Reduce spinlock detection limits now that spinlocks do backoff

[msimberg]

After #927 spin_k and yield_k go through iterations much slower. This means that the spinlock deadlock warnings may not be printed in any reasonable time with the old limits. This reduces the error limit by an order of magnitude, and the warning limit by two orders of magnitude (since this is typically the interesting message, and it's harmless when printed). The limits can still be changed with configuration options as before, and the detection is still opt-in.

[pika-bot]

Performance test report

pika Performance

Comparison

BENCHMARK	RESULT
Task Overhead - Create Thread Hierarchical - Latch	--

Info

Property	Before	After
pika Commit	0abc084	5415ea1
pika Datetime	2024-02-19T15:15:15+00:00	2024-04-18T15:42:43+00:00
Compiler	/apps/daint/SSL/pika/spack/lib/spack/env/clang/clang++ 11.0.1	/apps/daint/SSL/pika/spack/lib/spack/env/clang/clang++ 11.0.1
Hostname	nid00025	nid00025
Datetime	2024-02-19T16:26:16.072067+01:00	2024-04-18T17:49:51.107061+02:00
Clustername	daint	daint
Envfile

Explanation of Symbols

Symbol	MEANING
=	No performance change (confidence interval within ±1%)
(=)	Probably no performance change (confidence interval within ±2%)
(+)/(-)	Very small performance improvement/degradation (≤1%)
+/-	Small performance improvement/degradation (>10%)
++/--	Large performance improvement/degradation (>10%)
+++/---	Very large performance improvement/degradation (>10%)
?	Probably no change, but quite large uncertainty (confidence interval with ±5%)
??	Unclear result, very large uncertainty (±10%)
???	Something unexpected…