Add probe-queue-contention, which prices the tail claim - #90

Open

MikeGrier wants to merge 74 commits into

mainfrom

mikegrier/probe-queue-contention

Commits on Sep 14, 2026

feat(platform-probes): add probe-queue-contention, which prices the tail claim

Mike Grier
and
Copilot
committed
381303c
fix(platform-probes): restore the default-features build the probe's features took away

Mike Grier
and
Copilot
committed
d74a00c
Merge remote-tracking branch 'origin/main' into mikegrier/probe-queue-contention
Mike Grier
committed
d8783f9
fix(platform-probes): time the producers, not this thread, and refuse a spin on a dead queue

Mike Grier
and
Copilot
committed
d49a71f
docs(platform-probes): withdraw the "re-apportionment is free" claim

Mike Grier
and
Copilot
committed
0ca3ea1
docs(waitable-queues): drop "at no measured cost" from the layout guidance

Mike Grier
and
Copilot
committed
2ffea52
docs(platform-probes): treat a wide self-control as a defect report, not a wider ruler

Mike Grier
and
Copilot
committed
86e851b
docs(platform-probes): say what to try first on high variance, and where the floor is

Mike Grier
and
Copilot
committed
a9d7427
docs(platform-probes): record the sampling parameters as capture parameters

Mike Grier
and
Copilot
committed
123bdbe

Commits on Sep 15, 2026

docs(platform-probes): withdraw the drained widening figure too

Mike Grier
and
Copilot
committed
a4b752a
docs(waitable-queues): three more sites carrying the withdrawn cost claim

Mike Grier
and
Copilot
committed
0759b5c
fix(platform-probes): restore the i686 build this probe's features removed

Mike Grier
and
Copilot
committed
2d29bee
docs(waitable-queues): the last sites of the withdrawn claim, and a stale figure

Mike Grier
and
Copilot
committed
40226d7
fix(platform-probes): warn in the report when the build cannot measure

Mike Grier
and
Copilot
committed
56693ac
fix(platform-probes): stop the attribution correction stopping at one site

Mike Grier
and
Copilot
committed
b23d70e
fix(platform-probes): the "bounds it" correction was itself an over-claim

Mike Grier
and
Copilot
committed
88df746
docs: stop restating volatile figures outside the one table that owns them

Mike Grier
and
Copilot
committed
236c8d0
docs(platform-probes): an observation is reviewed against its procedure, not argued

Mike Grier
and
Copilot
committed
1a42d9a
fix(platform-probes): label the instrument as what it measures, not what it is about

Mike Grier
and
Copilot
committed
906aa01
docs: remove advice built on top of measurements

Mike Grier
and
Copilot
committed
0ad7465
docs: label pre-correction figures, and state the defect instead of recommending against it

Mike Grier
and
Copilot
committed
9c5f3e0
docs(platform-probes): a broken doc link, a missing capture field, and three more claims

Mike Grier
and
Copilot
committed
3647539
fix(platform-probes): gate dwcas on the instruction, not just the architecture

Mike Grier
and
Copilot
committed
ea090c5
docs(waitable-queues): fix a broken sentence, a too-strong claim, and a construction

Mike Grier
and
Copilot
committed
d973924
docs: derive the layout count, correct the ceiling, and order M4.4 before M4.2

Mike Grier
and
Copilot
committed
bba97ba
docs: Wide moves the recurrence to 2^64, it does not remove it

Mike Grier
and
Copilot
committed
a99108f
docs(waitable-queues): regenerate the published measurements with attribution

Mike Grier
and
Copilot
committed
b196db6
docs: the reservation column was a field ceiling, not a reachable count

Mike Grier
and
Copilot
committed
d5628ce
docs: the refusal counts do not say what that section said they say

Mike Grier
and
Copilot
committed
c9a9127
docs: mark D-17 superseded, and scope the forced-split premise to Balanced

Mike Grier
and
Copilot
committed
8b31650
docs: the 2^31 capacity is a 64-bit figure, and other target-dependence

Mike Grier
and
Copilot
committed
16c329e
test(platform-probes): cover measured_span and the report's renderer

Mike Grier
and
Copilot
committed
34b47d6
docs: require review feedback to be answered where it was raised

Mike Grier
and
Copilot
committed
c83d411
test(platform-probes): cover median_run's selection and the refusal column

Mike Grier
and
Copilot
committed
23f6e0f
docs(waitable-queues): withdraw the producer-count reservation bound as false

Mike Grier
and
Copilot
committed
e0bb76c
docs: record that restatement count, not prose volume, is the error surface

Mike Grier
and
Copilot
committed
b173198
docs: the wrap exposure is a count, not a rate, and 6.4x came from withdrawn data

Mike Grier
and
Copilot
committed
679dd8e

Commits on Sep 16, 2026

fix(platform-probes): the baseline has no pushes, so the units are per operation

Mike Grier
and
Copilot
committed
9f204bd
docs(waitable-queues): withdraw the control-band conclusion, which six sites made

Mike Grier
and
Copilot
committed
c846922
fix(platform-probes): carry the dispersion the crate's own contract requires

Mike Grier
and
Copilot
committed
fecd352
chore(topology): the placement tool is placement-probe, not probe-core-affinity

Mike Grier
and
Copilot
committed
3062c47
docs(waitable-queues): publish the dispersion, and stop naming a tool that does not exist

Mike Grier
and
Copilot
committed
81cca99
docs(platform-probes): record that the dispersion work landed, and how to run it

Mike Grier
and
Copilot
committed
da20e51
docs(waitable-queues): the rustdoc table still held the superseded capture

Mike Grier
and
Copilot
committed
fca9a9c
fix(platform-probes): the refusal counts do not rule the tail out

Mike Grier
and
Copilot
committed
0416f0f
docs(waitable-queues): D-41 still called the reservation half the wrong one to spend bits on

Mike Grier
and
Copilot
committed
cddef16
docs(platform-probes): split M4.2's completed half out as M4.5 and archive it

Mike Grier
and
Copilot
committed
bd82bf9
docs(waitable-queues): the horizons were still "measured" in seven places

Mike Grier
and
Copilot
committed
bc56465
fix(waitable-queues): restore backticks and a newline PowerShell ate

Mike Grier
and
Copilot
committed
0f36dfa
docs: a design note should not record which branch a checklist item is on

Mike Grier
and
Copilot
committed
2039307
fix(platform-probes): the derived ratios had no dispersion either

Mike Grier
and
Copilot
committed
32cb97a
docs: count the tabular/prose split instead of estimating it

Mike Grier
and
Copilot
committed
8950987
docs: prose carries the claim, an artifact carries the number

Mike Grier
and
Copilot
committed
e75bc7e
fix(platform-probes): ratio_bounds returned the inverse of what it documented

Mike Grier
and
Copilot
committed
a8d4c7f
docs: the field binds only once the queue is at least that large

Mike Grier
and
Copilot
committed
3761c63
docs: the section arguing against embedded figures was embedded figures

Mike Grier
and
Copilot
committed
ce21ff3
docs(waitable-queues): 2^64 is about 5,000 years, not "unreachable"

Mike Grier
and
Copilot
committed
893a279
fix(platform-probes): a zero numerator rendered 0.00x, which reads as a result

Mike Grier
and
Copilot
committed
f3d5757
docs: the capture date needs its offset, and the time-figure count moved

Mike Grier
and
Copilot
committed
10ef5fc
docs: apply the one-home rule to the prose that introduced it

Mike Grier
and
Copilot
committed
9340fcf
fix(platform-probes): an unmeasured spread rendered as perfect stability

Mike Grier
and
Copilot
committed
28978ba
style(platform-probes): reflow two assertions cargo fmt wanted wrapped

Mike Grier
and
Copilot
committed
ad2491b
fix(probes): stop a shape that never ran from publishing figures

Mike Grier
and
Copilot
committed
df16f01
docs(queues): stop calling a finite recurrence horizon an absolute one

Mike Grier
and
Copilot
committed
eb807ca
docs(probes): correct the M4.5 archive, which inverted its own lesson

Mike Grier
and
Copilot
committed
eee03aa
fix(probes): count measured layouts, and size the ratio column from its cells

Mike Grier
and
Copilot
committed
117a7ad
docs: state every claim-position horizon as finite and rate-dependent

Mike Grier
and
Copilot
committed
0510b92
docs: present the data and stop, rather than drawing the reader's conclusion

Mike Grier
and
Copilot
committed
65e74e1
docs: make rule 5 absolute, removing the carve-out that invited argument

Mike Grier
and
Copilot
committed
fd0864f
docs: draw rule 5's boundary at the slash count, which is where it is

Mike Grier
and
Copilot
committed
401733b
docs: remove a paragraph rule 5 said twice

Mike Grier
and
Copilot
committed
ae12bec
fix(probes): reject non-finite spans, and size the last two ratio tables

Mike Grier
and
Copilot
committed
b95dda8
docs: publish only the exposure figures the disclosed rate supports

Mike Grier
and
Copilot
committed
eef1544
fix(probes): stop the drained consumer when the producer phase unwinds

Mike Grier
and
Copilot
committed
c471adf