Held memory is memory.current less the page cache, which the kernel hands to another tenant when the host needs it. The score charges memory.current, as spec section 8.5 says.
When
Total
Anonymous
Shared
Page cache
Kernel
idle
1.69
1.40
0.00
0.28
0.01
peak
1.69
1.40
0.00
0.28
0.01
ready
1.48
1.05
0.00
0.41
0.01
ready_repetition
0.50
0.30
0.00
0.19
0.01
Latency, warm single client
Class
Operation
p50 ms
p99 ms
Target ms
R1
point
0.87
2.04
10
R2
filter
2.38
5.17
200
R3
neighbourhood
1.18
7.74
50
R4
deep
0.99
2.42
500
R5
path
1.48
2.97
500
R6
pattern
1.15
7.36
2,000
R7
analytic
222.34
256.25
5,000
W1
insert
4.33
13.41
20
W2
insert edge
2.15
2.88
10
W3
batch
74.50
119.36
500
U1
modify
2.60
5.39
20
U2
rewire
3.50
7.48
30
D1
delete edge
2.91
8.70
20
D2
delete cascade
4.33
14.60
100
O1
person by email
65.34
70.33
20
O2
accounts by opening date
102.72
115.14
200
O3
people by surname and city
0.83
3.07
100
S1
amount band
593.92
613.38
1,000
S3
text search
116.48
120.58
2,000
J4
neighbours alike
1.59
3.19
500
R1w
point, worst case
0.70
1.12
10
R2w
filter, worst case
36.86
155.52
200
R3w
neighbourhood, worst case
9.63
24.43
50
R4w
deep, worst case
5.52
10.88
500
R5w
path, worst case
1.47
3.35
500
R6w
pattern, worst case
1.80
9.21
2,000
R7w
analytic, worst case
220.03
232.45
5,000
Operational surface: 2 points
Item
Points
Source
runtime dependency: java
1
audited
index index_f8efaa80: RANGE index on Account(country, status, balance)
# Neo4j 5.26 Community Edition, the LTS line (spec §11.3).
#
# The image is pinned by digest (spec §5.3): the multi-architecture index of
# docker.io/library/neo4j:5.26-community as of 20 October 2026, so each host
# pulls the same release built for its own architecture.
name: neo4j
system: Neo4j Community
version: "5.26"
image: docker.io/library/neo4j@sha256:0a6e7aedf54a531f76286032d4365ead65b11cca05e4a15cdf23b8e53d6db02b
model: property-graph
protocols: [bolt]
query_language: cypher
# Neo4j flushes its transaction log on every commit by default, so commits
# are durable without settings (spec §9.2). Its isolation is read
# committed, weaker than the snapshot isolation the spec asks for; the run
# records that as a deviation.
durability:
settings: {}
isolation: read-committed
# The graph mapping of spec §7.2 brings a uniqueness constraint, with its
# index, on every node and relationship id. Beyond those, R2's filter needs
# one composite index. Neo4j sizes its heap and page cache from the
# container's memory by default, so nothing is sized by hand.
declared_surface:
services: 1
runtime_dependencies: [java]
secondary_indexes:
- "Account(country, status, balance)"
hand_sized_caches: []
non_default_settings: {}
maintenance_tasks: []
storage_areas: [data]
manual_steps: []
envelope:
size: S
memory_limit_gib: 8
plugin: neo4j
container:
port: 7687
data_path: /data
# The image reads an /import mount without changing its ownership, and
# points LOAD CSV at it.
dataset_path: /import
log_paths: [transactions]
env:
NEO4J_AUTH: neo4j/dpbench-neo4j
client:
user: neo4j
password: dpbench-neo4j
database: neo4j
Executed DDL
CREATE CONSTRAINT place_id FOR (n:Place) REQUIRE n.id IS UNIQUE;
CREATE CONSTRAINT person_id FOR (n:Person) REQUIRE n.id IS UNIQUE;
CREATE CONSTRAINT organization_id FOR (n:Organization) REQUIRE n.id IS UNIQUE;
CREATE CONSTRAINT account_id FOR (n:Account) REQUIRE n.id IS UNIQUE;
CREATE CONSTRAINT device_id FOR (n:Device) REQUIRE n.id IS UNIQUE;
CREATE CONSTRAINT tag_id FOR (n:Tag) REQUIRE n.id IS UNIQUE;
CREATE CONSTRAINT post_id FOR (n:Post) REQUIRE n.id IS UNIQUE;
CREATE CONSTRAINT knows_id FOR ()-[r:KNOWS]-() REQUIRE r.id IS UNIQUE;
CREATE CONSTRAINT works_at_id FOR ()-[r:WORKS_AT]-() REQUIRE r.id IS UNIQUE;
CREATE CONSTRAINT uses_id FOR ()-[r:USES]-() REQUIRE r.id IS UNIQUE;
CREATE CONSTRAINT transfer_id FOR ()-[r:TRANSFER]-() REQUIRE r.id IS UNIQUE;
CREATE CONSTRAINT interested_in_id FOR ()-[r:INTERESTED_IN]-() REQUIRE r.id IS UNIQUE;
CREATE CONSTRAINT likes_id FOR ()-[r:LIKES]-() REQUIRE r.id IS UNIQUE;
LOAD CSV WITH HEADERS FROM 'file:///data/place.csv' AS r
CALL { WITH r CREATE (:Place {id: toInteger(r.id), name: r.name, type: r.type}) } IN TRANSACTIONS OF 20000 ROWS;
LOAD CSV WITH HEADERS FROM 'file:///data/person.csv' AS r
CALL { WITH r CREATE (:Person {id: toInteger(r.id), first_name: r.first_name, last_name: r.last_name, birth_date: date(r.birth_date), created_at: datetime(r.created_at), risk_score: toFloat(r.risk_score), email: r.email}) } IN TRANSACTIONS OF 20000 ROWS;
LOAD CSV WITH HEADERS FROM 'file:///data/organization.csv' AS r
CALL { WITH r CREATE (:Organization {id: toInteger(r.id), name: r.name, type: r.type, founded: date(r.founded)}) } IN TRANSACTIONS OF 20000 ROWS;
LOAD CSV WITH HEADERS FROM 'file:///data/account.csv' AS r
CALL { WITH r CREATE (:Account {id: toInteger(r.id), opened_at: datetime(r.opened_at), balance: CASE WHEN r.balance IS NULL THEN null ELSE (CASE WHEN r.balance STARTS WITH '-' THEN -1 ELSE 1 END) * (toInteger(replace(split(r.balance, '.')[0], '-', '')) * 100 + toInteger(split(r.balance, '.')[1])) END, status: r.status, country: r.country}) } IN TRANSACTIONS OF 20000 ROWS;
LOAD CSV WITH HEADERS FROM 'file:///data/device.csv' AS r
CALL { WITH r CREATE (:Device {id: toInteger(r.id), kind: r.kind, os: r.os, first_seen: datetime(r.first_seen), fingerprint: r.fingerprint}) } IN TRANSACTIONS OF 20000 ROWS;
LOAD CSV WITH HEADERS FROM 'file:///data/tag.csv' AS r
CALL { WITH r CREATE (:Tag {id: toInteger(r.id), name: r.name, class: r.class}) } IN TRANSACTIONS OF 20000 ROWS;
LOAD CSV WITH HEADERS FROM 'file:///data/post.csv' AS r
CALL { WITH r CREATE (:Post {id: toInteger(r.id), created_at: datetime(r.created_at), length: toInteger(r.length), language: r.language, content: r.content}) } IN TRANSACTIONS OF 20000 ROWS;
LOAD CSV WITH HEADERS FROM 'file:///data/place.csv' AS r
CALL { WITH r WITH r WHERE r.part_of IS NOT NULL MATCH (a:Place {id: toInteger(r.id)}) MATCH (b:Place {id: toInteger(r.part_of)}) CREATE (a)-[:PART_OF]->(b) } IN TRANSACTIONS OF 20000 ROWS;
LOAD CSV WITH HEADERS FROM 'file:///data/person.csv' AS r
CALL { WITH r WITH r WHERE r.lives_in IS NOT NULL MATCH (a:Person {id: toInteger(r.id)}) MATCH (b:Place {id: toInteger(r.lives_in)}) CREATE (a)-[:LIVES_IN]->(b) } IN TRANSACTIONS OF 20000 ROWS;
LOAD CSV WITH HEADERS FROM 'file:///data/organization.csv' AS r
CALL { WITH r WITH r WHERE r.located_in IS NOT NULL MATCH (a:Organization {id: toInteger(r.id)}) MATCH (b:Place {id: toInteger(r.located_in)}) CREATE (a)-[:LOCATED_IN]->(b) } IN TRANSACTIONS OF 20000 ROWS;
LOAD CSV WITH HEADERS FROM 'file:///data/account.csv' AS r
CALL { WITH r WITH r WHERE r.owner_id IS NOT NULL MATCH (a:Account {id: toInteger(r.id)}) MATCH (b:Person {id: toInteger(r.owner_id)}) CREATE (a)-[:OWNED_BY]->(b) } IN TRANSACTIONS OF 20000 ROWS;
LOAD CSV WITH HEADERS FROM 'file:///data/account.csv' AS r
CALL { WITH r WITH r WHERE r.org_id IS NOT NULL MATCH (a:Account {id: toInteger(r.id)}) MATCH (b:Organization {id: toInteger(r.org_id)}) CREATE (a)-[:HELD_AT]->(b) } IN TRANSACTIONS OF 20000 ROWS;
LOAD CSV WITH HEADERS FROM 'file:///data/post.csv' AS r
CALL { WITH r WITH r WHERE r.author_id IS NOT NULL MATCH (a:Post {id: toInteger(r.id)}) MATCH (b:Person {id: toInteger(r.author_id)}) CREATE (a)-[:HAS_AUTHOR]->(b) } IN TRANSACTIONS OF 20000 ROWS;
LOAD CSV WITH HEADERS FROM 'file:///data/post.csv' AS r
CALL { WITH r WITH r WHERE r.tag_id IS NOT NULL MATCH (a:Post {id: toInteger(r.id)}) MATCH (b:Tag {id: toInteger(r.tag_id)}) CREATE (a)-[:ABOUT]->(b) } IN TRANSACTIONS OF 20000 ROWS;
LOAD CSV WITH HEADERS FROM 'file:///data/knows.csv' AS r
CALL { WITH r MATCH (a:Person {id: toInteger(r.src_id)}) MATCH (b:Person {id: toInteger(r.dst_id)}) CREATE (a)-[:KNOWS {id: toInteger(r.id), since: date(r.since), strength: toFloat(r.strength)}]->(b) } IN TRANSACTIONS OF 20000 ROWS;
LOAD CSV WITH HEADERS FROM 'file:///data/works_at.csv' AS r
CALL { WITH r MATCH (a:Person {id: toInteger(r.src_id)}) MATCH (b:Organization {id: toInteger(r.dst_id)}) CREATE (a)-[:WORKS_AT {id: toInteger(r.id), from_date: date(r.from_date), role: r.role}]->(b) } IN TRANSACTIONS OF 20000 ROWS;
LOAD CSV WITH HEADERS FROM 'file:///data/uses.csv' AS r
CALL { WITH r MATCH (a:Person {id: toInteger(r.src_id)}) MATCH (b:Device {id: toInteger(r.dst_id)}) CREATE (a)-[:USES {id: toInteger(r.id), first_seen: datetime(r.first_seen), last_seen: datetime(r.last_seen)}]->(b) } IN TRANSACTIONS OF 20000 ROWS;
LOAD CSV WITH HEADERS FROM 'file:///data/transfer.csv' AS r
CALL { WITH r MATCH (a:Account {id: toInteger(r.src_id)}) MATCH (b:Account {id: toInteger(r.dst_id)}) CREATE (a)-[:TRANSFER {id: toInteger(r.id), ts: datetime(r.ts), amount: CASE WHEN r.amount IS NULL THEN null ELSE (CASE WHEN r.amount STARTS WITH '-' THEN -1 ELSE 1 END) * (toInteger(replace(split(r.amount, '.')[0], '-', '')) * 100 + toInteger(split(r.amount, '.')[1])) END, currency: r.currency, channel: r.channel}]->(b) } IN TRANSACTIONS OF 20000 ROWS;
LOAD CSV WITH HEADERS FROM 'file:///data/interested_in.csv' AS r
CALL { WITH r MATCH (a:Person {id: toInteger(r.src_id)}) MATCH (b:Tag {id: toInteger(r.dst_id)}) CREATE (a)-[:INTERESTED_IN {id: toInteger(r.id), weight: toFloat(r.weight)}]->(b) } IN TRANSACTIONS OF 20000 ROWS;
LOAD CSV WITH HEADERS FROM 'file:///data/likes.csv' AS r
CALL { WITH r MATCH (a:Person {id: toInteger(r.src_id)}) MATCH (b:Post {id: toInteger(r.dst_id)}) CREATE (a)-[:LIKES {id: toInteger(r.id), ts: datetime(r.ts)}]->(b) } IN TRANSACTIONS OF 20000 ROWS;
CREATE INDEX FOR (x:Account) ON (x.country, x.status, x.balance);
CALL db.awaitIndexes(3600);
Settings that differ from a fresh container of the same image
None.
5. Queries
The adapter runs these statements, with parameters bound as $n.
O3
MATCH (p:Person {last_name: $last})-[:LIVES_IN]->(:Place {id: $city})
OPTIONAL MATCH (a:Account)-[:OWNED_BY]->(p)
RETURN count(DISTINCT p), count(a), coalesce(sum(a.balance), 0)
S3
MATCH (p:Post)
WHERE p.language = $language AND p.created_at >= $from AND p.created_at < $to AND p.content CONTAINS $word
RETURN count(p)
J5
MATCH (:Account {id: $id})-[t1:TRANSFER]->()-[t2:TRANSFER]->()-[t3:TRANSFER]->()-[t4:TRANSFER]->()
WHERE t1.ts >= $from AND t1.ts < $to AND t2.ts > t1.ts AND t3.ts > t2.ts AND t4.ts > t3.ts
AND t4.ts < t1.ts + duration({days: 30})
RETURN count(*)
J3
MATCH (p:Person {id: $id})-[:KNOWS]->(a:Person)-[:KNOWS]->(b:Person)-[:KNOWS]->(p)
WHERE a <> p AND b <> p AND a <> b
RETURN count(*)
J1
MATCH (p:Person {id: $id})-[:KNOWS*1..4]->(x:Person)
WHERE x <> p
RETURN count(DISTINCT x)
J2
MATCH (:Person {id: $id})-[:KNOWS]->(:Person)<-[:OWNED_BY]-(:Account)-[t:TRANSFER]->(b:Account)
WHERE t.ts >= $from AND t.ts < $to
MATCH (b)-[:OWNED_BY]->(:Person)-[:LIVES_IN]->(:Place)-[:PART_OF]->(region:Place)
WITH region, count(t) AS n, sum(t.amount) AS total
RETURN region.name, n, total
J4
MATCH (p:Person {id: $id})-[:LIVES_IN]->(:Place)<-[:LIVES_IN]-(o:Person)
WHERE o <> p AND EXISTS { (o)-[:WORKS_AT]->() } AND EXISTS { (p)-[:INTERESTED_IN]->(:Tag)<-[:INTERESTED_IN]-(o) }
RETURN count(DISTINCT o)
S2
MATCH (p:Person)
WHERE p.risk_score >= $risk AND p.birth_date >= date($from) AND p.birth_date < date($to)
RETURN p.id
S1
MATCH ()-[t:TRANSFER]->()
WHERE t.ts >= $from AND t.ts < $to AND t.channel = $channel AND t.amount >= $lo AND t.amount <= $hi
RETURN count(t), coalesce(sum(t.amount), 0)
O1
MATCH (p:Person {email: $email})
OPTIONAL MATCH (a:Account)-[:OWNED_BY]->(p)
RETURN p.id, p.first_name, p.last_name, a.id, a.balance, a.status
O2
MATCH (a:Account)
WHERE a.opened_at >= $from AND a.opened_at < $to AND a.balance >= $lo
RETURN a.id
S4
MATCH ()-[t:TRANSFER]->()
WHERE t.ts >= $from AND t.ts < $to
RETURN t.currency, t.channel, count(t), sum(t.amount)
R1
MATCH (p:Person {id: $id})
OPTIONAL MATCH (a:Account)-[:OWNED_BY]->(p)
RETURN p.id, p.first_name, p.last_name, a.id, a.balance, a.status
R2
MATCH (a:Account)
WHERE a.country = $country AND a.status = 'open' AND a.balance >= $lo AND a.balance <= $hi
RETURN a.id
R3
MATCH (:Person {id: $id})-[:KNOWS]->(f:Person)
OPTIONAL MATCH (a:Account)-[:OWNED_BY]->(f)
OPTIONAL MATCH (a)-[t:TRANSFER]->()
WHERE t.ts >= $from AND t.ts < $to
RETURN f.id, sum(t.amount)
R4
MATCH (s:Account {id: $id})-[ts:TRANSFER*1..3]->(x:Account)
WHERE all(t IN ts WHERE t.ts >= $from AND t.ts < $to) AND x <> s
RETURN DISTINCT x.id
R5
MATCH (a:Person {id: $src}), (b:Person {id: $dst})
MATCH p = shortestPath((a)-[:KNOWS*..6]->(b))
RETURN length(p)
R6
MATCH p = (a:Account {id: $id})-[ts:TRANSFER*3..5]->(a)
WHERE all(t IN ts WHERE t.ts >= $from AND t.ts < $to)
WITH a, nodes(p)[1..-1] AS mid
WHERE NOT a IN mid AND all(i IN range(0, size(mid) - 2) WHERE NOT mid[i] IN mid[i + 1..])
RETURN count(*)
R7
MATCH (o:Organization)<-[:HELD_AT]-(:Account)
WITH o, count(*) AS held ORDER BY held DESC, o.id ASC LIMIT 20
MATCH (o)<-[:HELD_AT]-(sa:Account)-[t:TRANSFER]->(da:Account)
WHERE t.ts >= $from AND t.ts < $to
RETURN substring(toString(t.ts), 0, 7), sa.country, da.country, sum(t.amount)
W1
CREATE (p:Person {id: $pid, first_name: $fn, last_name: $ln, birth_date: date('1990-01-01'), created_at: $ts, risk_score: 0.5})
WITH p MATCH (home:Place {id: 1}) CREATE (p)-[:LIVES_IN]->(home)
CREATE (:Account {id: $aid, opened_at: $ts, balance: 0, status: 'open', country: 'ZZ'})-[:OWNED_BY]->(p)
WITH p UNWIND $knows AS k
MATCH (d:Person {id: k.dst})
CREATE (p)-[:KNOWS {id: k.id, since: date($ts), strength: 0.5}]->(d)
W2
MATCH (s:Account {id: $src}) MATCH (d:Account {id: $dst})
CREATE (s)-[:TRANSFER {id: $id, ts: $ts, amount: $amount, currency: $currency, channel: $channel}]->(d)
W3
UNWIND $batch AS b
MATCH (s:Account {id: b.src}) MATCH (d:Account {id: b.dst})
CREATE (s)-[:TRANSFER {id: b.id, ts: b.ts, amount: b.amount, currency: b.currency, channel: b.channel}]->(d)
U1
MATCH (a:Account {id: $id}) SET a.balance = $balance, a.status = $status
U2
MATCH (p:Person {id: $id})
OPTIONAL MATCH (p)-[w:WORKS_AT]->()
DELETE w
WITH DISTINCT p
MATCH (o:Organization {id: $org})
CREATE (p)-[:WORKS_AT {id: $wid, from_date: date($ts), role: 'engineer'}]->(o)
D1
MATCH ()-[t:TRANSFER {id: $id}]->() DELETE t
D2
MATCH (p:Person {id: $id})
OPTIONAL MATCH (a:Account)-[:OWNED_BY]->(p)
OPTIONAL MATCH (post:Post)-[:HAS_AUTHOR]->(p)
WITH p, collect(DISTINCT a) + collect(DISTINCT post) AS owned
FOREACH (n IN owned | DETACH DELETE n)
DETACH DELETE p
6. Concurrency ladder, mixed
k
X ops/s
Q
X × Q
1
35.6
0.713
25.4
4 k*
75.5
0.727
54.9
16
80.4
0.284
22.8
64
85.1
0.054
4.6
7. Memory over the duty cycle
memory.current of the median repetition, sampled every 100 ms. Shaded bands are the busy windows; the rest is idle.
8. Validation
0 invalid operation instances.
Write validation read back 200 account, 286 batch, 200 person alive, 149 person deleted, 200 transfer alive, 200 transfer deleted, 200 works, with 0 kinds failing.
9. Deviations
dataset.scale: 3M is not a DP-Core scale factor (spec §6.6)
tiers: 15% off-index and 5% worst-case operations blended in; spec §9.3's mix is typical operations alone
profile: quick timings: ladders, duty cycle and warm-up are shorter than spec §8.5, §8.9 and §9.5 require
repetitions: 1; spec §9.5 publishes the median of 3
every container runs --privileged (DPBENCH_PRIVILEGED): the host's kernel has no BPF syscall, so the runtime cannot attach a cgroup v2 device filter; cpu, cpuset, memory and pids limits apply as before
the SUT runs at read-committed isolation; spec §9.2 asks for snapshot isolation or stronger
warm-up before P5 runs at 16 clients: spec §9.5 says k*, which only P7 finds
10. Phase detail
Import (P2): 3,000,000 rows in 116.1 s: load 21.0 s, keys 93.4 s, indexes 1.5 s. Peak memory during load 2.19 GiB.
Repetitions (P7, P8)
n
T
k*
Q at k*
M GiB
After work GiB
1
54.9
4
0.727
1.67
2.02
Coefficient of variation of DataPower across repetitions: 0.0%.
Memory floor (P9)
Limit GiB
Q
Outcome
8
0.727
ran
4
0.930
ran
2
0.931
ran
1
0.000
failed
Cold (P3) and warm (P5) single-client latency, ms
Class
Operation
Cold p50
Cold p99
Warm p50
Warm p99
R1
point
12.59
2,136.06
0.87
2.04
R2
filter
38.21
248.06
2.38
5.17
R3
neighbourhood
17.25
356.35
1.18
7.74
R4
deep
16.77
250.50
0.99
2.42
R5
path
10.00
231.04
1.48
2.97
R6
pattern
19.71
181.89
1.15
7.36
R7
analytic
282.11
2,185.22
222.34
256.25
W1
insert
12.71
672.25
4.33
13.41
W2
insert edge
5.90
127.94
2.15
2.88
W3
batch
173.69
752.13
74.50
119.36
U1
modify
6.52
92.22
2.60
5.39
U2
rewire
8.57
192.25
3.50
7.48
D1
delete edge
4.19
60.19
2.91
8.70
D2
delete cascade
10.48
153.09
4.33
14.60
O1
person by email
75.65
318.46
65.34
70.33
O2
accounts by opening date
99.52
187.13
102.72
115.14
O3
people by surname and city
15.26
117.18
0.83
3.07
S1
amount band
569.86
833.53
593.92
613.38
S3
text search
109.76
289.54
116.48
120.58
J4
neighbours alike
19.34
238.59
1.59
3.19
R1w
point, worst case
1.99
6.45
0.70
1.12
R2w
filter, worst case
46.27
51.68
36.86
155.52
R3w
neighbourhood, worst case
25.07
47.10
9.63
24.43
R4w
deep, worst case
13.89
17.18
5.52
10.88
R5w
path, worst case
5.26
6.01
1.47
3.35
R6w
pattern, worst case
8.78
16.57
1.80
9.21
R7w
analytic, worst case
226.18
243.46
220.03
232.45
Ladder (P6): read
k
X ops/s
Errors
Driver CPU
1
31.9
0
1.6%
4
66.4
0
3.6%
16
60.5
0
2.9%
64
47.7
0
2.0%
Ladder (P6): insert
k
X ops/s
Errors
Driver CPU
1
306.0
0
2.8%
4
588.6
0
4.4%
16
670.4
0
4.7%
64
428.2
0
3.2%
Ladder (P6): modify-delete
k
X ops/s
Errors
Driver CPU
1
847.0
0
4.3%
4
2,075.4
0
9.2%
16
4,354.2
0
16.2%
64
5,113.6
0
16.6%
Ladder (P6): write
k
X ops/s
Errors
Driver CPU
1
511.2
0
3.6%
4
1,068.4
0
6.3%
16
1,251.6
0
6.6%
64
517.4
0
3.6%
11. Run timeline
The run from its first phase to its last, every second. The driver counts every operation, ramps and warm-ups included, and the probe samples the SUT's cgroup every 100 ms. Bands mark the phases; P3, the cold run, and P4, the warm-up, stand out in teal. After the duty cycle the run asks the kernel to reclaim everything the SUT holds, and the drop in memory shows what the system gives back.
ThroughputValid operations per second, ramps and warm-ups included.Latencyp99 of every operation in each second, all classes together, on a log scale.MemoryThe SUT's cgroup memory, stacked: what the processes hold, shared memory such as a buffer pool, and the page cache the kernel can reclaim. The envelope allows 8 GiB.
process (anon)shared memorypage cachekernel
CPUCores the SUT kept busy.DiskBytes the SUT read from and wrote to its block devices.
readwritten
Phase
Starts
Seconds
Operations
ops/s
CPU cores
Peak GiB
Page cache at end GiB
P0 environment check
0 s
0.8
0
0.0
0.00
0.00
0.00
P1 schema definition
1 s
11.9
0
0.0
1.75
0.62
0.28
P2 import
13 s
142.8
0
0.0
1.39
2.08
0.00
P3 cold single-client run
155 s
44.1
223
5.1
1.62
1.62
0.70
P4 warm-up
200 s
12.4
682
55.0
1.85
1.82
0.71
P5 warm single-client run
212 s
30.7
560
18.3
1.03
1.85
0.73
P6 read, write and modify ladders
243 s
165.9
126,747
763.9
1.64
2.92
0.97
P7 mixed ladder
409 s
359.0
16,008
44.6
1.23
2.87
0.00
P9 memory floor search
768 s
40.2
1,766
44.0
1.79
1.12
0.09
P10 validation reads
808 s
5.1
0
0.0
1.95
0.92
0.21
P11 final footprint and surface audit
813 s
22.1
0
0.0
0.11
0.82
0.21
P12 teardown
835 s
7.0
0
0.0
0.01
0.82
0.21
12. From the cold restart
P3 restarts the SUT with an empty page cache and runs every class from one client. P4 warms the system up under load, and P5 runs every class again. These charts start at the restart, so they show how long the system takes to reach its warm speed and what it holds while it gets there.
Every single-client operationEach dot is one operation, on a log scale. P3 runs every class in turn on an empty page cache; P5 runs them again after the warm-up.
P3 coldP5 warm
ThroughputValid operations per second, ramps and warm-ups included.Latencyp99 of every operation in each second, all classes together, on a log scale.MemoryThe SUT's cgroup memory, stacked: what the processes hold, shared memory such as a buffer pool, and the page cache the kernel can reclaim. The envelope allows 8 GiB.
process (anon)shared memorypage cachekernel
CPUCores the SUT kept busy.DiskBytes the SUT read from and wrote to its block devices.
readwritten
Cold and warm mediansEach class's p50 on an empty page cache (P3) and after the warm-up (P5), log scale. The longer the line, the more the class depends on a warm cache.