ReperioNet benchmark results
Generated by benchmarks/run-matrix.sh 100000 10 on 2026-06-07 12:54 UTC.
CPU: Intel(R) Core(TM) i7-6850K CPU @ 3.60GHz (8 cores, 3597.912 MHz)
Memory: 31Gi total
OS: Linux 6.8.0-101-generic
.NET: 8.0.127
Documents per run: 100000 email-like texts (en/de/fr 5:3:2, ~860 bytes average)
Device classes are simulated via CPU affinity (taskset); per-core clock speed and storage
speed cannot be throttled from user space, so phone-class numbers are optimistic — they
model reduced parallelism, not slower silicon.
Profiles: full = trigram + stemming + phonetic + stored content; no-trigram drops
substring recall; compact additionally stores no content copy (no snippets);
smallest additionally drops phonetic codes and removes stop words from the stem stream.
Multi-token queries default to all-terms (AND) semantics with an automatic any-term
fallback when fewer than Limit candidates match (SearchQueryOptions.TermMatch); the
latency battery shows both modes side by side. The gain grows with term selectivity —
this corpus's tiny synthetic vocabulary makes every term pathologically common, so the
all-terms numbers here are a lower bound.
run
profile
cores
index docs/s
optimize
content
db size
db/content
needle p50
common-term p50
concurrent qps
add
peak mem
smoke
desktop / full
full
8
2,956
4.3 s
81.8 MiB
357.4 MiB
4.37x
5.5 ms
180.4 ms
21.0
1.1 ms
285 MiB
pass
desktop / no-trigram
no-trigram
8
6,410
1.1 s
81.8 MiB
166.9 MiB
2.04x
0.6 ms
76.0 ms
45.1
0.7 ms
282 MiB
pass
desktop / compact
compact
8
6,443
1.0 s
81.8 MiB
166.9 MiB
2.04x
0.4 ms
79.8 ms
44.5
0.5 ms
310 MiB
pass
desktop / smallest
smallest
8
7,655
0.8 s
81.8 MiB
150.6 MiB
1.84x
0.4 ms
77.6 ms
55.1
0.5 ms
274 MiB
pass
fast-phone (4 cores) / full
full
4
3,158
4.3 s
81.8 MiB
357.4 MiB
4.37x
5.2 ms
173.0 ms
16.8
0.7 ms
311 MiB
pass
fast-phone (4 cores) / no-trigram
no-trigram
4
6,677
1.0 s
81.8 MiB
166.9 MiB
2.04x
0.3 ms
76.6 ms
31.7
0.5 ms
289 MiB
pass
fast-phone (4 cores) / compact
compact
4
6,674
1.0 s
81.8 MiB
166.9 MiB
2.04x
0.4 ms
71.8 ms
31.4
0.6 ms
278 MiB
pass
fast-phone (4 cores) / smallest
smallest
4
8,441
0.6 s
81.8 MiB
150.6 MiB
1.84x
0.4 ms
78.0 ms
38.1
0.5 ms
216 MiB
pass
slow-phone (2 cores) / full
full
2
3,320
4.4 s
81.8 MiB
357.4 MiB
4.37x
5.5 ms
169.4 ms
11.8
0.8 ms
294 MiB
pass
slow-phone (2 cores) / no-trigram
no-trigram
2
7,105
1.0 s
81.8 MiB
166.9 MiB
2.04x
0.6 ms
74.1 ms
20.4
0.5 ms
229 MiB
pass
slow-phone (2 cores) / compact
compact
2
7,112
1.0 s
81.8 MiB
166.9 MiB
2.04x
0.6 ms
73.2 ms
20.7
0.5 ms
267 MiB
pass
slow-phone (2 cores) / smallest
smallest
2
9,209
0.7 s
81.8 MiB
150.6 MiB
1.84x
0.4 ms
82.0 ms
23.7
0.5 ms
232 MiB
pass
profile: full (trigram=True, storeContent=True, phonetic=True, removeStopWords=False)
usable cores: 8 (Intel(R) Core(TM) i7-6850K CPU @ 3.60GHz @ 3597.912 MHz)
documents: 100,000 (+3 planted), batch size 25,000
indexing: 2,956 docs/s , OptimizeAsync 4.3 s
raw content: 85,821,349 bytes (81.8 MiB) — database: 374,808,576 bytes (357.4 MiB) — ratio 4.37x
peak process memory: 285 MiB
concurrent throughput: 21.0 queries/s (8 workers) — AddAsync 1.1 ms, RemoveAsync 0.2 ms
smoke checks: pass
pattern
hits
cold
p50
p95
max
needle, phonetic off
1
11.6 ms
5.5 ms
9.3 ms
9.3 ms
needle (uid; phonetic collides)
50
82.8 ms
82.2 ms
116.0 ms
116.0 ms
common term (~50% of docs)
50
160.0 ms
180.4 ms
226.2 ms
226.2 ms
common term, fuzzy off
50
174.7 ms
161.1 ms
211.1 ms
211.1 ms
common term + snippets
50
204.4 ms
175.5 ms
215.5 ms
215.5 ms
common term, page offset 100
50
182.9 ms
175.0 ms
204.1 ms
204.1 ms
de stemmed inflection
50
171.8 ms
183.1 ms
212.9 ms
212.9 ms
de phonetic variant
1
0.6 ms
0.3 ms
0.3 ms
0.3 ms
multi-token, all terms (default)
50
164.2 ms
177.7 ms
202.8 ms
202.8 ms
multi-token, any term
50
239.8 ms
237.8 ms
264.4 ms
264.4 ms
multi-token with typo (fallback)
50
87.9 ms
92.7 ms
115.5 ms
115.5 ms
short query (prefix aid)
50
293.5 ms
285.0 ms
355.1 ms
355.1 ms
no-match term
0
1.2 ms
0.2 ms
0.3 ms
0.3 ms
profile: no-trigram (trigram=False, storeContent=True, phonetic=True, removeStopWords=False)
usable cores: 8 (Intel(R) Core(TM) i7-6850K CPU @ 3.60GHz @ 3597.912 MHz)
documents: 100,000 (+3 planted), batch size 25,000
indexing: 6,410 docs/s , OptimizeAsync 1.1 s
raw content: 85,821,349 bytes (81.8 MiB) — database: 175,009,792 bytes (166.9 MiB) — ratio 2.04x
peak process memory: 282 MiB
concurrent throughput: 45.1 queries/s (8 workers) — AddAsync 0.7 ms, RemoveAsync 0.2 ms
smoke checks: pass
pattern
hits
cold
p50
p95
max
needle, phonetic off
1
4.2 ms
0.6 ms
2.7 ms
2.7 ms
needle (uid; phonetic collides)
50
80.7 ms
84.5 ms
132.7 ms
132.7 ms
common term (~50% of docs)
50
74.3 ms
76.0 ms
106.3 ms
106.3 ms
common term, fuzzy off
50
73.8 ms
67.0 ms
75.1 ms
75.1 ms
common term + snippets
50
73.0 ms
78.6 ms
104.2 ms
104.2 ms
common term, page offset 100
50
74.0 ms
76.4 ms
117.1 ms
117.1 ms
de stemmed inflection
50
53.8 ms
58.9 ms
75.0 ms
75.0 ms
de phonetic variant
1
0.5 ms
0.2 ms
0.3 ms
0.3 ms
multi-token, all terms (default)
50
78.9 ms
78.5 ms
91.7 ms
91.7 ms
multi-token, any term
50
143.3 ms
146.7 ms
158.2 ms
158.2 ms
multi-token with typo (fallback)
50
95.6 ms
87.4 ms
149.8 ms
149.8 ms
short query (prefix aid)
50
297.7 ms
281.5 ms
314.5 ms
314.5 ms
no-match term
0
0.8 ms
0.1 ms
0.3 ms
0.3 ms
profile: compact (trigram=False, storeContent=False, phonetic=True, removeStopWords=False)
usable cores: 8 (Intel(R) Core(TM) i7-6850K CPU @ 3.60GHz @ 3597.912 MHz)
documents: 100,000 (+3 planted), batch size 25,000
indexing: 6,443 docs/s , OptimizeAsync 1.0 s
raw content: 85,821,349 bytes (81.8 MiB) — database: 175,009,792 bytes (166.9 MiB) — ratio 2.04x
peak process memory: 310 MiB
concurrent throughput: 44.5 queries/s (8 workers) — AddAsync 0.5 ms, RemoveAsync 0.2 ms
smoke checks: pass
pattern
hits
cold
p50
p95
max
needle, phonetic off
1
2.9 ms
0.4 ms
2.0 ms
2.0 ms
needle (uid; phonetic collides)
50
74.4 ms
75.2 ms
98.8 ms
98.8 ms
common term (~50% of docs)
50
78.4 ms
79.8 ms
93.6 ms
93.6 ms
common term, fuzzy off
50
73.9 ms
65.7 ms
99.0 ms
99.0 ms
common term + snippets
50
117.0 ms
78.2 ms
92.4 ms
92.4 ms
common term, page offset 100
50
97.9 ms
84.0 ms
113.7 ms
113.7 ms
de stemmed inflection
50
59.3 ms
60.0 ms
75.8 ms
75.8 ms
de phonetic variant
1
0.5 ms
0.2 ms
0.2 ms
0.2 ms
multi-token, all terms (default)
50
76.9 ms
77.7 ms
97.7 ms
97.7 ms
multi-token, any term
50
134.0 ms
140.3 ms
149.2 ms
149.2 ms
multi-token with typo (fallback)
50
83.2 ms
98.3 ms
139.4 ms
139.4 ms
short query (prefix aid)
50
324.5 ms
299.6 ms
329.0 ms
329.0 ms
no-match term
0
0.9 ms
0.1 ms
0.2 ms
0.2 ms
profile: smallest (trigram=False, storeContent=False, phonetic=False, removeStopWords=True)
usable cores: 8 (Intel(R) Core(TM) i7-6850K CPU @ 3.60GHz @ 3597.912 MHz)
documents: 100,000 (+3 planted), batch size 25,000
indexing: 7,655 docs/s , OptimizeAsync 0.8 s
raw content: 85,821,349 bytes (81.8 MiB) — database: 157,958,144 bytes (150.6 MiB) — ratio 1.84x
peak process memory: 274 MiB
concurrent throughput: 55.1 queries/s (8 workers) — AddAsync 0.5 ms, RemoveAsync 0.2 ms
smoke checks: pass
pattern
hits
cold
p50
p95
max
needle, phonetic off
1
3.1 ms
0.4 ms
3.6 ms
3.6 ms
needle (uid; phonetic collides)
1
0.8 ms
0.4 ms
0.4 ms
0.4 ms
common term (~50% of docs)
50
106.2 ms
77.6 ms
118.5 ms
118.5 ms
common term, fuzzy off
50
51.7 ms
53.0 ms
63.8 ms
63.8 ms
common term + snippets
50
65.3 ms
65.3 ms
73.2 ms
73.2 ms
common term, page offset 100
50
85.9 ms
67.3 ms
82.9 ms
82.9 ms
de stemmed inflection
50
59.4 ms
57.3 ms
72.3 ms
72.3 ms
de phonetic variant
0
0.8 ms
0.1 ms
0.1 ms
0.1 ms
multi-token, all terms (default)
50
79.8 ms
94.4 ms
128.4 ms
128.4 ms
multi-token, any term
50
108.7 ms
126.8 ms
155.7 ms
155.7 ms
multi-token with typo (fallback)
50
74.0 ms
68.6 ms
89.5 ms
89.5 ms
short query (prefix aid)
50
301.6 ms
281.7 ms
305.2 ms
305.2 ms
no-match term
0
0.3 ms
0.1 ms
0.1 ms
0.1 ms
fast-phone (4 cores) / full
profile: full (trigram=True, storeContent=True, phonetic=True, removeStopWords=False)
usable cores: 4 (Intel(R) Core(TM) i7-6850K CPU @ 3.60GHz @ 3597.912 MHz)
documents: 100,000 (+3 planted), batch size 25,000
indexing: 3,158 docs/s , OptimizeAsync 4.3 s
raw content: 85,821,349 bytes (81.8 MiB) — database: 374,808,576 bytes (357.4 MiB) — ratio 4.37x
peak process memory: 311 MiB
concurrent throughput: 16.8 queries/s (8 workers) — AddAsync 0.7 ms, RemoveAsync 0.2 ms
smoke checks: pass
pattern
hits
cold
p50
p95
max
needle, phonetic off
1
10.1 ms
5.2 ms
8.5 ms
8.5 ms
needle (uid; phonetic collides)
50
76.4 ms
80.9 ms
132.8 ms
132.8 ms
common term (~50% of docs)
50
169.4 ms
173.0 ms
287.7 ms
287.7 ms
common term, fuzzy off
50
164.6 ms
166.8 ms
203.2 ms
203.2 ms
common term + snippets
50
176.1 ms
178.6 ms
193.3 ms
193.3 ms
common term, page offset 100
50
162.8 ms
165.3 ms
192.6 ms
192.6 ms
de stemmed inflection
50
173.8 ms
167.8 ms
191.1 ms
191.1 ms
de phonetic variant
1
0.6 ms
0.2 ms
0.3 ms
0.3 ms
multi-token, all terms (default)
50
165.0 ms
168.0 ms
204.9 ms
204.9 ms
multi-token, any term
50
230.9 ms
232.0 ms
283.1 ms
283.1 ms
multi-token with typo (fallback)
50
87.4 ms
92.6 ms
135.7 ms
135.7 ms
short query (prefix aid)
50
287.6 ms
286.8 ms
324.1 ms
324.1 ms
no-match term
0
0.8 ms
0.1 ms
0.2 ms
0.2 ms
fast-phone (4 cores) / no-trigram
profile: no-trigram (trigram=False, storeContent=True, phonetic=True, removeStopWords=False)
usable cores: 4 (Intel(R) Core(TM) i7-6850K CPU @ 3.60GHz @ 3597.912 MHz)
documents: 100,000 (+3 planted), batch size 25,000
indexing: 6,677 docs/s , OptimizeAsync 1.0 s
raw content: 85,821,349 bytes (81.8 MiB) — database: 175,009,792 bytes (166.9 MiB) — ratio 2.04x
peak process memory: 289 MiB
concurrent throughput: 31.7 queries/s (8 workers) — AddAsync 0.5 ms, RemoveAsync 0.2 ms
smoke checks: pass
pattern
hits
cold
p50
p95
max
needle, phonetic off
1
3.0 ms
0.3 ms
1.9 ms
1.9 ms
needle (uid; phonetic collides)
50
73.3 ms
92.6 ms
115.0 ms
115.0 ms
common term (~50% of docs)
50
78.1 ms
76.6 ms
105.4 ms
105.4 ms
common term, fuzzy off
50
75.4 ms
62.0 ms
74.5 ms
74.5 ms
common term + snippets
50
70.3 ms
76.2 ms
108.4 ms
108.4 ms
common term, page offset 100
50
92.7 ms
75.2 ms
83.7 ms
83.7 ms
de stemmed inflection
50
57.2 ms
54.5 ms
82.9 ms
82.9 ms
de phonetic variant
1
0.5 ms
0.2 ms
0.2 ms
0.2 ms
multi-token, all terms (default)
50
85.1 ms
78.7 ms
90.0 ms
90.0 ms
multi-token, any term
50
141.8 ms
145.4 ms
200.1 ms
200.1 ms
multi-token with typo (fallback)
50
90.3 ms
90.8 ms
107.7 ms
107.7 ms
short query (prefix aid)
50
309.9 ms
280.0 ms
384.1 ms
384.1 ms
no-match term
0
0.8 ms
0.2 ms
0.3 ms
0.3 ms
fast-phone (4 cores) / compact
profile: compact (trigram=False, storeContent=False, phonetic=True, removeStopWords=False)
usable cores: 4 (Intel(R) Core(TM) i7-6850K CPU @ 3.60GHz @ 3597.912 MHz)
documents: 100,000 (+3 planted), batch size 25,000
indexing: 6,674 docs/s , OptimizeAsync 1.0 s
raw content: 85,821,349 bytes (81.8 MiB) — database: 175,009,792 bytes (166.9 MiB) — ratio 2.04x
peak process memory: 278 MiB
concurrent throughput: 31.4 queries/s (8 workers) — AddAsync 0.6 ms, RemoveAsync 0.2 ms
smoke checks: pass
pattern
hits
cold
p50
p95
max
needle, phonetic off
1
2.9 ms
0.4 ms
1.9 ms
1.9 ms
needle (uid; phonetic collides)
50
70.7 ms
86.3 ms
128.9 ms
128.9 ms
common term (~50% of docs)
50
81.6 ms
71.8 ms
102.2 ms
102.2 ms
common term, fuzzy off
50
69.5 ms
62.8 ms
82.4 ms
82.4 ms
common term + snippets
50
69.3 ms
74.1 ms
81.9 ms
81.9 ms
common term, page offset 100
50
94.6 ms
80.3 ms
89.8 ms
89.8 ms
de stemmed inflection
50
55.3 ms
58.8 ms
63.4 ms
63.4 ms
de phonetic variant
1
0.5 ms
0.3 ms
0.3 ms
0.3 ms
multi-token, all terms (default)
50
94.0 ms
81.5 ms
102.3 ms
102.3 ms
multi-token, any term
50
154.4 ms
156.0 ms
187.4 ms
187.4 ms
multi-token with typo (fallback)
50
92.4 ms
92.9 ms
120.3 ms
120.3 ms
short query (prefix aid)
50
293.0 ms
297.7 ms
395.5 ms
395.5 ms
no-match term
0
0.9 ms
0.1 ms
0.1 ms
0.1 ms
fast-phone (4 cores) / smallest
profile: smallest (trigram=False, storeContent=False, phonetic=False, removeStopWords=True)
usable cores: 4 (Intel(R) Core(TM) i7-6850K CPU @ 3.60GHz @ 3597.912 MHz)
documents: 100,000 (+3 planted), batch size 25,000
indexing: 8,441 docs/s , OptimizeAsync 0.6 s
raw content: 85,821,349 bytes (81.8 MiB) — database: 157,958,144 bytes (150.6 MiB) — ratio 1.84x
peak process memory: 216 MiB
concurrent throughput: 38.1 queries/s (8 workers) — AddAsync 0.5 ms, RemoveAsync 0.2 ms
smoke checks: pass
pattern
hits
cold
p50
p95
max
needle, phonetic off
1
4.2 ms
0.4 ms
3.6 ms
3.6 ms
needle (uid; phonetic collides)
1
0.9 ms
0.4 ms
0.5 ms
0.5 ms
common term (~50% of docs)
50
84.9 ms
78.0 ms
114.4 ms
114.4 ms
common term, fuzzy off
50
58.9 ms
56.4 ms
69.1 ms
69.1 ms
common term + snippets
50
63.7 ms
69.7 ms
89.4 ms
89.4 ms
common term, page offset 100
50
67.7 ms
67.4 ms
84.7 ms
84.7 ms
de stemmed inflection
50
65.6 ms
60.8 ms
69.1 ms
69.1 ms
de phonetic variant
0
0.8 ms
0.1 ms
0.1 ms
0.1 ms
multi-token, all terms (default)
50
83.1 ms
79.4 ms
98.5 ms
98.5 ms
multi-token, any term
50
114.9 ms
114.9 ms
124.3 ms
124.3 ms
multi-token with typo (fallback)
50
70.1 ms
75.2 ms
92.1 ms
92.1 ms
short query (prefix aid)
50
280.1 ms
273.8 ms
292.3 ms
292.3 ms
no-match term
0
0.4 ms
0.1 ms
0.1 ms
0.1 ms
slow-phone (2 cores) / full
profile: full (trigram=True, storeContent=True, phonetic=True, removeStopWords=False)
usable cores: 2 (Intel(R) Core(TM) i7-6850K CPU @ 3.60GHz @ 3597.912 MHz)
documents: 100,000 (+3 planted), batch size 25,000
indexing: 3,320 docs/s , OptimizeAsync 4.4 s
raw content: 85,821,349 bytes (81.8 MiB) — database: 374,808,576 bytes (357.4 MiB) — ratio 4.37x
peak process memory: 294 MiB
concurrent throughput: 11.8 queries/s (8 workers) — AddAsync 0.8 ms, RemoveAsync 0.3 ms
smoke checks: pass
pattern
hits
cold
p50
p95
max
needle, phonetic off
1
8.7 ms
5.5 ms
7.4 ms
7.4 ms
needle (uid; phonetic collides)
50
79.2 ms
77.5 ms
122.1 ms
122.1 ms
common term (~50% of docs)
50
171.7 ms
169.4 ms
204.1 ms
204.1 ms
common term, fuzzy off
50
148.7 ms
158.1 ms
208.9 ms
208.9 ms
common term + snippets
50
180.3 ms
170.6 ms
196.5 ms
196.5 ms
common term, page offset 100
50
173.3 ms
179.0 ms
197.1 ms
197.1 ms
de stemmed inflection
50
180.3 ms
184.0 ms
217.1 ms
217.1 ms
de phonetic variant
1
0.6 ms
0.3 ms
0.3 ms
0.3 ms
multi-token, all terms (default)
50
168.7 ms
179.3 ms
218.5 ms
218.5 ms
multi-token, any term
50
228.5 ms
231.1 ms
268.8 ms
268.8 ms
multi-token with typo (fallback)
50
96.5 ms
91.5 ms
112.0 ms
112.0 ms
short query (prefix aid)
50
302.6 ms
286.9 ms
335.9 ms
335.9 ms
no-match term
0
0.8 ms
0.1 ms
0.2 ms
0.2 ms
slow-phone (2 cores) / no-trigram
profile: no-trigram (trigram=False, storeContent=True, phonetic=True, removeStopWords=False)
usable cores: 2 (Intel(R) Core(TM) i7-6850K CPU @ 3.60GHz @ 3597.912 MHz)
documents: 100,000 (+3 planted), batch size 25,000
indexing: 7,105 docs/s , OptimizeAsync 1.0 s
raw content: 85,821,349 bytes (81.8 MiB) — database: 175,009,792 bytes (166.9 MiB) — ratio 2.04x
peak process memory: 229 MiB
concurrent throughput: 20.4 queries/s (8 workers) — AddAsync 0.5 ms, RemoveAsync 0.2 ms
smoke checks: pass
pattern
hits
cold
p50
p95
max
needle, phonetic off
1
4.3 ms
0.6 ms
2.6 ms
2.6 ms
needle (uid; phonetic collides)
50
99.8 ms
84.5 ms
153.1 ms
153.1 ms
common term (~50% of docs)
50
75.3 ms
74.1 ms
99.9 ms
99.9 ms
common term, fuzzy off
50
62.9 ms
64.3 ms
72.9 ms
72.9 ms
common term + snippets
50
82.0 ms
73.6 ms
104.0 ms
104.0 ms
common term, page offset 100
50
76.1 ms
74.7 ms
97.0 ms
97.0 ms
de stemmed inflection
50
69.4 ms
57.2 ms
78.2 ms
78.2 ms
de phonetic variant
1
0.5 ms
0.2 ms
0.2 ms
0.2 ms
multi-token, all terms (default)
50
72.4 ms
84.9 ms
96.3 ms
96.3 ms
multi-token, any term
50
176.3 ms
151.6 ms
174.5 ms
174.5 ms
multi-token with typo (fallback)
50
94.7 ms
89.0 ms
105.3 ms
105.3 ms
short query (prefix aid)
50
300.3 ms
285.5 ms
319.3 ms
319.3 ms
no-match term
0
0.7 ms
0.1 ms
0.1 ms
0.1 ms
slow-phone (2 cores) / compact
profile: compact (trigram=False, storeContent=False, phonetic=True, removeStopWords=False)
usable cores: 2 (Intel(R) Core(TM) i7-6850K CPU @ 3.60GHz @ 3597.912 MHz)
documents: 100,000 (+3 planted), batch size 25,000
indexing: 7,112 docs/s , OptimizeAsync 1.0 s
raw content: 85,821,349 bytes (81.8 MiB) — database: 175,009,792 bytes (166.9 MiB) — ratio 2.04x
peak process memory: 267 MiB
concurrent throughput: 20.7 queries/s (8 workers) — AddAsync 0.5 ms, RemoveAsync 0.2 ms
smoke checks: pass
pattern
hits
cold
p50
p95
max
needle, phonetic off
1
3.8 ms
0.6 ms
2.5 ms
2.5 ms
needle (uid; phonetic collides)
50
109.9 ms
68.5 ms
125.9 ms
125.9 ms
common term (~50% of docs)
50
75.6 ms
73.2 ms
99.9 ms
99.9 ms
common term, fuzzy off
50
58.8 ms
64.2 ms
94.5 ms
94.5 ms
common term + snippets
50
74.7 ms
75.4 ms
88.4 ms
88.4 ms
common term, page offset 100
50
76.6 ms
74.6 ms
88.6 ms
88.6 ms
de stemmed inflection
50
57.5 ms
59.8 ms
76.0 ms
76.0 ms
de phonetic variant
1
0.5 ms
0.2 ms
0.3 ms
0.3 ms
multi-token, all terms (default)
50
78.0 ms
80.1 ms
102.6 ms
102.6 ms
multi-token, any term
50
137.9 ms
144.5 ms
162.4 ms
162.4 ms
multi-token with typo (fallback)
50
90.0 ms
92.1 ms
113.1 ms
113.1 ms
short query (prefix aid)
50
282.3 ms
283.1 ms
310.6 ms
310.6 ms
no-match term
0
0.8 ms
0.1 ms
0.2 ms
0.2 ms
slow-phone (2 cores) / smallest
profile: smallest (trigram=False, storeContent=False, phonetic=False, removeStopWords=True)
usable cores: 2 (Intel(R) Core(TM) i7-6850K CPU @ 3.60GHz @ 3597.912 MHz)
documents: 100,000 (+3 planted), batch size 25,000
indexing: 9,209 docs/s , OptimizeAsync 0.7 s
raw content: 85,821,349 bytes (81.8 MiB) — database: 157,958,144 bytes (150.6 MiB) — ratio 1.84x
peak process memory: 232 MiB
concurrent throughput: 23.7 queries/s (8 workers) — AddAsync 0.5 ms, RemoveAsync 0.2 ms
smoke checks: pass
pattern
hits
cold
p50
p95
max
needle, phonetic off
1
2.9 ms
0.4 ms
3.8 ms
3.8 ms
needle (uid; phonetic collides)
1
0.8 ms
0.4 ms
0.4 ms
0.4 ms
common term (~50% of docs)
50
84.6 ms
82.0 ms
95.3 ms
95.3 ms
common term, fuzzy off
50
65.3 ms
54.2 ms
83.5 ms
83.5 ms
common term + snippets
50
69.2 ms
66.7 ms
73.3 ms
73.3 ms
common term, page offset 100
50
87.1 ms
66.0 ms
92.2 ms
92.2 ms
de stemmed inflection
50
49.7 ms
52.2 ms
59.7 ms
59.7 ms
de phonetic variant
0
0.7 ms
0.1 ms
0.3 ms
0.3 ms
multi-token, all terms (default)
50
91.7 ms
76.7 ms
100.2 ms
100.2 ms
multi-token, any term
50
109.2 ms
122.1 ms
140.2 ms
140.2 ms
multi-token with typo (fallback)
50
71.5 ms
67.7 ms
86.1 ms
86.1 ms
short query (prefix aid)
50
274.4 ms
270.7 ms
306.2 ms
306.2 ms
no-match term
0
0.4 ms
0.1 ms
0.1 ms
0.1 ms