1.0 |
| track/GPS L1CA, 8 sats, 5K @ 5 MHz – in-place, Int16 | 14 allocs: 1280 B | 14 allocs: 1280 B | 1.0 |
| track/GPS L1CA, 8 sats, 5K @ 5 MHz – in-place, OneBit | 14 allocs: 1168 B | 14 allocs: 1168 B | 1.0 |
| track/GPS L1CA, 8 sats, 5K @ 5 MHz – out-of-place, Float32 | 10 allocs: 4656 B | 10 allocs: 4656 B | 1.0 |
| track/GPS L1CA, 8 sats, 5K @ 5 MHz – out-of-place, Float32, threaded | 12 allocs: 4848 B | 12 allocs: 4848 B | 1.0 |
| track/Galileo E1B, 4 sats, 25K @ 25 MHz – in-place, Float32 | 0 allocs: 0 B | 0 allocs: 0 B | — |
| track/Galileo E1B, 4 sats, 25K @ 25 MHz – in-place, Float32, threaded | 2 allocs: 192 B | 2 allocs: 192 B | 1.0 |
| track/Galileo E1B, 4 sats, 25K @ 25 MHz – in-place, Int16 | 6 allocs: 640 B | 6 allocs: 640 B | 1.0 |
| track/Galileo E1B, 4 sats, 25K @ 25 MHz – out-of-place, Float32 | 10 allocs: 3056 B | 10 allocs: 3056 B | 1.0 |
| track/Galileo E1B, 4 sats, 25K @ 25 MHz – out-of-place, Float32, threaded | 12 allocs: 3248 B | 12 allocs: 3248 B | 1.0 |
| time_to_load | 200 allocs: 14176 B | 200 allocs: 14176 B | 1.0 |
Member
Author
|
CI follow-up after adding direct gap-handling coverage in d71924b:
I reviewed the Ubuntu and macOS benchmark comments. The bit-sync workloads affected by this PR are neutral on both platforms (roughly 1.00×), and allocations are unchanged. macOS package load improves from 442 μs to 247 μs. Ubuntu flags a 32.9 ns noise-observation microbenchmark and one threaded mixed-signal variant, but neither code path is changed here; macOS is neutral on both, and the paired Ubuntu threaded out-of-place variant moves by the same magnitude in the opposite direction. These are runner/scheduling variance rather than an actionable regression. |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Hardware correlator producers can observe elapsed signal time without having a valid prompt measurement. Add bit-clock APIs that advance those missing primary-code blocks directly, preserving bit and secondary-code phase without feeding fabricated zero correlators into discriminators, prompt filters, or C/N0 estimation.
The implementation also drains every completed bit when a record spans multiple symbol boundaries, preventing an overshoot from leaving the bit accumulator permanently stalled.
This is required by JuliaGNSS/GNSSReceiver.jl#129 and the hardware work tracked in JuliaGNSS/GNSSReceiver.jl#107.