ichthyo/LUMIERA.clone
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

Scheduler-test: document new instrumentation facility with simple test

Browse source 
...turns out rather challenging to come up with any test case,
that is both meaningful, simple to setup and understand, yet still
produces somewhat stable values. `IncidenceCount` seems most valuable
for investigation and direct inspection of results
This commit is contained in:
Fischlurch 2024-02-17 21:55:21 +01:00
parent 0e7bdcc5b5
commit 7efaf5f0cc
2 changed files with 46 additions and 81 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

106
tests/vault/gear/scheduler-stress-test.cpp
View file

@ -288,95 +288,43 @@ namespace test {
/** @test TODO verify detailed instrumentation of job invocations
* @todo WIP 2/24 🔁 define ⟶ implement
/** @test verify capability for instrumentation of job invocations
* @see IncidenceCount_test
* @todo WIP 2/24 ✔ define ⟶ ✔ implement
*/
void
verify_instrumentation()
{
MARK_TEST_FUN
TestChainLoad testLoad{800};
testLoad.setWeight(1)
.printTopologyDOT()
.printTopologyStatistics()
;
const size_t NODES = 20;
const size_t CORES = work::Config::COMPUTATION_CAPACITY;
auto LOAD_BASE = 5ms;
TestChainLoad testLoad{NODES};
BlockFlowAlloc bFlow;
EngineObserver watch;
Scheduler scheduler{bFlow, watch};
auto LOAD_BASE = 10ms;
// auto stressFac = 1.0;
// auto concurrency = 8;
//
ComputationalLoad cpuLoad;
cpuLoad.timeBase = 200us;
cpuLoad.calibrate();
//
double loadMicros = cpuLoad.invoke();
// double refTime = testLoad.calcRuntimeReference(LOAD_BASE);
SHOW_EXPR(loadMicros)
auto testSetup =
testLoad.setupSchedule(scheduler)
.withLoadTimeBase(LOAD_BASE)
.withLevelDuration(10us)
.withJobDeadline(2s)
// .withBaseExpense(500us)
// .withUpfrontPlanning()
.withInstrumentation()
// .withAdaptedSchedule (stressFac, concurrency)
;
testLoad.setWeight(1)
.setupSchedule(scheduler)
.withLoadTimeBase(LOAD_BASE)
.withJobDeadline(50ms)
.withInstrumentation() // activate an instrumentation bracket around each job invocation
;
double runTime = testSetup.launch_and_wait();
double expected = testSetup.getExpectedEndTime();
SHOW_EXPR(runTime)
SHOW_EXPR(expected)
//SHOW_EXPR(refTime)
auto stat = testSetup.getInvocationStatistic();
SHOW_EXPR(stat.cumulatedTime);
SHOW_EXPR(stat.activeTime);
SHOW_EXPR(stat.coveredTime);
SHOW_EXPR(stat.eventCnt);
SHOW_EXPR(stat.activationCnt);
SHOW_EXPR(stat.cntCase(0));
SHOW_EXPR(stat.cntCase(1));
SHOW_EXPR(stat.cntCase(2));
SHOW_EXPR(stat.cntCase(3));
SHOW_EXPR(stat.timeCase(0));
SHOW_EXPR(stat.timeCase(1));
SHOW_EXPR(stat.timeCase(2));
SHOW_EXPR(stat.timeCase(3));
SHOW_EXPR(stat.cntThread(0));
SHOW_EXPR(stat.cntThread(1));
SHOW_EXPR(stat.cntThread(2));
SHOW_EXPR(stat.cntThread(3));
SHOW_EXPR(stat.cntThread(4));
SHOW_EXPR(stat.cntThread(5));
SHOW_EXPR(stat.cntThread(6));
SHOW_EXPR(stat.cntThread(7));
SHOW_EXPR(stat.cntThread(8));
SHOW_EXPR(stat.cntThread(9));
SHOW_EXPR(stat.timeThread(0));
SHOW_EXPR(stat.timeThread(1));
SHOW_EXPR(stat.timeThread(2));
SHOW_EXPR(stat.timeThread(3));
SHOW_EXPR(stat.timeThread(4));
SHOW_EXPR(stat.timeThread(5));
SHOW_EXPR(stat.timeThread(6));
SHOW_EXPR(stat.timeThread(7));
SHOW_EXPR(stat.timeThread(8));
SHOW_EXPR(stat.timeThread(9));
SHOW_EXPR(stat.avgConcurrency);
SHOW_EXPR(stat.timeAtConc(0));
SHOW_EXPR(stat.timeAtConc(1));
SHOW_EXPR(stat.timeAtConc(2));
SHOW_EXPR(stat.timeAtConc(3));
SHOW_EXPR(stat.timeAtConc(4));
SHOW_EXPR(stat.timeAtConc(5));
SHOW_EXPR(stat.timeAtConc(6));
SHOW_EXPR(stat.timeAtConc(7));
SHOW_EXPR(stat.timeAtConc(8));
SHOW_EXPR(stat.timeAtConc(9));
}
auto stat = testSetup.getInvocationStatistic(); // retrieve observed invocation statistics
CHECK (runTime < stat.activeTime);
CHECK (isLimited (4900, stat.activeTime/NODES, 8000)); // should be close to 5000
CHECK (stat.coveredTime < runTime);
CHECK (NODES == stat.activationCnt); // each node activated once
CHECK (isLimited (CORES/2, stat.avgConcurrency, CORES)); // should ideally come close to hardware concurrency
CHECK (0 == stat.timeAtConc(0));
CHECK (0 == stat.timeAtConc(CORES+1));
CHECK (runTime/2 < stat.timeAtConc(CORES-1)+stat.timeAtConc(CORES));
} // should ideally spend most of the time at highes concurrency levels

21
wiki/thinkPad.ichthyo.mm
View file

@ -111431,8 +111431,25 @@ Date:&#160;&#160;&#160;Thu Apr 20 18:53:17 2023 +0200<br/>
<icon BUILTIN="button_ok"/>
</node>
</node>
<node BACKGROUND_COLOR="#eee5c3" COLOR="#990000" CREATED="1707755273281" ID="ID_787410697" MODIFIED="1707755280011" TEXT="geeignet Dokumentieren (Test)">
<icon BUILTIN="flag-yellow"/>
<node COLOR="#338800" CREATED="1707755273281" ID="ID_787410697" MODIFIED="1708202930833" TEXT="geeignet Dokumentieren (Test)">
<icon BUILTIN="button_ok"/>
<node CREATED="1708202932752" ID="ID_21164866" MODIFIED="1708202938068" TEXT="schwierig..">
<icon BUILTIN="smily_bad"/>
</node>
<node CREATED="1708202940088" ID="ID_1078042444" MODIFIED="1708202953985" TEXT="kann zumindest einen einfachen Scheduler-Lauf machen"/>
<node CREATED="1708202954868" ID="ID_575048126" MODIFIED="1708203062489" TEXT="kann dann aber nicht sonderlich viel verifizieren">
<richcontent TYPE="NOTE"><html>
<head>
</head>
<body>
<p>
...weil keine Festlegung auf eine bestimmte Zahl an Cores m&#246;glich ist und auch die Werte stark statistisch schwanken, besonders bei den relativ kurzen Laufzeiten, die hier aus praktischen Gr&#252;nden erforderlich sind, schon um das Test-Setup einfach zu halten
</p>
</body>
</html>
</richcontent>
</node>
</node>
</node>
</node>