v0.2.0 benchmark results

[uatuko]

Benchmarks

Scenarios

• 1a: Single client with no concurrent streams
• 1b: Single client with 10 concurrent streams
• 2a: 10 clients with no concurrent streams
• 2b: 10 clients with 10 concurrent streams
• 3a: 100 clients with no concurrent streams
• 3b: 100 clients with 10 concurrent streams

Results (req/s)

	1a	1b	2a	2b	3a	3b
single-threaded (db137df)	42k	193k	120k	449k	160k	452k
multi-threaded, 2 workers (db137df)	33k	166k	120k	431k	91k	425k
multi-threaded, hardware concurrency (db137df)	34k	164k	118k	450k	101k	408k

Benchmark details

Benchmarks were run on a MacBook Pro 2021 (M1 Max, 32GB), macOS 14.2.1 (23C71) against db137df.

Older benchmark results can be found in #2 and #20.

[uatuko]

Single-threaded

single-threaded

diff --git a/examples/helloworld/main.cpp b/examples/helloworld/main.cpp
index debc569..758fae3 100644
--- a/examples/helloworld/main.cpp
+++ b/examples/helloworld/main.cpp
@@ -35,7 +35,7 @@ int main() {
 	GreeterImpl greeter;
 	Service     service(greeter);
 
-	grpcxx::server server;
+	grpcxx::server server(0);
 	server.add(service);
 
 	std::printf("Listening on [127.0.0.1:7000] ...\n");

1a

❯ h2load --clients=1 --duration=3 \
  --header='Content-Type: application/grpc' \
  --data=examples/helloworld/testdata/hello.grpc-lpm.data \
  http://localhost:7000/helloworld.v1.Greeter/Hello
starting benchmark...
spawning thread #0: 1 total client(s). Timing-based test with 0s of warm-up time and 3s of main duration for measurements.
Warm-up started for thread #0.
progress: 100% of clients started
Warm-up phase is over for thread #0.
Main benchmark duration is started for thread #0.
Application protocol: h2c
Main benchmark duration is over for thread #0. Stopping all clients.
Stopped all clients for thread #0

finished in 3.00s, 42407.33 req/s, 2.47MB/s
requests: 127222 total, 127223 started, 127222 done, 127222 succeeded, 0 failed, 0 errored, 0 timeout
status codes: 127222 2xx, 0 3xx, 0 4xx, 0 5xx
traffic: 7.40MB (7761486) total, 248.49KB (254456) headers (space savings 94.74%), 3.76MB (3943882) data
                     min         max         mean         sd        +/- sd
time for request:       15us       301us        21us         5us    93.05%
time for connect:      339us       339us       339us         0us   100.00%
time to 1st byte:      633us       633us       633us         0us   100.00%
req/s           :   42406.41    42406.41    42406.41        0.00   100.00%

1b

❯ h2load --clients=1 --max-concurrent-streams=10 --duration=3 \
  --header='Content-Type: application/grpc' \
  --data=examples/helloworld/testdata/hello.grpc-lpm.data \
  http://localhost:7000/helloworld.v1.Greeter/Hello
starting benchmark...
spawning thread #0: 1 total client(s). Timing-based test with 0s of warm-up time and 3s of main duration for measurements.
Warm-up started for thread #0.
progress: 100% of clients started
Warm-up phase is over for thread #0.
Main benchmark duration is started for thread #0.
Application protocol: h2c
Main benchmark duration is over for thread #0. Stopping all clients.
Stopped all clients for thread #0

finished in 3.00s, 193210.00 req/s, 11.24MB/s
requests: 579630 total, 579640 started, 579630 done, 579630 succeeded, 0 failed, 0 errored, 0 timeout
status codes: 579630 2xx, 0 3xx, 0 4xx, 0 5xx
traffic: 33.72MB (35361611) total, 1.11MB (1159272) headers (space savings 94.74%), 17.14MB (17968530) data
                     min         max         mean         sd        +/- sd
time for request:       37us       419us        46us         5us    83.31%
time for connect:       30us        30us        30us         0us   100.00%
time to 1st byte:      135us       135us       135us         0us   100.00%
req/s           :  193206.21   193206.21   193206.21        0.00   100.00%

2a

❯ h2load --clients=10 --duration=3 \
  --header='Content-Type: application/grpc' \
  --data=examples/helloworld/testdata/hello.grpc-lpm.data \
  http://localhost:7000/helloworld.v1.Greeter/Hello
starting benchmark...
spawning thread #0: 10 total client(s). Timing-based test with 0s of warm-up time and 3s of main duration for measurements.
Warm-up started for thread #0.
progress: 10% of clients started
progress: 20% of clients started
progress: 30% of clients started
progress: 40% of clients started
progress: 50% of clients started
progress: 60% of clients started
progress: 70% of clients started
progress: 80% of clients started
progress: 90% of clients started
progress: 100% of clients started
Warm-up phase is over for thread #0.
Main benchmark duration is started for thread #0.
Application protocol: h2c
Main benchmark duration is over for thread #0. Stopping all clients.
Stopped all clients for thread #0

finished in 3.00s, 120498.67 req/s, 7.01MB/s
requests: 361496 total, 361506 started, 361496 done, 361496 succeeded, 0 failed, 0 errored, 0 timeout
status codes: 361496 2xx, 0 3xx, 0 4xx, 0 5xx
traffic: 21.03MB (22054196) total, 706.16KB (723112) headers (space savings 94.74%), 10.69MB (11206376) data
                     min         max         mean         sd        +/- sd
time for request:       14us       220us        70us        21us    72.21%
time for connect:       21us       109us        49us        31us    80.00%
time to 1st byte:      164us       263us       211us        33us    50.00%
req/s           :   12035.92    12075.06    12049.48       10.80    80.00%

2b

❯ h2load --clients=10 --max-concurrent-streams=10 --duration=3 \
  --header='Content-Type: application/grpc' \
  --data=examples/helloworld/testdata/hello.grpc-lpm.data \
  http://localhost:7000/helloworld.v1.Greeter/Hello
starting benchmark...
spawning thread #0: 10 total client(s). Timing-based test with 0s of warm-up time and 3s of main duration for measurements.
Warm-up started for thread #0.
progress: 10% of clients started
progress: 20% of clients started
progress: 30% of clients started
progress: 40% of clients started
progress: 50% of clients started
progress: 60% of clients started
progress: 70% of clients started
progress: 80% of clients started
progress: 90% of clients started
progress: 100% of clients started
Warm-up phase is over for thread #0.
Main benchmark duration is started for thread #0.
Application protocol: h2c
Main benchmark duration is over for thread #0. Stopping all clients.
Stopped all clients for thread #0

finished in 3.00s, 449450.00 req/s, 26.15MB/s
requests: 1348350 total, 1348450 started, 1348350 done, 1348350 succeeded, 0 failed, 0 errored, 0 timeout
status codes: 1348350 2xx, 0 3xx, 0 4xx, 0 5xx
traffic: 78.45MB (82259375) total, 2.57MB (2696820) headers (space savings 94.74%), 39.86MB (41798850) data
                     min         max         mean         sd        +/- sd
time for request:       56us       581us       209us        73us    61.50%
time for connect:       22us       219us        59us        57us    90.00%
time to 1st byte:      246us       524us       360us        97us    70.00%
req/s           :   43781.72    47234.24    44943.32      971.42    80.00%

3a

❯ h2load --clients=100 --duration=3 \
  --header='Content-Type: application/grpc' \
  --data=examples/helloworld/testdata/hello.grpc-lpm.data \
  http://localhost:7000/helloworld.v1.Greeter/Hello
starting benchmark...
spawning thread #0: 100 total client(s). Timing-based test with 0s of warm-up time and 3s of main duration for measurements.
Warm-up started for thread #0.
progress: 10% of clients started
progress: 20% of clients started
progress: 30% of clients started
progress: 40% of clients started
progress: 50% of clients started
progress: 60% of clients started
progress: 70% of clients started
progress: 80% of clients started
progress: 90% of clients started
progress: 100% of clients started
Warm-up phase is over for thread #0.
Main benchmark duration is started for thread #0.
Application protocol: h2c
Main benchmark duration is over for thread #0. Stopping all clients.
Stopped all clients for thread #0

finished in 3.00s, 160147.67 req/s, 9.32MB/s
requests: 480443 total, 480543 started, 480443 done, 480443 succeeded, 0 failed, 0 errored, 0 timeout
status codes: 480443 2xx, 0 3xx, 0 4xx, 0 5xx
traffic: 27.96MB (29314323) total, 939.54KB (962086) headers (space savings 94.73%), 14.20MB (14893733) data
                     min         max         mean         sd        +/- sd
time for request:       53us      2.20ms       421us       147us    77.20%
time for connect:       23us       317us       169us        86us    57.00%
time to 1st byte:      911us      1.60ms      1.13ms       216us    78.00%
req/s           :    1592.89     1607.40     1601.18        2.77    65.00%

3b

❯ h2load --clients=100 --max-concurrent-streams=10 --duration=3 \
  --header='Content-Type: application/grpc' \
  --data=examples/helloworld/testdata/hello.grpc-lpm.data \
  http://localhost:7000/helloworld.v1.Greeter/Hello
starting benchmark...
spawning thread #0: 100 total client(s). Timing-based test with 0s of warm-up time and 3s of main duration for measurements.
Warm-up started for thread #0.
progress: 10% of clients started
progress: 20% of clients started
progress: 30% of clients started
progress: 40% of clients started
progress: 50% of clients started
progress: 60% of clients started
progress: 70% of clients started
progress: 80% of clients started
progress: 90% of clients started
progress: 100% of clients started
Warm-up phase is over for thread #0.
Main benchmark duration is started for thread #0.
Application protocol: h2c
Main benchmark duration is over for thread #0. Stopping all clients.
Stopped all clients for thread #0

finished in 3.00s, 451906.67 req/s, 26.29MB/s
requests: 1355720 total, 1356720 started, 1355720 done, 1355720 succeeded, 0 failed, 0 errored, 0 timeout
status codes: 1355720 2xx, 0 3xx, 0 4xx, 0 5xx
traffic: 78.88MB (82712473) total, 2.59MB (2712640) headers (space savings 94.73%), 40.08MB (42027320) data
                     min         max         mean         sd        +/- sd
time for request:      265us      5.20ms      2.20ms       367us    96.24%
time for connect:       22us       494us       259us       137us    58.00%
time to 1st byte:     1.92ms      3.92ms      2.63ms       585us    50.00%
req/s           :    2756.53     4652.13     4518.36      390.41    90.00%

[uatuko]

Multi-threaded, 2 workers

2 workers

diff --git a/examples/helloworld/main.cpp b/examples/helloworld/main.cpp
index debc569..a04f98d 100644
--- a/examples/helloworld/main.cpp
+++ b/examples/helloworld/main.cpp
@@ -35,7 +35,7 @@ int main() {
 	GreeterImpl greeter;
 	Service     service(greeter);
 
-	grpcxx::server server;
+	grpcxx::server server(2);
 	server.add(service);
 
 	std::printf("Listening on [127.0.0.1:7000] ...\n");

1a

❯ h2load --clients=1 --duration=3 \
  --header='Content-Type: application/grpc' \
  --data=examples/helloworld/testdata/hello.grpc-lpm.data \
  http://localhost:7000/helloworld.v1.Greeter/Hello
starting benchmark...
spawning thread #0: 1 total client(s). Timing-based test with 0s of warm-up time and 3s of main duration for measurements.
Warm-up started for thread #0.
progress: 100% of clients started
Warm-up phase is over for thread #0.
Main benchmark duration is started for thread #0.
Application protocol: h2c
Main benchmark duration is over for thread #0. Stopping all clients.
Stopped all clients for thread #0

finished in 3.00s, 33163.67 req/s, 1.93MB/s
requests: 99491 total, 99492 started, 99491 done, 99491 succeeded, 0 failed, 0 errored, 0 timeout
status codes: 99491 2xx, 0 3xx, 0 4xx, 0 5xx
traffic: 5.79MB (6069700) total, 194.33KB (198994) headers (space savings 94.74%), 2.94MB (3084221) data
                     min         max         mean         sd        +/- sd
time for request:       20us       174us        28us         4us    89.53%
time for connect:      120us       120us       120us         0us   100.00%
time to 1st byte:      249us       249us       249us         0us   100.00%
req/s           :   33163.06    33163.06    33163.06        0.00   100.00%

1b

❯ h2load --clients=1 --max-concurrent-streams=10 --duration=3 \
  --header='Content-Type: application/grpc' \
  --data=examples/helloworld/testdata/hello.grpc-lpm.data \
  http://localhost:7000/helloworld.v1.Greeter/Hello
starting benchmark...
spawning thread #0: 1 total client(s). Timing-based test with 0s of warm-up time and 3s of main duration for measurements.
Warm-up started for thread #0.
progress: 100% of clients started
Warm-up phase is over for thread #0.
Main benchmark duration is started for thread #0.
Application protocol: h2c
Main benchmark duration is over for thread #0. Stopping all clients.
Stopped all clients for thread #0

finished in 3.00s, 165506.67 req/s, 9.63MB/s
requests: 496520 total, 496530 started, 496520 done, 496520 succeeded, 0 failed, 0 errored, 0 timeout
status codes: 496520 2xx, 0 3xx, 0 4xx, 0 5xx
traffic: 28.89MB (30291303) total, 969.78KB (993052) headers (space savings 94.74%), 14.68MB (15392120) data
                     min         max         mean         sd        +/- sd
time for request:       42us       331us        54us         5us    78.47%
time for connect:       26us        26us        26us         0us   100.00%
time to 1st byte:      142us       142us       142us         0us   100.00%
req/s           :  165502.97   165502.97   165502.97        0.00   100.00%

2a

❯ h2load --clients=10 --duration=3 \
  --header='Content-Type: application/grpc' \
  --data=examples/helloworld/testdata/hello.grpc-lpm.data \
  http://localhost:7000/helloworld.v1.Greeter/Hello
starting benchmark...
spawning thread #0: 10 total client(s). Timing-based test with 0s of warm-up time and 3s of main duration for measurements.
Warm-up started for thread #0.
progress: 10% of clients started
progress: 20% of clients started
progress: 30% of clients started
progress: 40% of clients started
progress: 50% of clients started
progress: 60% of clients started
progress: 70% of clients started
progress: 80% of clients started
progress: 90% of clients started
progress: 100% of clients started
Warm-up phase is over for thread #0.
Main benchmark duration is started for thread #0.
Application protocol: h2c
Main benchmark duration is over for thread #0. Stopping all clients.
Stopped all clients for thread #0

finished in 3.00s, 120382.33 req/s, 7.00MB/s
requests: 361147 total, 361157 started, 361147 done, 361147 succeeded, 0 failed, 0 errored, 0 timeout
status codes: 361147 2xx, 0 3xx, 0 4xx, 0 5xx
traffic: 21.01MB (22032907) total, 705.48KB (722414) headers (space savings 94.74%), 10.68MB (11195557) data
                     min         max         mean         sd        +/- sd
time for request:       15us       311us        66us        17us    70.72%
time for connect:       20us       104us        39us        23us    90.00%
time to 1st byte:      145us       206us       187us        19us    90.00%
req/s           :   12030.99    12045.89    12037.77        4.04    70.00%

2b

❯ h2load --clients=10 --max-concurrent-streams=10 --duration=3 \
  --header='Content-Type: application/grpc' \
  --data=examples/helloworld/testdata/hello.grpc-lpm.data \
  http://localhost:7000/helloworld.v1.Greeter/Hello
starting benchmark...
spawning thread #0: 10 total client(s). Timing-based test with 0s of warm-up time and 3s of main duration for measurements.
Warm-up started for thread #0.
progress: 10% of clients started
progress: 20% of clients started
progress: 30% of clients started
progress: 40% of clients started
progress: 50% of clients started
progress: 60% of clients started
progress: 70% of clients started
progress: 80% of clients started
progress: 90% of clients started
progress: 100% of clients started
Warm-up phase is over for thread #0.
Main benchmark duration is started for thread #0.
Application protocol: h2c
Main benchmark duration is over for thread #0. Stopping all clients.
Stopped all clients for thread #0

finished in 3.00s, 430950.00 req/s, 25.07MB/s
requests: 1292850 total, 1292950 started, 1292850 done, 1292850 succeeded, 0 failed, 0 errored, 0 timeout
status codes: 1292850 2xx, 0 3xx, 0 4xx, 0 5xx
traffic: 75.22MB (78873459) total, 2.47MB (2585820) headers (space savings 94.74%), 38.22MB (40078350) data
                     min         max         mean         sd        +/- sd
time for request:       42us       850us       180us        46us    66.08%
time for connect:       26us       241us        68us        62us    90.00%
time to 1st byte:      287us       463us       380us        70us    50.00%
req/s           :   43075.75    43108.97    43093.66        9.05    80.00%

3a

❯ h2load --clients=100 --duration=3 \
  --header='Content-Type: application/grpc' \
  --data=examples/helloworld/testdata/hello.grpc-lpm.data \
  http://localhost:7000/helloworld.v1.Greeter/Hello
starting benchmark...
spawning thread #0: 100 total client(s). Timing-based test with 0s of warm-up time and 3s of main duration for measurements.
Warm-up started for thread #0.
progress: 10% of clients started
progress: 20% of clients started
progress: 30% of clients started
progress: 40% of clients started
progress: 50% of clients started
progress: 60% of clients started
progress: 70% of clients started
progress: 80% of clients started
progress: 90% of clients started
progress: 100% of clients started
Warm-up phase is over for thread #0.
Main benchmark duration is started for thread #0.
Application protocol: h2c
Main benchmark duration is over for thread #0. Stopping all clients.
Stopped all clients for thread #0

finished in 3.00s, 90955.00 req/s, 5.29MB/s
requests: 272865 total, 272965 started, 272865 done, 272865 succeeded, 0 failed, 0 errored, 0 timeout
status codes: 272865 2xx, 0 3xx, 0 4xx, 0 5xx
traffic: 15.88MB (16650765) total, 534.11KB (546930) headers (space savings 94.73%), 8.07MB (8458815) data
                     min         max         mean         sd        +/- sd
time for request:       52us      1.67ms       734us       254us    63.79%
time for connect:       22us       332us       177us        91us    57.00%
time to 1st byte:     1.17ms      2.01ms      1.50ms       319us    72.00%
req/s           :     908.95      909.61      909.29        0.24    46.00%

3b

❯ h2load --clients=100 --max-concurrent-streams=10 --duration=3 \
  --header='Content-Type: application/grpc' \
  --data=examples/helloworld/testdata/hello.grpc-lpm.data \
  http://localhost:7000/helloworld.v1.Greeter/Hello
starting benchmark...
spawning thread #0: 100 total client(s). Timing-based test with 0s of warm-up time and 3s of main duration for measurements.
Warm-up started for thread #0.
progress: 10% of clients started
progress: 20% of clients started
progress: 30% of clients started
progress: 40% of clients started
progress: 50% of clients started
progress: 60% of clients started
progress: 70% of clients started
progress: 80% of clients started
progress: 90% of clients started
progress: 100% of clients started
Warm-up phase is over for thread #0.
Main benchmark duration is started for thread #0.
Application protocol: h2c
Main benchmark duration is over for thread #0. Stopping all clients.
Stopped all clients for thread #0

finished in 3.00s, 424666.67 req/s, 24.71MB/s
requests: 1274000 total, 1275000 started, 1274000 done, 1274000 succeeded, 0 failed, 0 errored, 0 timeout
status codes: 1274000 2xx, 0 3xx, 0 4xx, 0 5xx
traffic: 74.13MB (77726500) total, 2.43MB (2549200) headers (space savings 94.73%), 37.66MB (39494000) data
                     min         max         mean         sd        +/- sd
time for request:       60us      3.16ms      1.46ms       580us    56.40%
time for connect:       27us       563us       296us       160us    58.00%
time to 1st byte:     2.40ms      3.06ms      2.73ms       192us    57.00%
req/s           :    4243.78     4245.82     4244.65        0.60    63.00%

[uatuko]

Multi-threaded, hardware concurrency

hardware concurrency

1a

❯ h2load --clients=1 --duration=3 \
  --header='Content-Type: application/grpc' \
  --data=examples/helloworld/testdata/hello.grpc-lpm.data \
  http://localhost:7000/helloworld.v1.Greeter/Hello
starting benchmark...
spawning thread #0: 1 total client(s). Timing-based test with 0s of warm-up time and 3s of main duration for measurements.
Warm-up started for thread #0.
progress: 100% of clients started
Warm-up phase is over for thread #0.
Main benchmark duration is started for thread #0.
Application protocol: h2c
Main benchmark duration is over for thread #0. Stopping all clients.
Stopped all clients for thread #0

finished in 3.00s, 34289.67 req/s, 2.00MB/s
requests: 102869 total, 102870 started, 102869 done, 102869 succeeded, 0 failed, 0 errored, 0 timeout
status codes: 102869 2xx, 0 3xx, 0 4xx, 0 5xx
traffic: 5.99MB (6275784) total, 200.93KB (205750) headers (space savings 94.74%), 3.04MB (3188939) data
                     min         max         mean         sd        +/- sd
time for request:       20us       494us        27us         4us    91.58%
time for connect:      208us       208us       208us         0us   100.00%
time to 1st byte:      409us       409us       409us         0us   100.00%
req/s           :   34288.67    34288.67    34288.67        0.00   100.00%

1b

❯ h2load --clients=1 --max-concurrent-streams=10 --duration=3 \
  --header='Content-Type: application/grpc' \
  --data=examples/helloworld/testdata/hello.grpc-lpm.data \
  http://localhost:7000/helloworld.v1.Greeter/Hello
starting benchmark...
spawning thread #0: 1 total client(s). Timing-based test with 0s of warm-up time and 3s of main duration for measurements.
Warm-up started for thread #0.
progress: 100% of clients started
Warm-up phase is over for thread #0.
Main benchmark duration is started for thread #0.
Application protocol: h2c
Main benchmark duration is over for thread #0. Stopping all clients.
Stopped all clients for thread #0

finished in 3.00s, 164196.67 req/s, 9.55MB/s
requests: 492590 total, 492600 started, 492590 done, 492590 succeeded, 0 failed, 0 errored, 0 timeout
status codes: 492590 2xx, 0 3xx, 0 4xx, 0 5xx
traffic: 28.66MB (30051547) total, 962.10KB (985192) headers (space savings 94.74%), 14.56MB (15270290) data
                     min         max         mean         sd        +/- sd
time for request:       42us       545us        54us         5us    81.22%
time for connect:       45us        45us        45us         0us   100.00%
time to 1st byte:      153us       153us       153us         0us   100.00%
req/s           :  164193.76   164193.76   164193.76        0.00   100.00%

2a

❯ h2load --clients=10 --duration=3 \
  --header='Content-Type: application/grpc' \
  --data=examples/helloworld/testdata/hello.grpc-lpm.data \
  http://localhost:7000/helloworld.v1.Greeter/Hello
starting benchmark...
spawning thread #0: 10 total client(s). Timing-based test with 0s of warm-up time and 3s of main duration for measurements.
Warm-up started for thread #0.
progress: 10% of clients started
progress: 20% of clients started
progress: 30% of clients started
progress: 40% of clients started
progress: 50% of clients started
progress: 60% of clients started
progress: 70% of clients started
progress: 80% of clients started
progress: 90% of clients started
progress: 100% of clients started
Warm-up phase is over for thread #0.
Main benchmark duration is started for thread #0.
Application protocol: h2c
Main benchmark duration is over for thread #0. Stopping all clients.
Stopped all clients for thread #0

finished in 3.00s, 118214.67 req/s, 6.88MB/s
requests: 354644 total, 354654 started, 354644 done, 354644 succeeded, 0 failed, 0 errored, 0 timeout
status codes: 354644 2xx, 0 3xx, 0 4xx, 0 5xx
traffic: 20.63MB (21636224) total, 692.78KB (709408) headers (space savings 94.74%), 10.48MB (10993964) data
                     min         max         mean         sd        +/- sd
time for request:       18us      1.07ms        68us        20us    73.25%
time for connect:       22us       107us        42us        23us    90.00%
time to 1st byte:      175us       287us       229us        41us    50.00%
req/s           :   11815.94    11825.05    11821.09        3.27    50.00%

2b

❯ h2load --clients=10 --max-concurrent-streams=10 --duration=3 \
  --header='Content-Type: application/grpc' \
  --data=examples/helloworld/testdata/hello.grpc-lpm.data \
  http://localhost:7000/helloworld.v1.Greeter/Hello
starting benchmark...
spawning thread #0: 10 total client(s). Timing-based test with 0s of warm-up time and 3s of main duration for measurements.
Warm-up started for thread #0.
progress: 10% of clients started
progress: 20% of clients started
progress: 30% of clients started
progress: 40% of clients started
progress: 50% of clients started
progress: 60% of clients started
progress: 70% of clients started
progress: 80% of clients started
progress: 90% of clients started
progress: 100% of clients started
Warm-up phase is over for thread #0.
Main benchmark duration is started for thread #0.
Application protocol: h2c
Main benchmark duration is over for thread #0. Stopping all clients.
Stopped all clients for thread #0

finished in 3.00s, 450420.00 req/s, 26.21MB/s
requests: 1351260 total, 1351360 started, 1351260 done, 1351260 succeeded, 0 failed, 0 errored, 0 timeout
status codes: 1351260 2xx, 0 3xx, 0 4xx, 0 5xx
traffic: 78.62MB (82436950) total, 2.58MB (2702640) headers (space savings 94.74%), 39.95MB (41889060) data
                     min         max         mean         sd        +/- sd
time for request:       41us      1.09ms       169us        45us    71.95%
time for connect:       23us       238us        62us        63us    90.00%
time to 1st byte:      279us       551us       368us        87us    80.00%
req/s           :   45024.20    45053.48    45039.23        8.26    60.00%

3a

❯ h2load --clients=100 --duration=3 \
  --header='Content-Type: application/grpc' \
  --data=examples/helloworld/testdata/hello.grpc-lpm.data \
  http://localhost:7000/helloworld.v1.Greeter/Hello
starting benchmark...
spawning thread #0: 100 total client(s). Timing-based test with 0s of warm-up time and 3s of main duration for measurements.
Warm-up started for thread #0.
progress: 10% of clients started
progress: 20% of clients started
progress: 30% of clients started
progress: 40% of clients started
progress: 50% of clients started
progress: 60% of clients started
progress: 70% of clients started
progress: 80% of clients started
progress: 90% of clients started
progress: 100% of clients started
Warm-up phase is over for thread #0.
Main benchmark duration is started for thread #0.
Application protocol: h2c
Main benchmark duration is over for thread #0. Stopping all clients.
Stopped all clients for thread #0

finished in 3.00s, 101368.00 req/s, 5.90MB/s
requests: 304104 total, 304204 started, 304104 done, 304104 succeeded, 0 failed, 0 errored, 0 timeout
status codes: 304104 2xx, 0 3xx, 0 4xx, 0 5xx
traffic: 17.70MB (18556344) total, 595.13KB (609408) headers (space savings 94.73%), 8.99MB (9427224) data
                     min         max         mean         sd        +/- sd
time for request:       51us      2.47ms       760us       338us    63.68%
time for connect:       25us       355us       191us        94us    60.00%
time to 1st byte:     1.11ms      1.91ms      1.45ms       286us    57.00%
req/s           :    1013.00     1014.44     1013.44        0.26    74.00%

3b

❯ h2load --clients=100 --max-concurrent-streams=10 --duration=3 \
  --header='Content-Type: application/grpc' \
  --data=examples/helloworld/testdata/hello.grpc-lpm.data \
  http://localhost:7000/helloworld.v1.Greeter/Hello
starting benchmark...
spawning thread #0: 100 total client(s). Timing-based test with 0s of warm-up time and 3s of main duration for measurements.
Warm-up started for thread #0.
progress: 10% of clients started
progress: 20% of clients started
progress: 30% of clients started
progress: 40% of clients started
progress: 50% of clients started
progress: 60% of clients started
progress: 70% of clients started
progress: 80% of clients started
progress: 90% of clients started
progress: 100% of clients started
Warm-up phase is over for thread #0.
Main benchmark duration is started for thread #0.
Application protocol: h2c
Main benchmark duration is over for thread #0. Stopping all clients.
Stopped all clients for thread #0

finished in 3.00s, 408493.33 req/s, 23.77MB/s
requests: 1225480 total, 1226480 started, 1225480 done, 1225480 succeeded, 0 failed, 0 errored, 0 timeout
status codes: 1225480 2xx, 0 3xx, 0 4xx, 0 5xx
traffic: 71.30MB (74766780) total, 2.34MB (2452160) headers (space savings 94.73%), 36.23MB (37989880) data
                     min         max         mean         sd        +/- sd
time for request:       70us      3.28ms      1.72ms       470us    65.91%
time for connect:       24us      2.27ms       292us       248us    95.00%
time to 1st byte:     2.38ms      4.61ms      2.73ms       268us    82.00%
req/s           :    4079.49     4084.78     4082.05        2.27    52.00%