fix(bench): adjust unpaced newPayload latency#23607

Closed

mediocregopher wants to merge 1 commit into

mainfrom

fix/bench-newpayload-latency

Member

mediocregopher commented Apr 17, 2026 •

edited

Loading

When the benchmark runs without wait-time, subtract persistence_wait from the reported newPayload latency so backpressure from prior blocks does not inflate the per-block number. When wait-time is configured, keep the raw server timing returned by reth_newPayload.

Co-Authored-By: Brian Picciano 933154+mediocregopher@users.noreply.github.com

Prompted by: Brian

mediocregopher requested review from Rjected, mattsse and shekhirin as code owners

April 17, 2026 08:45

github-project-automation Bot added this to Reth Tracker

github-project-automation Bot moved this to Backlog in Reth Tracker

mediocregopher added the C-benchmark label

mediocregopher self-assigned this

Member

decofe commented Apr 17, 2026 •

edited

Loading

cc @mediocregopher

✅ Benchmark complete! View job

Benchmark Results

⚠️ Feature is 1 commit behind fix/bench-newpayload-latency. Consider rebasing for accurate results.

Metric	`fix/bench-newpayload-latency`	`fix/bench-newpayload-latency`	Change
Mean	768.53ms	817.17ms	+6.33% ❌ (±4.07%)
StdDev	165.23ms	157.75ms
P50	743.90ms	815.48ms	+9.62% ❌ (±4.56%)
P90	970.08ms	1018.00ms	+4.94% ⚪ (±10.05%)
P99	1370.12ms	1347.56ms	-1.65% ⚪ (±21.68%)
Mgas/s	1377.17	1292.33	-6.16% ❌ (±3.39%)
Wall Clock	156.80s	166.31s	+6.06% ❌ (±4.17%)
Persist Wait	0.88ms	0.21ms	-76.18% ⚪ (±99.06%)

100 big blocks, 10 warmup, BAL: feature

Wait Time Breakdown

Persistence Wait

Metric	`fix/bench-newpayload-latency`	`fix/bench-newpayload-latency`
Mean	0.88ms	0.21ms
P50	0.06ms	0.05ms
P95	0.08ms	0.08ms

Trie Cache Update Wait

Metric	`fix/bench-newpayload-latency`	`fix/bench-newpayload-latency`
Mean	0.00ms	0.00ms
P50	0.00ms	0.00ms
P95	0.00ms	0.00ms

Execution Cache Update Wait

Metric	`fix/bench-newpayload-latency`	`fix/bench-newpayload-latency`
Mean	0.00ms	0.00ms
P50	0.00ms	0.00ms
P95	0.00ms	0.00ms

Grafana Dashboard

Charts

Latency, Throughput & Diff

Wait Time Breakdown

Gas vs Latency

Grafana Dashboard

View real-time metrics

Member

decofe commented Apr 17, 2026 •

edited

Loading

cc @mediocregopher

✅ Benchmark complete! View job

Benchmark Results

⚠️ Feature is 1 commit behind fix/bench-newpayload-latency. Consider rebasing for accurate results.

Metric	`fix/bench-newpayload-latency`	`fix/bench-newpayload-latency`	Change
Mean	768.70ms	827.84ms	+7.69% ❌ (±4.22%)
StdDev	173.79ms	166.60ms
P50	739.73ms	825.04ms	+11.53% ❌ (±5.41%)
P90	975.29ms	1052.42ms	+7.91% ⚪ (±10.91%)
P99	1367.81ms	1270.28ms	-7.13% ⚪ (±16.08%)
Mgas/s	1381.82	1280.57	-7.33% ❌ (±3.54%)
Wall Clock	156.80s	168.40s	+7.40% ❌ (±4.35%)
Persist Wait	0.26ms	0.31ms	+19.66% ⚪ (±189.65%)

100 big blocks, 10 warmup, wait time: 500ms, BAL: feature

Wait Time Breakdown

Persistence Wait

Metric	`fix/bench-newpayload-latency`	`fix/bench-newpayload-latency`
Mean	0.26ms	0.31ms
P50	0.06ms	0.06ms
P95	0.08ms	0.09ms

Trie Cache Update Wait

Metric	`fix/bench-newpayload-latency`	`fix/bench-newpayload-latency`
Mean	0.00ms	0.00ms
P50	0.00ms	0.00ms
P95	0.00ms	0.00ms

Execution Cache Update Wait

Metric	`fix/bench-newpayload-latency`	`fix/bench-newpayload-latency`
Mean	0.00ms	0.00ms
P50	0.00ms	0.00ms
P95	0.00ms	0.00ms

Grafana Dashboard

Charts

Latency, Throughput & Diff

Wait Time Breakdown

Gas vs Latency

Grafana Dashboard

View real-time metrics


          fix(bench): adjust unpaced newPayload latency

b11aba4

Subtract persistence wait from reported newPayload latency when the benchmark runs without wait-time. Paced runs keep the raw server timing returned by reth_newPayload.

Co-authored-by: Brian Picciano <933154+mediocregopher@users.noreply.github.com>
Co-authored-by: Amp <amp@ampcode.com>
Amp-Thread-ID: https://ampcode.com/threads/T-019d9a98-5803-738f-b05a-065e4d7452a8

mediocregopher force-pushed the fix/bench-newpayload-latency branch from dd60baf to b11aba4 Compare

April 17, 2026 09:04

Member

decofe commented Apr 17, 2026 •

edited

Loading

cc @mediocregopher

✅ Benchmark complete! View job

Benchmark Results

Metric	`fix/bench-newpayload-latency`	`fix/bench-newpayload-latency`	Change
Mean	29.02ms	29.13ms	+0.38% ⚪ (±2.15%)
StdDev	13.41ms	14.47ms
P50	26.50ms	26.51ms	+0.03% ⚪ (±2.02%)
P90	45.00ms	45.54ms	+1.21% ⚪ (±3.27%)
P99	82.42ms	85.78ms	+4.09% ⚪ (±20.48%)
Mgas/s	1061.82	1065.83	+0.38% ⚪ (±1.16%)
Wall Clock	29.93s	30.08s	+0.51% ⚪ (±2.12%)
Persist Wait	26.15ms	26.05ms	-0.37% ⚪ (±26.71%)

500 blocks, 200 warmup

Wait Time Breakdown

Persistence Wait

Metric	`fix/bench-newpayload-latency`	`fix/bench-newpayload-latency`
Mean	26.15ms	26.05ms
P50	0.02ms	0.02ms
P95	232.95ms	248.71ms

Trie Cache Update Wait

Metric	`fix/bench-newpayload-latency`	`fix/bench-newpayload-latency`
Mean	0.52ms	0.47ms
P50	0.00ms	0.00ms
P95	2.53ms	2.17ms

Execution Cache Update Wait

Metric	`fix/bench-newpayload-latency`	`fix/bench-newpayload-latency`
Mean	0.00ms	0.00ms
P50	0.00ms	0.00ms
P95	0.00ms	0.00ms

Grafana Dashboard

Charts

Latency, Throughput & Diff

Wait Time Breakdown

Gas vs Latency

Grafana Dashboard

View real-time metrics

Member

decofe commented Apr 17, 2026 •

edited

Loading

cc @mediocregopher

✅ Benchmark complete! View job

Benchmark Results

Metric	`fix/bench-newpayload-latency`	`fix/bench-newpayload-latency`	Change
Mean	24.81ms	24.77ms	-0.15% ⚪ (±0.90%)
StdDev	11.09ms	11.28ms
P50	23.11ms	23.16ms	+0.21% ⚪ (±1.38%)
P90	40.62ms	40.33ms	-0.70% ⚪ (±1.78%)
P99	55.47ms	54.47ms	-1.81% ⚪ (±21.86%)
Mgas/s	1237.71	1239.88	+0.18% ⚪ (±0.43%)
Wall Clock	25.58s	25.54s	-0.15% ⚪ (±0.88%)
Persist Wait	0.03ms	0.03ms	+0.21% ⚪ (±2.77%)

500 blocks, 200 warmup, wait time: 500ms

Wait Time Breakdown

Persistence Wait

Metric	`fix/bench-newpayload-latency`	`fix/bench-newpayload-latency`
Mean	0.03ms	0.03ms
P50	0.03ms	0.03ms
P95	0.04ms	0.04ms

Trie Cache Update Wait

Metric	`fix/bench-newpayload-latency`	`fix/bench-newpayload-latency`
Mean	0.00ms	0.00ms
P50	0.00ms	0.00ms
P95	0.00ms	0.00ms

Execution Cache Update Wait

Metric	`fix/bench-newpayload-latency`	`fix/bench-newpayload-latency`
Mean	0.00ms	0.00ms
P50	0.00ms	0.00ms
P95	0.00ms	0.00ms

Grafana Dashboard

Charts

Latency, Throughput & Diff

Wait Time Breakdown

Gas vs Latency

Grafana Dashboard

View real-time metrics

Member

decofe commented Apr 17, 2026 •

edited

Loading

cc @mediocregopher

❌ Benchmark failed while unknown step. View logs

Member

decofe commented Apr 17, 2026 •

edited

Loading

cc @mediocregopher

✅ Benchmark complete! View job

Benchmark Results

Metric	`fix/bench-newpayload-latency`	`fix/bench-newpayload-latency`	Change
Mean	40.54ms	40.20ms	-0.82% ⚪ (±7.06%)
StdDev	35.48ms	35.93ms
P50	29.38ms	28.71ms	-2.26% ⚪ (±3.46%)
P90	80.43ms	79.91ms	-0.65% ⚪ (±15.96%)
P99	197.89ms	198.43ms	+0.27% ⚪ (±14.03%)
Mgas/s	942.53	948.35	+0.62% ⚪ (±2.72%)
Wall Clock	41.69s	41.27s	-1.01% ⚪ (±6.87%)
Persist Wait	8.42ms	8.56ms	+1.66% ⚪ (±32.15%)

500 blocks, 200 warmup, wait time: 50ms

Wait Time Breakdown

Persistence Wait

Metric	`fix/bench-newpayload-latency`	`fix/bench-newpayload-latency`
Mean	8.42ms	8.56ms
P50	0.03ms	0.03ms
P95	80.20ms	85.71ms

Trie Cache Update Wait

Metric	`fix/bench-newpayload-latency`	`fix/bench-newpayload-latency`
Mean	0.12ms	0.16ms
P50	0.00ms	0.00ms
P95	0.40ms	0.72ms

Execution Cache Update Wait

Metric	`fix/bench-newpayload-latency`	`fix/bench-newpayload-latency`
Mean	0.00ms	0.00ms
P50	0.00ms	0.00ms
P95	0.00ms	0.00ms

Grafana Dashboard

Charts

Latency, Throughput & Diff

Wait Time Breakdown

Gas vs Latency

Grafana Dashboard

View real-time metrics

Member Author

mediocregopher commented Apr 20, 2026

Superseded by #23646, which reverts #23541 and #23578 instead.

mediocregopher closed this

github-project-automation Bot moved this from Backlog to Done in Reth Tracker

emmajam deleted the fix/bench-newpayload-latency branch

May 1, 2026 10:08

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels