Skip to content

Commit cad2bbb

Browse files
ethanglaserethanglaser
committed
move all tools/benchmark_inputs to public
1 parent a95c2ab commit cad2bbb

File tree

7 files changed

+1374
-0
lines changed

7 files changed

+1374
-0
lines changed

‎tools/benchmark_inputs/inverted/static-lvq-regression.toml‎

Lines changed: 129 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,129 @@
1+
# Copyright 2024 Intel Corporation
2+
#
3+
# Licensed under the Apache License, Version 2.0 (the "License");
4+
# you may not use this file except in compliance with the License.
5+
# You may obtain a copy of the License at
6+
#
7+
# http://www.apache.org/licenses/LICENSE-2.0
8+
#
9+
# Unless required by applicable law or agreed to in writing, software
10+
# distributed under the License is distributed on an "AS IS" BASIS,
11+
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
12+
# See the License for the specific language governing permissions and
13+
# limitations under the License.
14+
15+
[[inverted_static_memory_build]]
16+
__schema__ = 'benchmark_inverted_memory_build_job'
17+
__version__ = 'v0.0.1'
18+
centroids_directory = ''
19+
data = 'dpr/dpr_1m_f16.fvecs'
20+
data_type = 'float16'
21+
description = 'dpr-1m'
22+
distance = 'MIP'
23+
groundtruth = 'dpr/dpr_1m_groundtruth_ip_10k.ivecs'
24+
ndims = 768
25+
num_build_threads = 72
26+
queries = 'dpr/dpr_queries_10k_f32.fvecs'
27+
query_type = 'float32'
28+
save_directory = ''
29+
search_configs = []
30+
strategy = 'sparse'
31+
32+
[inverted_static_memory_build.clustering_parameters]
33+
__schema__ = 'clustering_parameters'
34+
__version__ = 'v0.0.0'
35+
batchsize = 100000
36+
epsilon = 1.0
37+
max_cluster_size = 0
38+
max_replicas = 7
39+
num_intermediate_results = 64
40+
percent_centroids = 0.12
41+
refinement_alpha = 0.94999999999999996
42+
search_window_size = 50
43+
seed = 12648430
44+
45+
[inverted_static_memory_build.dataset]
46+
__schema__ = 'benchmark_dataset_abstract'
47+
__version__ = 'v0.0.0'
48+
kind = 'uncompressed'
49+
50+
[inverted_static_memory_build.dataset.dataset]
51+
__schema__ = 'benchmark_dataset_uncompressed'
52+
__version__ = 'v0.0.0'
53+
data_type = 'float16'
54+
55+
[inverted_static_memory_build.primary_build_parameters]
56+
__schema__ = 'vamana_build_parameters'
57+
__version__ = 'v0.0.1'
58+
alpha = 0.94999999999999996
59+
graph_max_degree = 64
60+
max_candidate_pool_size = 750
61+
name = 'vamana build parameters'
62+
prune_to = 60
63+
use_full_search_history = true
64+
window_size = 200
65+
66+
[inverted_static_memory_build.search_parameters]
67+
__schema__ = 'benchmark_search_parameters'
68+
__version__ = 'v0.0.0'
69+
num_neighbors = 10
70+
target_recalls = [ 0.80000000000000004, 0.84999999999999998, 0.90000000000000002 ]
71+
72+
[[inverted_static_memory_build]]
73+
__schema__ = 'benchmark_inverted_memory_build_job'
74+
__version__ = 'v0.0.1'
75+
centroids_directory = ''
76+
data = 'dpr/dpr_1m_f16.fvecs'
77+
data_type = 'float16'
78+
description = 'dpr-1m'
79+
distance = 'MIP'
80+
groundtruth = 'dpr/dpr_1m_groundtruth_ip_10k.ivecs'
81+
ndims = 768
82+
num_build_threads = 72
83+
queries = 'dpr/dpr_queries_10k_f32.fvecs'
84+
query_type = 'float32'
85+
save_directory = ''
86+
search_configs = []
87+
strategy = 'sparse'
88+
89+
[inverted_static_memory_build.clustering_parameters]
90+
__schema__ = 'clustering_parameters'
91+
__version__ = 'v0.0.0'
92+
batchsize = 100000
93+
epsilon = 1.0
94+
max_cluster_size = 0
95+
max_replicas = 8
96+
num_intermediate_results = 64
97+
percent_centroids = 0.12
98+
refinement_alpha = 0.94999999999999996
99+
search_window_size = 50
100+
seed = 12648430
101+
102+
[inverted_static_memory_build.dataset]
103+
__schema__ = 'benchmark_dataset_abstract'
104+
__version__ = 'v0.0.0'
105+
kind = 'lvq'
106+
107+
[inverted_static_memory_build.dataset.dataset]
108+
__schema__ = 'benchmark_dataset_lvq'
109+
__version__ = 'v0.0.0'
110+
primary = 8
111+
residual = 0
112+
strategy = 'sequential'
113+
114+
[inverted_static_memory_build.primary_build_parameters]
115+
__schema__ = 'vamana_build_parameters'
116+
__version__ = 'v0.0.1'
117+
alpha = 0.94999999999999996
118+
graph_max_degree = 64
119+
max_candidate_pool_size = 750
120+
name = 'vamana build parameters'
121+
prune_to = 60
122+
use_full_search_history = true
123+
window_size = 200
124+
125+
[inverted_static_memory_build.search_parameters]
126+
__schema__ = 'benchmark_search_parameters'
127+
__version__ = 'v0.0.0'
128+
num_neighbors = 10
129+
target_recalls = [ 0.90000000000000002 ]

‎tools/benchmark_inputs/vamana/iterator-input.toml‎

Lines changed: 91 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -100,3 +100,94 @@ query_type = 'float32'
100100
scale_buffer_capacity = 10
101101
scale_search_window = 10
102102
restart_searches = false
103+
104+
[[vamana_iterator_v1]]
105+
__schema__ = 'svsbenchmark_vamana_iterator'
106+
__version__ = 'v0.0.0'
107+
config = '/export/data/datasets/indexes/dpr-1m/config/'
108+
data = '/export/data/datasets/indexes/dpr-1m/data/'
109+
distance = 'MIP'
110+
graph = '/export/data/datasets/indexes/dpr-1m/graph/'
111+
groundtruth = '/export/data/datasets/dpr/dpr_1m_groundtruth_ip_10k_1000neighors.ivecs'
112+
ndims = 768
113+
queries = '/export/data/datasets/dpr/dpr_queries_10k_f32.fvecs'
114+
query_type = 'float32'
115+
116+
[vamana_iterator_v1.dataset]
117+
__schema__ = 'benchmark_dataset_abstract'
118+
__version__ = 'v0.0.0'
119+
kind = 'lvq'
120+
121+
[vamana_iterator_v1.dataset.dataset]
122+
__schema__ = 'benchmark_dataset_lvq'
123+
__version__ = 'v0.0.0'
124+
primary = 4
125+
residual = 8
126+
strategy = "turbo<16x8>"
127+
128+
[vamana_iterator_v1.parameters]
129+
__schema__ = 'svsbenchamrk_isp'
130+
__version__ = 'v0.0.0'
131+
num_batches = 10
132+
query_subsample = 1000
133+
target_recalls = [ 0.9 ]
134+
135+
[[vamana_iterator_v1.parameters.schedules]]
136+
__schema__ = 'svsbench_vamana_iter_schedule'
137+
__version__ = 'v0.0.0'
138+
batch_size_start = 10
139+
enable_filter_after = -1
140+
scale_batch_size = 0
141+
scale_buffer_capacity = 10
142+
scale_search_window = 10
143+
restart_searches = true
144+
145+
[[vamana_iterator_v1.parameters.schedules]]
146+
__schema__ = 'svsbench_vamana_iter_schedule'
147+
__version__ = 'v0.0.0'
148+
batch_size_start = 10
149+
enable_filter_after = -1
150+
scale_batch_size = 0
151+
scale_buffer_capacity = 10
152+
scale_search_window = 10
153+
restart_searches = false
154+
155+
[[vamana_iterator_v1.parameters.schedules]]
156+
__schema__ = 'svsbench_vamana_iter_schedule'
157+
__version__ = 'v0.0.0'
158+
batch_size_start = 10
159+
enable_filter_after = 0
160+
scale_batch_size = 0
161+
scale_buffer_capacity = 10
162+
scale_search_window = 10
163+
restart_searches = false
164+
165+
[[vamana_iterator_v1.parameters.schedules]]
166+
__schema__ = 'svsbench_vamana_iter_schedule'
167+
__version__ = 'v0.0.0'
168+
batch_size_start = 10
169+
enable_filter_after = 1
170+
scale_batch_size = 0
171+
scale_buffer_capacity = 10
172+
scale_search_window = 10
173+
restart_searches = false
174+
175+
[[vamana_iterator_v1.parameters.schedules]]
176+
__schema__ = 'svsbench_vamana_iter_schedule'
177+
__version__ = 'v0.0.0'
178+
batch_size_start = 10
179+
enable_filter_after = 2
180+
scale_batch_size = 0
181+
scale_buffer_capacity = 10
182+
scale_search_window = 10
183+
restart_searches = false
184+
185+
[[vamana_iterator_v1.parameters.schedules]]
186+
__schema__ = 'svsbench_vamana_iter_schedule'
187+
__version__ = 'v0.0.0'
188+
batch_size_start = 10
189+
enable_filter_after = 3
190+
scale_batch_size = 0
191+
scale_buffer_capacity = 10
192+
scale_search_window = 10
193+
restart_searches = false

0 commit comments

Comments
 (0)