Improve NativeCallTargetDefinition marshalling #4326

tonyredondo · 2023-06-29T13:20:34Z

Summary of changes

This PR removes all arrays allocations for parameter types in NativeCallTargetDefinition structs, and also removes the multiple Marshall.Alloc* calls replacing it for a single allocation of a big segment of unmanaged memory that is freed at shutdown. ~~For netcoreapp we don't process the string at all, instead we use the ReadOnlySpan trick with the string already marshalled as UTF16 with no allocations.~~ (Reverted because it was taking more time than a normal unmanaged allocation + string marshalling, from 20ms to 1ms)

Reason for change

Currently we have 368 integrations being registered at startup, each integrations allocates an array object for the parameters types. Each parameter in this array calls allocates unmanaged memory to store the parameter type name. This creates overhead in the startup process that we are trying to reduce.

Test coverage

Tests has been updated with the new .ctor and source generator output.

tracer/src/Datadog.Trace/ClrProfiler/NativeCallTargetDefinition.cs

tracer/src/Datadog.Trace/ClrProfiler/NativeCallTargetDefinition2.cs

…llTargetDefinition strings.

…rshalling

kevingosse · 2023-07-03T14:04:42Z

tracer/src/Datadog.Trace/ClrProfiler/InstrumentationDefinitions.cs

-            {
-                def.Dispose();
-            }
+            Instrumentations = Array.Empty<NativeCallTargetDefinition2>();


I don't really mind, but note that using Array.Empty here is not beneficial. Because there will only ever be one of those empty arrays created, we still have the same number of allocations in total but we add one more generic type.

tracer/src/Datadog.Trace/ClrProfiler/NativeCallTargetUnmanagedMemoryHelper.cs

datadog-ddstaging · 2023-07-03T15:06:04Z

Datadog Report

Branch report: tony/improve-nativecalltargetdefinition-marshalling
Commit report: dc639a2

✅ dd-trace-dotnet: 0 Failed, 0 New Flaky, 369791 Passed, 1466 Skipped, 31m 38.4s Wall Time

andrewlock · 2023-07-03T16:46:18Z

Throughput/Crank Report:zap:

Throughput results for AspNetCoreSimpleController comparing the following branches/commits:

Cases where throughput results for the PR are worse than latest master (5% drop or greater), results are shown in red.

Note that these results are based on a single point-in-time result for each branch. For full results, see one of the many, many dashboards!

gantt
    title Throughput Linux x64 (Total requests) 
    dateFormat  X
    axisFormat %s
    section Baseline
    This PR (4326) (11.112M)   : 0, 11112274
    master (11.090M)   : 0, 11090030
    benchmarks/2.32.0 (11.375M)   : 0, 11374525
    benchmarks/2.9.0 (10.962M)   : 0, 10962162

    section Automatic
    This PR (4326) (7.734M)   : 0, 7734183
    master (7.769M)   : 0, 7769343
    benchmarks/2.32.0 (8.070M)   : 0, 8070318
    benchmarks/2.9.0 (7.961M)   : 0, 7961307

    section Trace stats
    master (7.770M)   : 0, 7769944
    benchmarks/2.32.0 (8.090M)   : 0, 8090308

    section Manual
    This PR (4326) (9.807M)   : 0, 9806681
    master (10.029M)   : 0, 10028784
    benchmarks/2.32.0 (10.358M)   : 0, 10357559

    section Manual + Automatic
    This PR (4326) (7.430M)   : 0, 7430236
    master (7.362M)   : 0, 7362416
    benchmarks/2.32.0 (7.698M)   : 0, 7697947

    section Version Conflict
    master (6.692M)   : 0, 6691750
    benchmarks/2.32.0 (6.996M)   : 0, 6995640

gantt
    title Throughput Linux arm64 (Total requests) 
    dateFormat  X
    axisFormat %s
    section Baseline
    This PR (4326) (9.768M)   : 0, 9767885
    master (9.720M)   : 0, 9720071
    benchmarks/2.32.0 (9.619M)   : 0, 9619353
    benchmarks/2.9.0 (9.727M)   : 0, 9726575

    section Automatic
    This PR (4326) (6.649M)   : 0, 6648941
    master (6.740M)   : 0, 6740171
    benchmarks/2.32.0 (6.682M)   : 0, 6681889

    section Trace stats
    master (6.648M)   : 0, 6647613
    benchmarks/2.32.0 (6.899M)   : 0, 6899182

    section Manual
    This PR (4326) (8.433M)   : 0, 8432810
    master (8.566M)   : 0, 8565697
    benchmarks/2.32.0 (8.545M)   : 0, 8545110

    section Manual + Automatic
    This PR (4326) (6.484M)   : 0, 6484162
    master (6.424M)   : 0, 6424293
    benchmarks/2.32.0 (6.707M)   : 0, 6707147

    section Version Conflict
    master (5.900M)   : 0, 5899730
    benchmarks/2.32.0 (6.004M)   : 0, 6004290

gantt
    title Throughput Windows x64 (Total requests) 
    dateFormat  X
    axisFormat %s
    section Baseline
    This PR (4326) (10.217M)   : 0, 10216532
    master (10.384M)   : 0, 10384309
    benchmarks/2.32.0 (10.272M)   : 0, 10272136
    benchmarks/2.9.0 (10.841M)   : 0, 10841230

    section Automatic
    This PR (4326) (7.879M)   : 0, 7878966
    master (7.421M)   : 0, 7421270
    benchmarks/2.32.0 (7.559M)   : 0, 7559361
    benchmarks/2.9.0 (7.862M)   : 0, 7862268

    section Trace stats
    master (7.549M)   : 0, 7548894
    benchmarks/2.32.0 (7.562M)   : 0, 7562409

    section Manual
    This PR (4326) (9.497M)   : 0, 9496554
    master (9.511M)   : 0, 9510968
    benchmarks/2.32.0 (9.530M)   : 0, 9530406

    section Manual + Automatic
    This PR (4326) (7.330M)   : 0, 7329708
    master (7.324M)   : 0, 7324257
    benchmarks/2.32.0 (7.279M)   : 0, 7279188

    section Version Conflict
    master (6.695M)   : 0, 6694521
    benchmarks/2.32.0 (6.827M)   : 0, 6827118

gantt
    title Throughput Linux x64 (ASM) (Total requests) 
    dateFormat  X
    axisFormat %s
    section Baseline
    This PR (4326) (7.689M)   : 0, 7688887
    master (7.539M)   : 0, 7538814
    benchmarks/2.32.0 (7.327M)   : 0, 7327005
    benchmarks/2.9.0 (7.934M)   : 0, 7933677

    section No attack
    This PR (4326) (2.186M)   : 0, 2186449
    master (2.195M)   : 0, 2195229
    benchmarks/2.32.0 (2.076M)   : 0, 2076414
    benchmarks/2.9.0 (3.287M)   : 0, 3286770

    section Attack
    This PR (4326) (1.885M)   : 0, 1884721
    master (1.882M)   : 0, 1882207
    benchmarks/2.32.0 (1.786M)   : 0, 1786036
    benchmarks/2.9.0 (2.600M)   : 0, 2600407

    section Blocking
    This PR (4326) (3.787M)   : 0, 3786962
    master (3.814M)   : 0, 3813723
    benchmarks/2.32.0 (3.490M)   : 0, 3490178

…rshalling

andrewlock · 2023-07-04T15:40:05Z

Execution-Time Benchmarks Report ⏱️

Execution-time results for samples comparing the following branches/commits:

Execution-time benchmarks measure the whole time it takes to execute a program. And are intended to measure the one-off costs. Cases where the execution time results for the PR are worse than latest master results are shown in red. The following thresholds were used for comparing the execution times:

Welch test with statistical test for significance of 5%
Only results indicating a difference greater than 5% and 5 ms are considered.

Note that these results are based on a single point-in-time result for each branch. For full results, see the dashboard.

Graphs show the p99 interval based on the mean and StdDev of the test run, as well as the mean value of the run (shown as a diamond below the graph).

gantt
    title Execution time (ms) FakeDbCommand (.NET Framework 4.6.2) 
    dateFormat  X
    axisFormat %s
    todayMarker off
    section Baseline
    This PR (4326) - mean (3,023ms)  : 2966, 3080
     .   : milestone, 3023,
    master - mean (3,020ms)  : 2943, 3097
     .   : milestone, 3020,

    section CallTarget+Inlining+NGEN
    This PR (4326) - mean (3,842ms)  : 3752, 3933
     .   : milestone, 3842,
    master - mean (3,809ms)  : 3754, 3864
     .   : milestone, 3809,

gantt
    title Execution time (ms) FakeDbCommand (.NET Core 3.1) 
    dateFormat  X
    axisFormat %s
    todayMarker off
    section Baseline
    This PR (4326) - mean (3,138ms)  : 3041, 3236
     .   : milestone, 3138,
    master - mean (3,143ms)  : 3061, 3226
     .   : milestone, 3143,

    section CallTarget+Inlining+NGEN
    This PR (4326) - mean (3,630ms)  : 3551, 3709
     .   : milestone, 3630,
    master - mean (3,647ms)  : 3575, 3719
     .   : milestone, 3647,

gantt
    title Execution time (ms) FakeDbCommand (.NET 6) 
    dateFormat  X
    axisFormat %s
    todayMarker off
    section Baseline
    This PR (4326) - mean (3,111ms)  : 2997, 3224
     .   : milestone, 3111,
    master - mean (3,121ms)  : 3006, 3236
     .   : milestone, 3121,

    section CallTarget+Inlining+NGEN
    This PR (4326) - mean (3,606ms)  : 3521, 3691
     .   : milestone, 3606,
    master - mean (3,610ms)  : 3533, 3687
     .   : milestone, 3610,

gantt
    title Execution time (ms) HttpMessageHandler (.NET Framework 4.6.2) 
    dateFormat  X
    axisFormat %s
    todayMarker off
    section Baseline
    This PR (4326) - mean (190ms)  : 188, 193
     .   : milestone, 190,
    master - mean (190ms)  : 186, 193
     .   : milestone, 190,

    section CallTarget+Inlining+NGEN
    This PR (4326) - mean (1,116ms)  : 1023, 1208
     .   : milestone, 1116,
    master - mean (1,092ms)  : 1070, 1114
     .   : milestone, 1092,

gantt
    title Execution time (ms) HttpMessageHandler (.NET Core 3.1) 
    dateFormat  X
    axisFormat %s
    todayMarker off
    section Baseline
    This PR (4326) - mean (372ms)  : 368, 377
     .   : milestone, 372,
    master - mean (370ms)  : 365, 375
     .   : milestone, 370,

    section CallTarget+Inlining+NGEN
    This PR (4326) - mean (1,171ms)  : 1136, 1205
     .   : milestone, 1171,
    master - mean (1,163ms)  : 1112, 1213
     .   : milestone, 1163,

gantt
    title Execution time (ms) HttpMessageHandler (.NET 6) 
    dateFormat  X
    axisFormat %s
    todayMarker off
    section Baseline
    This PR (4326) - mean (360ms)  : 354, 366
     .   : milestone, 360,
    master - mean (360ms)  : 354, 365
     .   : milestone, 360,

    section CallTarget+Inlining+NGEN
    This PR (4326) - mean (1,126ms)  : 1099, 1152
     .   : milestone, 1126,
    master - mean (1,122ms)  : 1086, 1159
     .   : milestone, 1122,

andrewlock · 2023-07-04T15:53:08Z

Benchmarks Report 🐌

Benchmarks for #4326 compared to master:

2 benchmarks are faster, with geometric mean 1.121
All benchmarks have the same allocations

The following thresholds were used for comparing the benchmark speeds:

Mann–Whitney U test with statistical test for significance of 5%
Only results indicating a difference greater than 10% and 0.3 ns are considered.

Allocation changes below 0.5% are ignored.

Benchmark details

Benchmarks.Trace.AgentWriterBenchmark - Same speed ✔️ Same allocations ✔️

Raw results

Branch	Method	Toolchain	Mean	StdError	StdDev	Gen 0	Allocated
master	`WriteAndFlushEnrichedTraces`	net6.0	469μs	297ns	1.15μs	0	2.62 KB
master	`WriteAndFlushEnrichedTraces`	netcoreapp3.1	646μs	285ns	1.1μs	0	2.62 KB
master	`WriteAndFlushEnrichedTraces`	net472	792μs	363ns	1.36μs	0.396	3.22 KB
#4326	`WriteAndFlushEnrichedTraces`	net6.0	481μs	144ns	558ns	0	2.62 KB
#4326	`WriteAndFlushEnrichedTraces`	netcoreapp3.1	650μs	181ns	653ns	0	2.63 KB
#4326	`WriteAndFlushEnrichedTraces`	net472	804μs	368ns	1.38μs	0.401	3.22 KB

Benchmarks.Trace.AppSecBodyBenchmark - Same speed ✔️ Same allocations ✔️

Raw results

Branch	Method	Toolchain	Mean	StdError	StdDev	Gen 0	Gen 1	Allocated
master	`AllCycleSimpleBody`	net6.0	39.1μs	34.9ns	135ns	0.0196	0	1.65 KB
master	`AllCycleSimpleBody`	netcoreapp3.1	41.5μs	55.1ns	199ns	0.0205	0	1.63 KB
master	`AllCycleSimpleBody`	net472	41μs	8.68ns	33.6ns	0.266	0	1.69 KB
master	`AllCycleMoreComplexBody`	net6.0	220μs	122ns	471ns	0.109	0	9.22 KB
master	`AllCycleMoreComplexBody`	netcoreapp3.1	228μs	453ns	1.63μs	0.115	0	9.12 KB
master	`AllCycleMoreComplexBody`	net472	238μs	53.6ns	208ns	1.42	0	9.28 KB
master	`ObjectExtractorSimpleBody`	net6.0	126ns	0.0824ns	0.308ns	0.00397	0	280 B
master	`ObjectExtractorSimpleBody`	netcoreapp3.1	174ns	0.227ns	0.878ns	0.00375	0	272 B
master	`ObjectExtractorSimpleBody`	net472	145ns	0.116ns	0.435ns	0.0446	0	281 B
master	`ObjectExtractorMoreComplexBody`	net6.0	2.98μs	0.964ns	3.73ns	0.0537	0	3.88 KB
master	`ObjectExtractorMoreComplexBody`	netcoreapp3.1	4.35μs	2.28ns	8.82ns	0.0521	0	3.78 KB
master	`ObjectExtractorMoreComplexBody`	net472	4.17μs	2.05ns	7.95ns	0.618	0.00624	3.89 KB
#4326	`AllCycleSimpleBody`	net6.0	39.3μs	33.6ns	130ns	0.0192	0	1.65 KB
#4326	`AllCycleSimpleBody`	netcoreapp3.1	41.9μs	208ns	858ns	0.0199	0	1.63 KB
#4326	`AllCycleSimpleBody`	net472	41.1μs	51.9ns	180ns	0.266	0	1.69 KB
#4326	`AllCycleMoreComplexBody`	net6.0	221μs	64.6ns	242ns	0.11	0	9.22 KB
#4326	`AllCycleMoreComplexBody`	netcoreapp3.1	229μs	241ns	903ns	0.113	0	9.12 KB
#4326	`AllCycleMoreComplexBody`	net472	238μs	64.1ns	248ns	1.43	0	9.28 KB
#4326	`ObjectExtractorSimpleBody`	net6.0	122ns	0.0392ns	0.136ns	0.00395	0	280 B
#4326	`ObjectExtractorSimpleBody`	netcoreapp3.1	170ns	0.318ns	1.23ns	0.0037	0	272 B
#4326	`ObjectExtractorSimpleBody`	net472	149ns	0.176ns	0.68ns	0.0446	0	281 B
#4326	`ObjectExtractorMoreComplexBody`	net6.0	2.97μs	1.15ns	4.45ns	0.055	0	3.88 KB
#4326	`ObjectExtractorMoreComplexBody`	netcoreapp3.1	4.14μs	3.05ns	11ns	0.0518	0	3.78 KB
#4326	`ObjectExtractorMoreComplexBody`	net472	4.24μs	2.29ns	8.89ns	0.617	0.00636	3.89 KB

Benchmarks.Trace.AspNetCoreBenchmark - Same speed ✔️ Same allocations ✔️

Raw results

Branch	Method	Toolchain	Mean	StdError	StdDev	Gen 0	Allocated
master	`SendRequest`	net6.0	167μs	102ns	380ns	0.167	18.03 KB
master	`SendRequest`	netcoreapp3.1	185μs	299ns	1.16μs	0.185	20.09 KB
master	`SendRequest`	net472	0ns	0ns	0ns	0	0 b
#4326	`SendRequest`	net6.0	166μs	300ns	1.16μs	0.248	18.03 KB
#4326	`SendRequest`	netcoreapp3.1	190μs	226ns	846ns	0.187	20.09 KB
#4326	`SendRequest`	net472	0.00115ns	0.000376ns	0.00146ns	0	0 b

Benchmarks.Trace.CIVisibilityProtocolWriterBenchmark - Same speed ✔️ Same allocations ✔️

Raw results

Branch	Method	Toolchain	Mean	StdError	StdDev	Gen 0	Gen 1	Gen 2	Allocated
master	`WriteAndFlushEnrichedTraces`	net6.0	514μs	410ns	1.59μs	0.5	0	0	41.49 KB
master	`WriteAndFlushEnrichedTraces`	netcoreapp3.1	631μs	1.18μs	4.27μs	0.312	0	0	41.65 KB
master	`WriteAndFlushEnrichedTraces`	net472	796μs	3.9μs	16.1μs	8.25	2.36	0.393	53.25 KB
#4326	`WriteAndFlushEnrichedTraces`	net6.0	528μs	454ns	1.76μs	0.765	0	0	41.67 KB
#4326	`WriteAndFlushEnrichedTraces`	netcoreapp3.1	623μs	1.58μs	6.13μs	0.319	0	0	41.54 KB
#4326	`WriteAndFlushEnrichedTraces`	net472	778μs	2.47μs	9.55μs	8.1	2.31	0.386	53.25 KB

Benchmarks.Trace.DbCommandBenchmark - Faster 🎉 Same allocations ✔️

Faster 🎉 in #4326

Benchmark	base/diff	Base Median (ns)	Diff Median (ns)	Modality
Benchmarks.Trace.DbCommandBenchmark.ExecuteNonQuery‑netcoreapp3.1	1.128	1,260.16	1,116.71

Raw results

Branch	Method	Toolchain	Mean	StdError	StdDev	Gen 0	Gen 1	Allocated
master	`ExecuteNonQuery`	net6.0	891ns	0.473ns	1.64ns	0.00988	0	720 B
master	`ExecuteNonQuery`	netcoreapp3.1	1.26μs	0.355ns	1.37ns	0.00942	0	720 B
master	`ExecuteNonQuery`	net472	1.34μs	0.625ns	2.42ns	0.108	0	682 B
#4326	`ExecuteNonQuery`	net6.0	912ns	0.324ns	1.25ns	0.0101	0	720 B
#4326	`ExecuteNonQuery`	netcoreapp3.1	1.12μs	0.907ns	3.51ns	0.0098	0	720 B
#4326	`ExecuteNonQuery`	net472	1.41μs	0.684ns	2.65ns	0.108	0.000706	682 B

Benchmarks.Trace.ElasticsearchBenchmark - Same speed ✔️ Same allocations ✔️

Raw results

Branch	Method	Toolchain	Mean	StdError	StdDev	Gen 0	Gen 1	Allocated
master	`CallElasticsearch`	net6.0	1.06μs	0.788ns	2.84ns	0.0132	0	944 B
master	`CallElasticsearch`	netcoreapp3.1	1.35μs	0.444ns	1.72ns	0.0128	0	944 B
master	`CallElasticsearch`	net472	1.96μs	0.441ns	1.65ns	0.151	0.000982	955 B
master	`CallElasticsearchAsync`	net6.0	998ns	0.374ns	1.4ns	0.013	0	920 B
master	`CallElasticsearchAsync`	netcoreapp3.1	1.31μs	0.471ns	1.7ns	0.0132	0	992 B
master	`CallElasticsearchAsync`	net472	2.16μs	0.679ns	2.45ns	0.16	0.00108	1.01 KB
#4326	`CallElasticsearch`	net6.0	973ns	0.2ns	0.749ns	0.0133	0	944 B
#4326	`CallElasticsearch`	netcoreapp3.1	1.32μs	0.599ns	2.32ns	0.0125	0	944 B
#4326	`CallElasticsearch`	net472	1.96μs	0.327ns	1.18ns	0.152	0.000979	955 B
#4326	`CallElasticsearchAsync`	net6.0	1.02μs	0.421ns	1.63ns	0.0129	0	920 B
#4326	`CallElasticsearchAsync`	netcoreapp3.1	1.34μs	0.915ns	3.42ns	0.0134	0	992 B
#4326	`CallElasticsearchAsync`	net472	2.16μs	0.299ns	1.16ns	0.16	0	1.01 KB

Benchmarks.Trace.GraphQLBenchmark - Same speed ✔️ Same allocations ✔️

Raw results

Branch	Method	Toolchain	Mean	StdError	StdDev	Gen 0	Gen 1	Allocated
master	`ExecuteAsync`	net6.0	1.04μs	0.4ns	1.5ns	0.0128	0	912 B
master	`ExecuteAsync`	netcoreapp3.1	1.33μs	0.338ns	1.26ns	0.0122	0	912 B
master	`ExecuteAsync`	net472	1.51μs	0.612ns	2.29ns	0.138	0.000756	875 B
#4326	`ExecuteAsync`	net6.0	1.06μs	4.52ns	20.2ns	0.0127	0	912 B
#4326	`ExecuteAsync`	netcoreapp3.1	1.36μs	0.723ns	2.8ns	0.0123	0	912 B
#4326	`ExecuteAsync`	net472	1.52μs	0.552ns	2.14ns	0.139	0	875 B

Benchmarks.Trace.HttpClientBenchmark - Same speed ✔️ Same allocations ✔️

Raw results

Branch	Method	Toolchain	Mean	StdError	StdDev	Gen 0	Allocated
master	`SendAsync`	net6.0	3.73μs	1.68ns	6.51ns	0.026	1.9 KB
master	`SendAsync`	netcoreapp3.1	4.39μs	1.67ns	6.46ns	0.0329	2.43 KB
master	`SendAsync`	net472	6.86μs	12.5ns	46.8ns	0.476	3 KB
#4326	`SendAsync`	net6.0	3.47μs	1.35ns	5.21ns	0.0259	1.9 KB
#4326	`SendAsync`	netcoreapp3.1	4.17μs	1.77ns	6.63ns	0.0333	2.43 KB
#4326	`SendAsync`	net472	6.84μs	2.36ns	8.83ns	0.476	3 KB

Benchmarks.Trace.ILoggerBenchmark - Same speed ✔️ Same allocations ✔️

Raw results

Branch	Method	Toolchain	Mean	StdError	StdDev	Gen 0	Allocated
master	`EnrichedLog`	net6.0	1.29μs	0.651ns	2.52ns	0.0226	1.62 KB
master	`EnrichedLog`	netcoreapp3.1	2.06μs	0.808ns	3.13ns	0.0219	1.62 KB
master	`EnrichedLog`	net472	2.23μs	1.09ns	4.09ns	0.244	1.54 KB
#4326	`EnrichedLog`	net6.0	1.4μs	0.805ns	3.01ns	0.0224	1.62 KB
#4326	`EnrichedLog`	netcoreapp3.1	1.92μs	0.644ns	2.41ns	0.0221	1.62 KB
#4326	`EnrichedLog`	net472	2.4μs	2.79ns	10.8ns	0.245	1.54 KB

Benchmarks.Trace.Log4netBenchmark - Same speed ✔️ Same allocations ✔️

Raw results

Branch	Method	Toolchain	Mean	StdError	StdDev	Gen 0	Gen 1	Allocated
master	`EnrichedLog`	net6.0	111μs	138ns	515ns	0.0551	0	4.21 KB
master	`EnrichedLog`	netcoreapp3.1	117μs	275ns	1.06μs	0.0582	0	4.21 KB
master	`EnrichedLog`	net472	148μs	235ns	911ns	0.657	0.219	4.38 KB
#4326	`EnrichedLog`	net6.0	112μs	183ns	711ns	0.0564	0	4.21 KB
#4326	`EnrichedLog`	netcoreapp3.1	119μs	142ns	550ns	0	0	4.21 KB
#4326	`EnrichedLog`	net472	149μs	104ns	404ns	0.668	0.223	4.38 KB

Benchmarks.Trace.NLogBenchmark - Same speed ✔️ Same allocations ✔️

Raw results

Branch	Method	Toolchain	Mean	StdError	StdDev	Gen 0	Gen 1	Allocated
master	`EnrichedLog`	net6.0	3.13μs	0.89ns	3.33ns	0.0516	0	3.71 KB
master	`EnrichedLog`	netcoreapp3.1	3.95μs	1.44ns	5.58ns	0.0506	0	3.69 KB
master	`EnrichedLog`	net472	4.75μs	2.2ns	7.93ns	0.526	0.00237	3.31 KB
#4326	`EnrichedLog`	net6.0	3.16μs	0.839ns	3.14ns	0.052	0	3.71 KB
#4326	`EnrichedLog`	netcoreapp3.1	3.87μs	1.08ns	4.04ns	0.0502	0	3.69 KB
#4326	`EnrichedLog`	net472	4.76μs	1.45ns	5.63ns	0.525	0.0024	3.31 KB

Benchmarks.Trace.RedisBenchmark - Same speed ✔️ Same allocations ✔️

Raw results

Branch	Method	Toolchain	Mean	StdError	StdDev	Gen 0	Gen 1	Allocated
master	`SendReceive`	net6.0	1.11μs	0.883ns	3.3ns	0.0155	0	1.11 KB
master	`SendReceive`	netcoreapp3.1	1.38μs	1.08ns	4.02ns	0.0152	0	1.11 KB
master	`SendReceive`	net472	1.72μs	1.3ns	4.88ns	0.177	0.000855	1.12 KB
#4326	`SendReceive`	net6.0	1.15μs	0.482ns	1.8ns	0.0156	0	1.11 KB
#4326	`SendReceive`	netcoreapp3.1	1.44μs	0.548ns	2.05ns	0.015	0	1.11 KB
#4326	`SendReceive`	net472	1.84μs	2.19ns	8.49ns	0.177	0	1.12 KB

Benchmarks.Trace.SerilogBenchmark - Same speed ✔️ Same allocations ✔️

Raw results

Branch	Method	Toolchain	Mean	StdError	StdDev	Gen 0	Allocated
master	`EnrichedLog`	net6.0	2.58μs	0.865ns	3.35ns	0.0211	1.53 KB
master	`EnrichedLog`	netcoreapp3.1	3.64μs	1.14ns	4.1ns	0.0199	1.58 KB
master	`EnrichedLog`	net472	4.18μs	1.51ns	5.65ns	0.309	1.96 KB
#4326	`EnrichedLog`	net6.0	2.6μs	0.524ns	1.96ns	0.0207	1.53 KB
#4326	`EnrichedLog`	netcoreapp3.1	3.63μs	1.55ns	5.59ns	0.02	1.58 KB
#4326	`EnrichedLog`	net472	4.09μs	1.5ns	5.82ns	0.311	1.96 KB

Benchmarks.Trace.SpanBenchmark - Same speed ✔️ Same allocations ✔️

Raw results

Branch	Method	Toolchain	Mean	StdError	StdDev	Gen 0	Allocated
master	`StartFinishSpan`	net6.0	396ns	0.15ns	0.56ns	0.00748	536 B
master	`StartFinishSpan`	netcoreapp3.1	478ns	0.19ns	0.685ns	0.00724	536 B
master	`StartFinishSpan`	net472	643ns	0.205ns	0.794ns	0.0852	538 B
master	`StartFinishScope`	net6.0	538ns	0.127ns	0.456ns	0.0093	656 B
master	`StartFinishScope`	netcoreapp3.1	630ns	0.214ns	0.828ns	0.00907	656 B
master	`StartFinishScope`	net472	898ns	0.554ns	2.14ns	0.0981	618 B
#4326	`StartFinishSpan`	net6.0	377ns	0.135ns	0.486ns	0.00753	536 B
#4326	`StartFinishSpan`	netcoreapp3.1	484ns	0.206ns	0.798ns	0.00724	536 B
#4326	`StartFinishSpan`	net472	612ns	0.232ns	0.898ns	0.0853	538 B
#4326	`StartFinishScope`	net6.0	536ns	0.131ns	0.491ns	0.00922	656 B
#4326	`StartFinishScope`	netcoreapp3.1	673ns	0.388ns	1.5ns	0.00911	656 B
#4326	`StartFinishScope`	net472	831ns	0.371ns	1.44ns	0.0978	618 B

Benchmarks.Trace.TraceAnnotationsBenchmark - Faster 🎉 Same allocations ✔️

Faster 🎉 in #4326

Benchmark	base/diff	Base Median (ns)	Diff Median (ns)	Modality
Benchmarks.Trace.TraceAnnotationsBenchmark.RunOnMethodBegin‑net6.0	1.114	608.85	546.74

Raw results

Branch	Method	Toolchain	Mean	StdError	StdDev	Gen 0	Allocated
master	`RunOnMethodBegin`	net6.0	609ns	0.263ns	0.986ns	0.00915	656 B
master	`RunOnMethodBegin`	netcoreapp3.1	858ns	0.349ns	1.21ns	0.00905	656 B
master	`RunOnMethodBegin`	net472	911ns	0.232ns	0.868ns	0.0982	618 B
#4326	`RunOnMethodBegin`	net6.0	547ns	0.174ns	0.651ns	0.00911	656 B
#4326	`RunOnMethodBegin`	netcoreapp3.1	803ns	0.625ns	2.34ns	0.0088	656 B
#4326	`RunOnMethodBegin`	net472	993ns	0.301ns	1.13ns	0.0979	618 B

tonyredondo added 2 commits June 29, 2023 14:28

Improve NativeCallTargetDefinition Marshalling

7d7d551

changes

2796771

tonyredondo self-assigned this Jun 29, 2023

github-advanced-security bot found potential problems Jun 29, 2023

View reviewed changes

tracer/src/Datadog.Trace/ClrProfiler/NativeCallTargetDefinition.cs Fixed Show fixed Hide fixed

tracer/src/Datadog.Trace/ClrProfiler/NativeCallTargetDefinition2.cs Fixed Show fixed Hide fixed

tonyredondo added 3 commits June 29, 2023 17:30

Introducing the UnmanagedMemorySegment class to allocate all NativeCa…

c8c23b3

…llTargetDefinition strings.

Merge branch 'master' into tony/improve-nativecalltargetdefinition-ma…

b0de67e

…rshalling

Fix

cf7ac2f

DataDog deleted a comment from datadog-ddstaging bot Jun 29, 2023

DataDog deleted a comment from andrewlock Jun 29, 2023

tonyredondo added 4 commits June 29, 2023 18:18

Merge branch 'master' into tony/improve-nativecalltargetdefinition-ma…

aad97dc

…rshalling

Merge branch 'master' into tony/improve-nativecalltargetdefinition-ma…

b8b7ae4

…rshalling

Changes

6cf20ff

Remove array allocations for parameters types

fbc612d

DataDog deleted a comment from datadog-ddstaging bot Jun 30, 2023

DataDog deleted a comment from andrewlock Jun 30, 2023

tonyredondo added 2 commits June 30, 2023 11:47

fix tests

19aad48

Merge branch 'master' into tony/improve-nativecalltargetdefinition-ma…

0b45ff0

…rshalling

DataDog deleted a comment from datadog-ddstaging bot Jun 30, 2023

DataDog deleted a comment from andrewlock Jun 30, 2023

tonyredondo added 2 commits June 30, 2023 11:55

changes

166df3b

Changes

a4932ad

tonyredondo changed the title ~~Tony/improve nativecalltargetdefinition marshalling~~ Improve NativeCallTargetDefinition marshalling Jun 30, 2023

changes

ec39029

DataDog deleted a comment from andrewlock Jul 3, 2023

Net Core optimization.

d7c130b

DataDog deleted a comment from andrewlock Jul 3, 2023

tonyredondo added 2 commits July 3, 2023 16:35

Reverted some changes

9c081b5

Remove unused code

f03bc2b

DataDog deleted a comment from datadog-ddstaging bot Jul 3, 2023

DataDog deleted a comment from andrewlock Jul 3, 2023

Merge branch 'master' into tony/improve-nativecalltargetdefinition-ma…

823204c

…rshalling

kevingosse reviewed Jul 3, 2023

View reviewed changes

This comment has been minimized.

Sign in to view

Changes from review

74dbe71

kevingosse approved these changes Jul 3, 2023

View reviewed changes

This comment has been minimized.

Sign in to view

tonyredondo added 2 commits July 4, 2023 16:57

Merge branch 'master' into tony/improve-nativecalltargetdefinition-ma…

73805f2

…rshalling

fixes

dc639a2

tonyredondo merged commit 28352a5 into master Jul 4, 2023

tonyredondo deleted the tony/improve-nativecalltargetdefinition-marshalling branch July 4, 2023 17:40

github-actions bot added this to the vNext milestone Jul 4, 2023

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Improve NativeCallTargetDefinition marshalling #4326

Improve NativeCallTargetDefinition marshalling #4326

tonyredondo commented Jun 29, 2023 •

edited

Loading

kevingosse Jul 3, 2023

datadog-ddstaging bot commented Jul 3, 2023 •

edited

Loading

This comment has been minimized.

This comment has been minimized.

This comment has been minimized.

This comment has been minimized.

andrewlock commented Jul 3, 2023

andrewlock commented Jul 4, 2023

andrewlock commented Jul 4, 2023

Raw results

Raw results

Raw results

Raw results

Faster 🎉 in #4326

Raw results

Raw results

Raw results

Raw results

Raw results

Raw results

Raw results

Raw results

Raw results

Raw results

Faster 🎉 in #4326

Raw results

Improve NativeCallTargetDefinition marshalling #4326

Improve NativeCallTargetDefinition marshalling #4326

Conversation

tonyredondo commented Jun 29, 2023 • edited Loading

Summary of changes

Reason for change

Test coverage

kevingosse Jul 3, 2023

Choose a reason for hiding this comment

datadog-ddstaging bot commented Jul 3, 2023 • edited Loading

Datadog Report

This comment has been minimized.

This comment has been minimized.

This comment has been minimized.

This comment has been minimized.

andrewlock commented Jul 3, 2023

Throughput/Crank Report:zap:

andrewlock commented Jul 4, 2023

Execution-Time Benchmarks Report ⏱️

andrewlock commented Jul 4, 2023

Benchmarks Report 🐌

Benchmark details

Raw results

Raw results

Raw results

Raw results

Faster 🎉 in #4326

Raw results

Raw results

Raw results

Raw results

Raw results

Raw results

Raw results

Raw results

Raw results

Raw results

Faster 🎉 in #4326

Raw results

tonyredondo commented Jun 29, 2023 •

edited

Loading

datadog-ddstaging bot commented Jul 3, 2023 •

edited

Loading