From 8731d137c54d5d7911583a6240b7ce44813867f1 Mon Sep 17 00:00:00 2001 From: Vissidarte-Herman Date: Tue, 26 Dec 2023 13:44:28 +0800 Subject: [PATCH 1/7] Added a benchmark report on sift --- docs/benchmark.md | 25 ++++++++++++++++--------- 1 file changed, 16 insertions(+), 9 deletions(-) diff --git a/docs/benchmark.md b/docs/benchmark.md index c8f50edf0a..a41223fcc3 100644 --- a/docs/benchmark.md +++ b/docs/benchmark.md @@ -1,6 +1,6 @@ # Benchmark -**Infinity** provides Python script for sift1m and gist1m dataset benchmark. +Infinity provides a Python script for benchmarking the SIFT1M and GIST1M datasets. ## Get the Infinity binary file @@ -9,9 +9,9 @@ git clone https://github.com/infiniflow/infinity.git cd infinity ``` -## Download the benchmark file +## Download the benchmark files -Download via wget. +To obtain the benchmark files, you have the option to download them using the wget command. ```sh #download sift benchmark @@ -21,16 +21,16 @@ wget ftp://ftp.irisa.fr/local/texmex/corpus/gist.tar.gz ``` -or visit [http://corpus-texmex.irisa.fr/](http://corpus-texmex.irisa.fr/) to download manually. +Alternatively, you can manually download the benchmark files by visiting [http://corpus-texmex.irisa.fr/](http://corpus-texmex.irisa.fr/). ```sh -#uncompress and move benchmark file +# Unzip and move the SIFT1M benchmark file. tar -zxvf sift.tar.gz mv sift/sift_base.fvecs test/data/benchmark/sift_1m/sift_base.fvecs mv sift/sift_query.fvecs test/data/benchmark/sift_1m/sift_query.fvecs mv sift/sift_groundtruth.ivecs test/data/benchmark/sift_1m/sift_groundtruth.ivecs - +# Unzip and move the GIST1M benchmark file. tar -zxvf gist.tar.gz mv gist/gist_base.fvecs test/data/benchmark/gist_1m/gist_base.fvecs mv gist/gist_query.fvecs test/data/benchmark/gist_1m/gist_query.fvecs @@ -74,10 +74,17 @@ options: -r ROUNDS, --rounds ROUNDS -d DATA_SET, --data DATA_SET -# ROUNDS refers to the number of times that Python runs the benchmark. The result is the average time for all runs. +# ROUNDS indicates the number of times Python executes the benchmark, and the result represents the average duration for each run. -# The following command means run benchmark with one thread, for one time using the sift dataset. +# Perform a benchmark on the SIFT1M dataset using a single thread, running it only once. python remote_benchmark.py -t 1 -r 1 -d sift_1m - +# Perform a benchmark on the GIST1M dataset using a single thread, running it only once. python remote_benchmark.py -t 1 -r 1 -d gist_1m ``` +## A SIFT benchmark report + +- **Hardware**: Intel i5-12500H, 16C, 16GB +- **Operating system**: Ubuntu 22.04 +- **Dataset**: SIFT1M; **topk**: 100; **recall**: 97%+ +- **QPS**: 10,305 +- **P99 Latency**: 0.3 ms \ No newline at end of file From 8e45122576a034a957936450cdfc665358e82b32 Mon Sep 17 00:00:00 2001 From: Vissidarte-Herman Date: Tue, 26 Dec 2023 14:10:21 +0800 Subject: [PATCH 2/7] Added Build Infinity --- .history/docs/benchmark_20231226133852.md | 90 ++++++++++++++++++++++ .history/docs/benchmark_20231226140016.md | 92 +++++++++++++++++++++++ .history/docs/benchmark_20231226140905.md | 86 +++++++++++++++++++++ .history/docs/benchmark_20231226140937.md | 86 +++++++++++++++++++++ .history/docs/benchmark_20231226141001.md | 86 +++++++++++++++++++++ docs/benchmark.md | 36 ++++----- 6 files changed, 456 insertions(+), 20 deletions(-) create mode 100644 .history/docs/benchmark_20231226133852.md create mode 100644 .history/docs/benchmark_20231226140016.md create mode 100644 .history/docs/benchmark_20231226140905.md create mode 100644 .history/docs/benchmark_20231226140937.md create mode 100644 .history/docs/benchmark_20231226141001.md diff --git a/.history/docs/benchmark_20231226133852.md b/.history/docs/benchmark_20231226133852.md new file mode 100644 index 0000000000..a41223fcc3 --- /dev/null +++ b/.history/docs/benchmark_20231226133852.md @@ -0,0 +1,90 @@ +# Benchmark + +Infinity provides a Python script for benchmarking the SIFT1M and GIST1M datasets. + +## Get the Infinity binary file + +```sh +git clone https://github.com/infiniflow/infinity.git +cd infinity +``` + +## Download the benchmark files + +To obtain the benchmark files, you have the option to download them using the wget command. + +```sh +#download sift benchmark +wget ftp://ftp.irisa.fr/local/texmex/corpus/sift.tar.gz +#download gist benchmark +wget ftp://ftp.irisa.fr/local/texmex/corpus/gist.tar.gz + +``` + +Alternatively, you can manually download the benchmark files by visiting [http://corpus-texmex.irisa.fr/](http://corpus-texmex.irisa.fr/). + +```sh +# Unzip and move the SIFT1M benchmark file. +tar -zxvf sift.tar.gz +mv sift/sift_base.fvecs test/data/benchmark/sift_1m/sift_base.fvecs +mv sift/sift_query.fvecs test/data/benchmark/sift_1m/sift_query.fvecs +mv sift/sift_groundtruth.ivecs test/data/benchmark/sift_1m/sift_groundtruth.ivecs + +# Unzip and move the GIST1M benchmark file. +tar -zxvf gist.tar.gz +mv gist/gist_base.fvecs test/data/benchmark/gist_1m/gist_base.fvecs +mv gist/gist_query.fvecs test/data/benchmark/gist_1m/gist_query.fvecs +mv gist/gist_groundtruth.ivecs test/data/benchmark/gist_1m/gist_groundtruth.ivecs + +``` + +## Benchmark dependencies + +```sh +cd python + +pip install -r requirements.txt +python setup.py bdist_wheel +pip install dist/infinity_sdk-0.1.0.dev1-py3-none-any.whl +``` + +## Start Infinity + +See the [README.md](https://github.com/infiniflow/infinity/blob/main/README.md) to start Infinity. + +## Import data + +```sh +cd benchmark + +options: + -h, --help show this help message and exit + -d DATA_SET, --data DATA_SET + +python remote_benchmark_import.py -d sift_1m +python remote_benchmark_import.py -d gist_1m +``` + +## Run benchmark + +```sh +options: + -h, --help show this help message and exit + -t THREADS, --threads THREADS + -r ROUNDS, --rounds ROUNDS + -d DATA_SET, --data DATA_SET + +# ROUNDS indicates the number of times Python executes the benchmark, and the result represents the average duration for each run. + +# Perform a benchmark on the SIFT1M dataset using a single thread, running it only once. +python remote_benchmark.py -t 1 -r 1 -d sift_1m +# Perform a benchmark on the GIST1M dataset using a single thread, running it only once. +python remote_benchmark.py -t 1 -r 1 -d gist_1m +``` +## A SIFT benchmark report + +- **Hardware**: Intel i5-12500H, 16C, 16GB +- **Operating system**: Ubuntu 22.04 +- **Dataset**: SIFT1M; **topk**: 100; **recall**: 97%+ +- **QPS**: 10,305 +- **P99 Latency**: 0.3 ms \ No newline at end of file diff --git a/.history/docs/benchmark_20231226140016.md b/.history/docs/benchmark_20231226140016.md new file mode 100644 index 0000000000..174dd1059b --- /dev/null +++ b/.history/docs/benchmark_20231226140016.md @@ -0,0 +1,92 @@ +# Benchmark + +Infinity provides a Python script for benchmarking the SIFT1M and GIST1M datasets. + +## Build Infinity + +You can either build Infinity + +```sh +git clone https://github.com/infiniflow/infinity.git +cd infinity +``` + +## Download the Benchmark datasets + +To obtain the benchmark files, you have the option to download them using the wget command. + +```sh +#download sift benchmark +wget ftp://ftp.irisa.fr/local/texmex/corpus/sift.tar.gz +#download gist benchmark +wget ftp://ftp.irisa.fr/local/texmex/corpus/gist.tar.gz + +``` + +Alternatively, you can manually download the benchmark files by visiting [http://corpus-texmex.irisa.fr/](http://corpus-texmex.irisa.fr/). + +```sh +# Unzip and move the SIFT1M benchmark file. +tar -zxvf sift.tar.gz +mv sift/sift_base.fvecs test/data/benchmark/sift_1m/sift_base.fvecs +mv sift/sift_query.fvecs test/data/benchmark/sift_1m/sift_query.fvecs +mv sift/sift_groundtruth.ivecs test/data/benchmark/sift_1m/sift_groundtruth.ivecs + +# Unzip and move the GIST1M benchmark file. +tar -zxvf gist.tar.gz +mv gist/gist_base.fvecs test/data/benchmark/gist_1m/gist_base.fvecs +mv gist/gist_query.fvecs test/data/benchmark/gist_1m/gist_query.fvecs +mv gist/gist_groundtruth.ivecs test/data/benchmark/gist_1m/gist_groundtruth.ivecs + +``` + +## Benchmark dependencies + +```sh +cd python + +pip install -r requirements.txt +python setup.py bdist_wheel +pip install dist/infinity_sdk-0.1.0.dev1-py3-none-any.whl +``` + +## Start Infinity + +See the [README.md](https://github.com/infiniflow/infinity/blob/main/README.md) to start Infinity. + +## Import the Benchmark datasets + +```sh +cd benchmark + +options: + -h, --help show this help message and exit + -d DATA_SET, --data DATA_SET + +python remote_benchmark_import.py -d sift_1m +python remote_benchmark_import.py -d gist_1m +``` + +## Run Benchmark + +```sh +options: + -h, --help show this help message and exit + -t THREADS, --threads THREADS + -r ROUNDS, --rounds ROUNDS + -d DATA_SET, --data DATA_SET + +# ROUNDS indicates the number of times Python executes the benchmark, and the result represents the average duration for each run. + +# Perform a benchmark on the SIFT1M dataset using a single thread, running it only once. +python remote_benchmark.py -t 1 -r 1 -d sift_1m +# Perform a benchmark on the GIST1M dataset using a single thread, running it only once. +python remote_benchmark.py -t 1 -r 1 -d gist_1m +``` +## A SIFT benchmark report + +- **Hardware**: Intel i5-12500H, 16C, 16GB +- **Operating system**: Ubuntu 22.04 +- **Dataset**: SIFT1M; **topk**: 100; **recall**: 97%+ +- **QPS**: 10,305 +- **P99 Latency**: 0.3 ms \ No newline at end of file diff --git a/.history/docs/benchmark_20231226140905.md b/.history/docs/benchmark_20231226140905.md new file mode 100644 index 0000000000..11805eeac9 --- /dev/null +++ b/.history/docs/benchmark_20231226140905.md @@ -0,0 +1,86 @@ +# Benchmark + +Infinity provides a Python script for benchmarking the SIFT1M and GIST1M datasets. + +## Build Infinity + +You have two options for building Infinity. Choose the option that best fit your needs: + +- [Build Infinity using Docker](../README.md) +- [Build from source](./build_from_source.md) + +## Download the Benchmark datasets + +To obtain the benchmark files, you have the option to download them using the wget command. + +```sh +#download sift benchmark +wget ftp://ftp.irisa.fr/local/texmex/corpus/sift.tar.gz +#download gist benchmark +wget ftp://ftp.irisa.fr/local/texmex/corpus/gist.tar.gz + +``` + +Alternatively, you can manually download the benchmark files by visiting [http://corpus-texmex.irisa.fr/](http://corpus-texmex.irisa.fr/). + +```sh +# Unzip and move the SIFT1M benchmark file. +tar -zxvf sift.tar.gz +mv sift/sift_base.fvecs test/data/benchmark/sift_1m/sift_base.fvecs +mv sift/sift_query.fvecs test/data/benchmark/sift_1m/sift_query.fvecs +mv sift/sift_groundtruth.ivecs test/data/benchmark/sift_1m/sift_groundtruth.ivecs + +# Unzip and move the GIST1M benchmark file. +tar -zxvf gist.tar.gz +mv gist/gist_base.fvecs test/data/benchmark/gist_1m/gist_base.fvecs +mv gist/gist_query.fvecs test/data/benchmark/gist_1m/gist_query.fvecs +mv gist/gist_groundtruth.ivecs test/data/benchmark/gist_1m/gist_groundtruth.ivecs + +``` + +## Benchmark dependencies + +```sh +cd python + +pip install -r requirements.txt +python setup.py bdist_wheel +pip install dist/infinity_sdk-0.1.0.dev1-py3-none-any.whl +``` + +## Import the Benchmark datasets + +```sh +cd benchmark + +options: + -h, --help show this help message and exit + -d DATA_SET, --data DATA_SET + +python remote_benchmark_import.py -d sift_1m +python remote_benchmark_import.py -d gist_1m +``` + +## Run Benchmark + +```sh +options: + -h, --help show this help message and exit + -t THREADS, --threads THREADS + -r ROUNDS, --rounds ROUNDS + -d DATA_SET, --data DATA_SET + +# ROUNDS indicates the number of times Python executes the benchmark, and the result represents the average duration for each run. + +# Perform a benchmark on the SIFT1M dataset using a single thread, running it only once. +python remote_benchmark.py -t 1 -r 1 -d sift_1m +# Perform a benchmark on the GIST1M dataset using a single thread, running it only once. +python remote_benchmark.py -t 1 -r 1 -d gist_1m +``` +## A SIFT benchmark report + +- **Hardware**: Intel i5-12500H, 16C, 16GB +- **Operating system**: Ubuntu 22.04 +- **Dataset**: SIFT1M; **topk**: 100; **recall**: 97%+ +- **QPS**: 10,305 +- **P99 Latency**: 0.3 ms \ No newline at end of file diff --git a/.history/docs/benchmark_20231226140937.md b/.history/docs/benchmark_20231226140937.md new file mode 100644 index 0000000000..b6d8b2f98e --- /dev/null +++ b/.history/docs/benchmark_20231226140937.md @@ -0,0 +1,86 @@ +# Benchmark + +Infinity provides a Python script for benchmarking the SIFT1M and GIST1M datasets. + +## Build Infinity + +You have two options for building Infinity. Choose the option that best fit your needs: + +- [Build Infinity using Docker](../README.md) +- [Build from source](./build_from_source.md) + +## Download the Benchmark datasets + +To obtain the benchmark files, you have the option to download them using the wget command. + +```sh +#download sift benchmark +wget ftp://ftp.irisa.fr/local/texmex/corpus/sift.tar.gz +#download gist benchmark +wget ftp://ftp.irisa.fr/local/texmex/corpus/gist.tar.gz + +``` + +Alternatively, you can manually download the benchmark files by visiting [http://corpus-texmex.irisa.fr/](http://corpus-texmex.irisa.fr/). + +```sh +# Unzip and move the SIFT1M benchmark file. +tar -zxvf sift.tar.gz +mv sift/sift_base.fvecs test/data/benchmark/sift_1m/sift_base.fvecs +mv sift/sift_query.fvecs test/data/benchmark/sift_1m/sift_query.fvecs +mv sift/sift_groundtruth.ivecs test/data/benchmark/sift_1m/sift_groundtruth.ivecs + +# Unzip and move the GIST1M benchmark file. +tar -zxvf gist.tar.gz +mv gist/gist_base.fvecs test/data/benchmark/gist_1m/gist_base.fvecs +mv gist/gist_query.fvecs test/data/benchmark/gist_1m/gist_query.fvecs +mv gist/gist_groundtruth.ivecs test/data/benchmark/gist_1m/gist_groundtruth.ivecs + +``` + +## Benchmark dependencies + +```sh +cd python + +pip install -r requirements.txt +python setup.py bdist_wheel +pip install dist/infinity_sdk-0.1.0.dev1-py3-none-any.whl +``` + +## Import the Benchmark datasets + +```sh +cd benchmark + +# options: +# -h, --help show this help message and exit +# -d DATA_SET, --data DATA_SET + +python remote_benchmark_import.py -d sift_1m +python remote_benchmark_import.py -d gist_1m +``` + +## Run Benchmark + +```sh +# options: +# -h, --help show this help message and exit +# -t THREADS, --threads THREADS +# -r ROUNDS, --rounds ROUNDS +# -d DATA_SET, --data DATA_SET + +# ROUNDS indicates the number of times Python executes the benchmark, and the result represents the average duration for each run. + +# Perform a benchmark on the SIFT1M dataset using a single thread, running it only once. +python remote_benchmark.py -t 1 -r 1 -d sift_1m +# Perform a benchmark on the GIST1M dataset using a single thread, running it only once. +python remote_benchmark.py -t 1 -r 1 -d gist_1m +``` +## A SIFT benchmark report + +- **Hardware**: Intel i5-12500H, 16C, 16GB +- **Operating system**: Ubuntu 22.04 +- **Dataset**: SIFT1M; **topk**: 100; **recall**: 97%+ +- **QPS**: 10,305 +- **P99 Latency**: 0.3 ms \ No newline at end of file diff --git a/.history/docs/benchmark_20231226141001.md b/.history/docs/benchmark_20231226141001.md new file mode 100644 index 0000000000..0e223fbbdd --- /dev/null +++ b/.history/docs/benchmark_20231226141001.md @@ -0,0 +1,86 @@ +# Benchmark + +Infinity provides a Python script for benchmarking the SIFT1M and GIST1M datasets. + +## Build Infinity + +You have two options for building Infinity. Choose the option that best fits your needs: + +- [Build Infinity using Docker](../README.md) +- [Build from source](./build_from_source.md) + +## Download the Benchmark datasets + +To obtain the benchmark files, you have the option to download them using the wget command. + +```sh +#download sift benchmark +wget ftp://ftp.irisa.fr/local/texmex/corpus/sift.tar.gz +#download gist benchmark +wget ftp://ftp.irisa.fr/local/texmex/corpus/gist.tar.gz + +``` + +Alternatively, you can manually download the benchmark files by visiting [http://corpus-texmex.irisa.fr/](http://corpus-texmex.irisa.fr/). + +```sh +# Unzip and move the SIFT1M benchmark file. +tar -zxvf sift.tar.gz +mv sift/sift_base.fvecs test/data/benchmark/sift_1m/sift_base.fvecs +mv sift/sift_query.fvecs test/data/benchmark/sift_1m/sift_query.fvecs +mv sift/sift_groundtruth.ivecs test/data/benchmark/sift_1m/sift_groundtruth.ivecs + +# Unzip and move the GIST1M benchmark file. +tar -zxvf gist.tar.gz +mv gist/gist_base.fvecs test/data/benchmark/gist_1m/gist_base.fvecs +mv gist/gist_query.fvecs test/data/benchmark/gist_1m/gist_query.fvecs +mv gist/gist_groundtruth.ivecs test/data/benchmark/gist_1m/gist_groundtruth.ivecs + +``` + +## Benchmark dependencies + +```sh +cd python + +pip install -r requirements.txt +python setup.py bdist_wheel +pip install dist/infinity_sdk-0.1.0.dev1-py3-none-any.whl +``` + +## Import the Benchmark datasets + +```sh +cd benchmark + +# options: +# -h, --help show this help message and exit +# -d DATA_SET, --data DATA_SET + +python remote_benchmark_import.py -d sift_1m +python remote_benchmark_import.py -d gist_1m +``` + +## Run Benchmark + +```sh +# options: +# -h, --help show this help message and exit +# -t THREADS, --threads THREADS +# -r ROUNDS, --rounds ROUNDS +# -d DATA_SET, --data DATA_SET + +# ROUNDS indicates the number of times Python executes the benchmark, and the result represents the average duration for each run. + +# Perform a benchmark on the SIFT1M dataset using a single thread, running it only once. +python remote_benchmark.py -t 1 -r 1 -d sift_1m +# Perform a benchmark on the GIST1M dataset using a single thread, running it only once. +python remote_benchmark.py -t 1 -r 1 -d gist_1m +``` +## A SIFT benchmark report + +- **Hardware**: Intel i5-12500H, 16C, 16GB +- **Operating system**: Ubuntu 22.04 +- **Dataset**: SIFT1M; **topk**: 100; **recall**: 97%+ +- **QPS**: 10,305 +- **P99 Latency**: 0.3 ms \ No newline at end of file diff --git a/docs/benchmark.md b/docs/benchmark.md index a41223fcc3..0e223fbbdd 100644 --- a/docs/benchmark.md +++ b/docs/benchmark.md @@ -2,14 +2,14 @@ Infinity provides a Python script for benchmarking the SIFT1M and GIST1M datasets. -## Get the Infinity binary file +## Build Infinity -```sh -git clone https://github.com/infiniflow/infinity.git -cd infinity -``` +You have two options for building Infinity. Choose the option that best fits your needs: + +- [Build Infinity using Docker](../README.md) +- [Build from source](./build_from_source.md) -## Download the benchmark files +## Download the Benchmark datasets To obtain the benchmark files, you have the option to download them using the wget command. @@ -48,31 +48,27 @@ python setup.py bdist_wheel pip install dist/infinity_sdk-0.1.0.dev1-py3-none-any.whl ``` -## Start Infinity - -See the [README.md](https://github.com/infiniflow/infinity/blob/main/README.md) to start Infinity. - -## Import data +## Import the Benchmark datasets ```sh cd benchmark -options: - -h, --help show this help message and exit - -d DATA_SET, --data DATA_SET +# options: +# -h, --help show this help message and exit +# -d DATA_SET, --data DATA_SET python remote_benchmark_import.py -d sift_1m python remote_benchmark_import.py -d gist_1m ``` -## Run benchmark +## Run Benchmark ```sh -options: - -h, --help show this help message and exit - -t THREADS, --threads THREADS - -r ROUNDS, --rounds ROUNDS - -d DATA_SET, --data DATA_SET +# options: +# -h, --help show this help message and exit +# -t THREADS, --threads THREADS +# -r ROUNDS, --rounds ROUNDS +# -d DATA_SET, --data DATA_SET # ROUNDS indicates the number of times Python executes the benchmark, and the result represents the average duration for each run. From e91abdd44e25a5aef651e8add4ee69e56b60f26f Mon Sep 17 00:00:00 2001 From: Vissidarte-Herman Date: Tue, 26 Dec 2023 14:13:10 +0800 Subject: [PATCH 3/7] Removed history --- .history/docs/benchmark_20231226133852.md | 90 ---------------------- .history/docs/benchmark_20231226140016.md | 92 ----------------------- .history/docs/benchmark_20231226140905.md | 86 --------------------- .history/docs/benchmark_20231226140937.md | 86 --------------------- .history/docs/benchmark_20231226141001.md | 86 --------------------- 5 files changed, 440 deletions(-) delete mode 100644 .history/docs/benchmark_20231226133852.md delete mode 100644 .history/docs/benchmark_20231226140016.md delete mode 100644 .history/docs/benchmark_20231226140905.md delete mode 100644 .history/docs/benchmark_20231226140937.md delete mode 100644 .history/docs/benchmark_20231226141001.md diff --git a/.history/docs/benchmark_20231226133852.md b/.history/docs/benchmark_20231226133852.md deleted file mode 100644 index a41223fcc3..0000000000 --- a/.history/docs/benchmark_20231226133852.md +++ /dev/null @@ -1,90 +0,0 @@ -# Benchmark - -Infinity provides a Python script for benchmarking the SIFT1M and GIST1M datasets. - -## Get the Infinity binary file - -```sh -git clone https://github.com/infiniflow/infinity.git -cd infinity -``` - -## Download the benchmark files - -To obtain the benchmark files, you have the option to download them using the wget command. - -```sh -#download sift benchmark -wget ftp://ftp.irisa.fr/local/texmex/corpus/sift.tar.gz -#download gist benchmark -wget ftp://ftp.irisa.fr/local/texmex/corpus/gist.tar.gz - -``` - -Alternatively, you can manually download the benchmark files by visiting [http://corpus-texmex.irisa.fr/](http://corpus-texmex.irisa.fr/). - -```sh -# Unzip and move the SIFT1M benchmark file. -tar -zxvf sift.tar.gz -mv sift/sift_base.fvecs test/data/benchmark/sift_1m/sift_base.fvecs -mv sift/sift_query.fvecs test/data/benchmark/sift_1m/sift_query.fvecs -mv sift/sift_groundtruth.ivecs test/data/benchmark/sift_1m/sift_groundtruth.ivecs - -# Unzip and move the GIST1M benchmark file. -tar -zxvf gist.tar.gz -mv gist/gist_base.fvecs test/data/benchmark/gist_1m/gist_base.fvecs -mv gist/gist_query.fvecs test/data/benchmark/gist_1m/gist_query.fvecs -mv gist/gist_groundtruth.ivecs test/data/benchmark/gist_1m/gist_groundtruth.ivecs - -``` - -## Benchmark dependencies - -```sh -cd python - -pip install -r requirements.txt -python setup.py bdist_wheel -pip install dist/infinity_sdk-0.1.0.dev1-py3-none-any.whl -``` - -## Start Infinity - -See the [README.md](https://github.com/infiniflow/infinity/blob/main/README.md) to start Infinity. - -## Import data - -```sh -cd benchmark - -options: - -h, --help show this help message and exit - -d DATA_SET, --data DATA_SET - -python remote_benchmark_import.py -d sift_1m -python remote_benchmark_import.py -d gist_1m -``` - -## Run benchmark - -```sh -options: - -h, --help show this help message and exit - -t THREADS, --threads THREADS - -r ROUNDS, --rounds ROUNDS - -d DATA_SET, --data DATA_SET - -# ROUNDS indicates the number of times Python executes the benchmark, and the result represents the average duration for each run. - -# Perform a benchmark on the SIFT1M dataset using a single thread, running it only once. -python remote_benchmark.py -t 1 -r 1 -d sift_1m -# Perform a benchmark on the GIST1M dataset using a single thread, running it only once. -python remote_benchmark.py -t 1 -r 1 -d gist_1m -``` -## A SIFT benchmark report - -- **Hardware**: Intel i5-12500H, 16C, 16GB -- **Operating system**: Ubuntu 22.04 -- **Dataset**: SIFT1M; **topk**: 100; **recall**: 97%+ -- **QPS**: 10,305 -- **P99 Latency**: 0.3 ms \ No newline at end of file diff --git a/.history/docs/benchmark_20231226140016.md b/.history/docs/benchmark_20231226140016.md deleted file mode 100644 index 174dd1059b..0000000000 --- a/.history/docs/benchmark_20231226140016.md +++ /dev/null @@ -1,92 +0,0 @@ -# Benchmark - -Infinity provides a Python script for benchmarking the SIFT1M and GIST1M datasets. - -## Build Infinity - -You can either build Infinity - -```sh -git clone https://github.com/infiniflow/infinity.git -cd infinity -``` - -## Download the Benchmark datasets - -To obtain the benchmark files, you have the option to download them using the wget command. - -```sh -#download sift benchmark -wget ftp://ftp.irisa.fr/local/texmex/corpus/sift.tar.gz -#download gist benchmark -wget ftp://ftp.irisa.fr/local/texmex/corpus/gist.tar.gz - -``` - -Alternatively, you can manually download the benchmark files by visiting [http://corpus-texmex.irisa.fr/](http://corpus-texmex.irisa.fr/). - -```sh -# Unzip and move the SIFT1M benchmark file. -tar -zxvf sift.tar.gz -mv sift/sift_base.fvecs test/data/benchmark/sift_1m/sift_base.fvecs -mv sift/sift_query.fvecs test/data/benchmark/sift_1m/sift_query.fvecs -mv sift/sift_groundtruth.ivecs test/data/benchmark/sift_1m/sift_groundtruth.ivecs - -# Unzip and move the GIST1M benchmark file. -tar -zxvf gist.tar.gz -mv gist/gist_base.fvecs test/data/benchmark/gist_1m/gist_base.fvecs -mv gist/gist_query.fvecs test/data/benchmark/gist_1m/gist_query.fvecs -mv gist/gist_groundtruth.ivecs test/data/benchmark/gist_1m/gist_groundtruth.ivecs - -``` - -## Benchmark dependencies - -```sh -cd python - -pip install -r requirements.txt -python setup.py bdist_wheel -pip install dist/infinity_sdk-0.1.0.dev1-py3-none-any.whl -``` - -## Start Infinity - -See the [README.md](https://github.com/infiniflow/infinity/blob/main/README.md) to start Infinity. - -## Import the Benchmark datasets - -```sh -cd benchmark - -options: - -h, --help show this help message and exit - -d DATA_SET, --data DATA_SET - -python remote_benchmark_import.py -d sift_1m -python remote_benchmark_import.py -d gist_1m -``` - -## Run Benchmark - -```sh -options: - -h, --help show this help message and exit - -t THREADS, --threads THREADS - -r ROUNDS, --rounds ROUNDS - -d DATA_SET, --data DATA_SET - -# ROUNDS indicates the number of times Python executes the benchmark, and the result represents the average duration for each run. - -# Perform a benchmark on the SIFT1M dataset using a single thread, running it only once. -python remote_benchmark.py -t 1 -r 1 -d sift_1m -# Perform a benchmark on the GIST1M dataset using a single thread, running it only once. -python remote_benchmark.py -t 1 -r 1 -d gist_1m -``` -## A SIFT benchmark report - -- **Hardware**: Intel i5-12500H, 16C, 16GB -- **Operating system**: Ubuntu 22.04 -- **Dataset**: SIFT1M; **topk**: 100; **recall**: 97%+ -- **QPS**: 10,305 -- **P99 Latency**: 0.3 ms \ No newline at end of file diff --git a/.history/docs/benchmark_20231226140905.md b/.history/docs/benchmark_20231226140905.md deleted file mode 100644 index 11805eeac9..0000000000 --- a/.history/docs/benchmark_20231226140905.md +++ /dev/null @@ -1,86 +0,0 @@ -# Benchmark - -Infinity provides a Python script for benchmarking the SIFT1M and GIST1M datasets. - -## Build Infinity - -You have two options for building Infinity. Choose the option that best fit your needs: - -- [Build Infinity using Docker](../README.md) -- [Build from source](./build_from_source.md) - -## Download the Benchmark datasets - -To obtain the benchmark files, you have the option to download them using the wget command. - -```sh -#download sift benchmark -wget ftp://ftp.irisa.fr/local/texmex/corpus/sift.tar.gz -#download gist benchmark -wget ftp://ftp.irisa.fr/local/texmex/corpus/gist.tar.gz - -``` - -Alternatively, you can manually download the benchmark files by visiting [http://corpus-texmex.irisa.fr/](http://corpus-texmex.irisa.fr/). - -```sh -# Unzip and move the SIFT1M benchmark file. -tar -zxvf sift.tar.gz -mv sift/sift_base.fvecs test/data/benchmark/sift_1m/sift_base.fvecs -mv sift/sift_query.fvecs test/data/benchmark/sift_1m/sift_query.fvecs -mv sift/sift_groundtruth.ivecs test/data/benchmark/sift_1m/sift_groundtruth.ivecs - -# Unzip and move the GIST1M benchmark file. -tar -zxvf gist.tar.gz -mv gist/gist_base.fvecs test/data/benchmark/gist_1m/gist_base.fvecs -mv gist/gist_query.fvecs test/data/benchmark/gist_1m/gist_query.fvecs -mv gist/gist_groundtruth.ivecs test/data/benchmark/gist_1m/gist_groundtruth.ivecs - -``` - -## Benchmark dependencies - -```sh -cd python - -pip install -r requirements.txt -python setup.py bdist_wheel -pip install dist/infinity_sdk-0.1.0.dev1-py3-none-any.whl -``` - -## Import the Benchmark datasets - -```sh -cd benchmark - -options: - -h, --help show this help message and exit - -d DATA_SET, --data DATA_SET - -python remote_benchmark_import.py -d sift_1m -python remote_benchmark_import.py -d gist_1m -``` - -## Run Benchmark - -```sh -options: - -h, --help show this help message and exit - -t THREADS, --threads THREADS - -r ROUNDS, --rounds ROUNDS - -d DATA_SET, --data DATA_SET - -# ROUNDS indicates the number of times Python executes the benchmark, and the result represents the average duration for each run. - -# Perform a benchmark on the SIFT1M dataset using a single thread, running it only once. -python remote_benchmark.py -t 1 -r 1 -d sift_1m -# Perform a benchmark on the GIST1M dataset using a single thread, running it only once. -python remote_benchmark.py -t 1 -r 1 -d gist_1m -``` -## A SIFT benchmark report - -- **Hardware**: Intel i5-12500H, 16C, 16GB -- **Operating system**: Ubuntu 22.04 -- **Dataset**: SIFT1M; **topk**: 100; **recall**: 97%+ -- **QPS**: 10,305 -- **P99 Latency**: 0.3 ms \ No newline at end of file diff --git a/.history/docs/benchmark_20231226140937.md b/.history/docs/benchmark_20231226140937.md deleted file mode 100644 index b6d8b2f98e..0000000000 --- a/.history/docs/benchmark_20231226140937.md +++ /dev/null @@ -1,86 +0,0 @@ -# Benchmark - -Infinity provides a Python script for benchmarking the SIFT1M and GIST1M datasets. - -## Build Infinity - -You have two options for building Infinity. Choose the option that best fit your needs: - -- [Build Infinity using Docker](../README.md) -- [Build from source](./build_from_source.md) - -## Download the Benchmark datasets - -To obtain the benchmark files, you have the option to download them using the wget command. - -```sh -#download sift benchmark -wget ftp://ftp.irisa.fr/local/texmex/corpus/sift.tar.gz -#download gist benchmark -wget ftp://ftp.irisa.fr/local/texmex/corpus/gist.tar.gz - -``` - -Alternatively, you can manually download the benchmark files by visiting [http://corpus-texmex.irisa.fr/](http://corpus-texmex.irisa.fr/). - -```sh -# Unzip and move the SIFT1M benchmark file. -tar -zxvf sift.tar.gz -mv sift/sift_base.fvecs test/data/benchmark/sift_1m/sift_base.fvecs -mv sift/sift_query.fvecs test/data/benchmark/sift_1m/sift_query.fvecs -mv sift/sift_groundtruth.ivecs test/data/benchmark/sift_1m/sift_groundtruth.ivecs - -# Unzip and move the GIST1M benchmark file. -tar -zxvf gist.tar.gz -mv gist/gist_base.fvecs test/data/benchmark/gist_1m/gist_base.fvecs -mv gist/gist_query.fvecs test/data/benchmark/gist_1m/gist_query.fvecs -mv gist/gist_groundtruth.ivecs test/data/benchmark/gist_1m/gist_groundtruth.ivecs - -``` - -## Benchmark dependencies - -```sh -cd python - -pip install -r requirements.txt -python setup.py bdist_wheel -pip install dist/infinity_sdk-0.1.0.dev1-py3-none-any.whl -``` - -## Import the Benchmark datasets - -```sh -cd benchmark - -# options: -# -h, --help show this help message and exit -# -d DATA_SET, --data DATA_SET - -python remote_benchmark_import.py -d sift_1m -python remote_benchmark_import.py -d gist_1m -``` - -## Run Benchmark - -```sh -# options: -# -h, --help show this help message and exit -# -t THREADS, --threads THREADS -# -r ROUNDS, --rounds ROUNDS -# -d DATA_SET, --data DATA_SET - -# ROUNDS indicates the number of times Python executes the benchmark, and the result represents the average duration for each run. - -# Perform a benchmark on the SIFT1M dataset using a single thread, running it only once. -python remote_benchmark.py -t 1 -r 1 -d sift_1m -# Perform a benchmark on the GIST1M dataset using a single thread, running it only once. -python remote_benchmark.py -t 1 -r 1 -d gist_1m -``` -## A SIFT benchmark report - -- **Hardware**: Intel i5-12500H, 16C, 16GB -- **Operating system**: Ubuntu 22.04 -- **Dataset**: SIFT1M; **topk**: 100; **recall**: 97%+ -- **QPS**: 10,305 -- **P99 Latency**: 0.3 ms \ No newline at end of file diff --git a/.history/docs/benchmark_20231226141001.md b/.history/docs/benchmark_20231226141001.md deleted file mode 100644 index 0e223fbbdd..0000000000 --- a/.history/docs/benchmark_20231226141001.md +++ /dev/null @@ -1,86 +0,0 @@ -# Benchmark - -Infinity provides a Python script for benchmarking the SIFT1M and GIST1M datasets. - -## Build Infinity - -You have two options for building Infinity. Choose the option that best fits your needs: - -- [Build Infinity using Docker](../README.md) -- [Build from source](./build_from_source.md) - -## Download the Benchmark datasets - -To obtain the benchmark files, you have the option to download them using the wget command. - -```sh -#download sift benchmark -wget ftp://ftp.irisa.fr/local/texmex/corpus/sift.tar.gz -#download gist benchmark -wget ftp://ftp.irisa.fr/local/texmex/corpus/gist.tar.gz - -``` - -Alternatively, you can manually download the benchmark files by visiting [http://corpus-texmex.irisa.fr/](http://corpus-texmex.irisa.fr/). - -```sh -# Unzip and move the SIFT1M benchmark file. -tar -zxvf sift.tar.gz -mv sift/sift_base.fvecs test/data/benchmark/sift_1m/sift_base.fvecs -mv sift/sift_query.fvecs test/data/benchmark/sift_1m/sift_query.fvecs -mv sift/sift_groundtruth.ivecs test/data/benchmark/sift_1m/sift_groundtruth.ivecs - -# Unzip and move the GIST1M benchmark file. -tar -zxvf gist.tar.gz -mv gist/gist_base.fvecs test/data/benchmark/gist_1m/gist_base.fvecs -mv gist/gist_query.fvecs test/data/benchmark/gist_1m/gist_query.fvecs -mv gist/gist_groundtruth.ivecs test/data/benchmark/gist_1m/gist_groundtruth.ivecs - -``` - -## Benchmark dependencies - -```sh -cd python - -pip install -r requirements.txt -python setup.py bdist_wheel -pip install dist/infinity_sdk-0.1.0.dev1-py3-none-any.whl -``` - -## Import the Benchmark datasets - -```sh -cd benchmark - -# options: -# -h, --help show this help message and exit -# -d DATA_SET, --data DATA_SET - -python remote_benchmark_import.py -d sift_1m -python remote_benchmark_import.py -d gist_1m -``` - -## Run Benchmark - -```sh -# options: -# -h, --help show this help message and exit -# -t THREADS, --threads THREADS -# -r ROUNDS, --rounds ROUNDS -# -d DATA_SET, --data DATA_SET - -# ROUNDS indicates the number of times Python executes the benchmark, and the result represents the average duration for each run. - -# Perform a benchmark on the SIFT1M dataset using a single thread, running it only once. -python remote_benchmark.py -t 1 -r 1 -d sift_1m -# Perform a benchmark on the GIST1M dataset using a single thread, running it only once. -python remote_benchmark.py -t 1 -r 1 -d gist_1m -``` -## A SIFT benchmark report - -- **Hardware**: Intel i5-12500H, 16C, 16GB -- **Operating system**: Ubuntu 22.04 -- **Dataset**: SIFT1M; **topk**: 100; **recall**: 97%+ -- **QPS**: 10,305 -- **P99 Latency**: 0.3 ms \ No newline at end of file From 30ede4dbb470e8f2ec6d29f1804e5823434901d8 Mon Sep 17 00:00:00 2001 From: Vissidarte-Herman Date: Tue, 26 Dec 2023 14:22:26 +0800 Subject: [PATCH 4/7] minor updates --- .history/docs/benchmark_20231226141000.md | 86 +++++++++++++++++++++++ .history/docs/benchmark_20231226141737.md | 86 +++++++++++++++++++++++ .history/docs/benchmark_20231226142212.md | 86 +++++++++++++++++++++++ docs/benchmark.md | 4 +- 4 files changed, 260 insertions(+), 2 deletions(-) create mode 100644 .history/docs/benchmark_20231226141000.md create mode 100644 .history/docs/benchmark_20231226141737.md create mode 100644 .history/docs/benchmark_20231226142212.md diff --git a/.history/docs/benchmark_20231226141000.md b/.history/docs/benchmark_20231226141000.md new file mode 100644 index 0000000000..0e223fbbdd --- /dev/null +++ b/.history/docs/benchmark_20231226141000.md @@ -0,0 +1,86 @@ +# Benchmark + +Infinity provides a Python script for benchmarking the SIFT1M and GIST1M datasets. + +## Build Infinity + +You have two options for building Infinity. Choose the option that best fits your needs: + +- [Build Infinity using Docker](../README.md) +- [Build from source](./build_from_source.md) + +## Download the Benchmark datasets + +To obtain the benchmark files, you have the option to download them using the wget command. + +```sh +#download sift benchmark +wget ftp://ftp.irisa.fr/local/texmex/corpus/sift.tar.gz +#download gist benchmark +wget ftp://ftp.irisa.fr/local/texmex/corpus/gist.tar.gz + +``` + +Alternatively, you can manually download the benchmark files by visiting [http://corpus-texmex.irisa.fr/](http://corpus-texmex.irisa.fr/). + +```sh +# Unzip and move the SIFT1M benchmark file. +tar -zxvf sift.tar.gz +mv sift/sift_base.fvecs test/data/benchmark/sift_1m/sift_base.fvecs +mv sift/sift_query.fvecs test/data/benchmark/sift_1m/sift_query.fvecs +mv sift/sift_groundtruth.ivecs test/data/benchmark/sift_1m/sift_groundtruth.ivecs + +# Unzip and move the GIST1M benchmark file. +tar -zxvf gist.tar.gz +mv gist/gist_base.fvecs test/data/benchmark/gist_1m/gist_base.fvecs +mv gist/gist_query.fvecs test/data/benchmark/gist_1m/gist_query.fvecs +mv gist/gist_groundtruth.ivecs test/data/benchmark/gist_1m/gist_groundtruth.ivecs + +``` + +## Benchmark dependencies + +```sh +cd python + +pip install -r requirements.txt +python setup.py bdist_wheel +pip install dist/infinity_sdk-0.1.0.dev1-py3-none-any.whl +``` + +## Import the Benchmark datasets + +```sh +cd benchmark + +# options: +# -h, --help show this help message and exit +# -d DATA_SET, --data DATA_SET + +python remote_benchmark_import.py -d sift_1m +python remote_benchmark_import.py -d gist_1m +``` + +## Run Benchmark + +```sh +# options: +# -h, --help show this help message and exit +# -t THREADS, --threads THREADS +# -r ROUNDS, --rounds ROUNDS +# -d DATA_SET, --data DATA_SET + +# ROUNDS indicates the number of times Python executes the benchmark, and the result represents the average duration for each run. + +# Perform a benchmark on the SIFT1M dataset using a single thread, running it only once. +python remote_benchmark.py -t 1 -r 1 -d sift_1m +# Perform a benchmark on the GIST1M dataset using a single thread, running it only once. +python remote_benchmark.py -t 1 -r 1 -d gist_1m +``` +## A SIFT benchmark report + +- **Hardware**: Intel i5-12500H, 16C, 16GB +- **Operating system**: Ubuntu 22.04 +- **Dataset**: SIFT1M; **topk**: 100; **recall**: 97%+ +- **QPS**: 10,305 +- **P99 Latency**: 0.3 ms \ No newline at end of file diff --git a/.history/docs/benchmark_20231226141737.md b/.history/docs/benchmark_20231226141737.md new file mode 100644 index 0000000000..5ec5178e0e --- /dev/null +++ b/.history/docs/benchmark_20231226141737.md @@ -0,0 +1,86 @@ +# Benchmark + +Infinity provides a Python script for benchmarking the SIFT1M and GIST1M datasets. + +## Build and start Infinity + +You have two options for building Infinity. Choose the option that best fits your needs: + +- [Build Infinity using Docker](../README.md) +- [Build from source](./build_from_source.md) + +## Download the Benchmark datasets + +To obtain the benchmark files, you have the option to download them using the wget command. + +```sh +#download sift benchmark +wget ftp://ftp.irisa.fr/local/texmex/corpus/sift.tar.gz +#download gist benchmark +wget ftp://ftp.irisa.fr/local/texmex/corpus/gist.tar.gz + +``` + +Alternatively, you can manually download the benchmark files by visiting [http://corpus-texmex.irisa.fr/](http://corpus-texmex.irisa.fr/). + +```sh +# Unzip and move the SIFT1M benchmark file. +tar -zxvf sift.tar.gz +mv sift/sift_base.fvecs test/data/benchmark/sift_1m/sift_base.fvecs +mv sift/sift_query.fvecs test/data/benchmark/sift_1m/sift_query.fvecs +mv sift/sift_groundtruth.ivecs test/data/benchmark/sift_1m/sift_groundtruth.ivecs + +# Unzip and move the GIST1M benchmark file. +tar -zxvf gist.tar.gz +mv gist/gist_base.fvecs test/data/benchmark/gist_1m/gist_base.fvecs +mv gist/gist_query.fvecs test/data/benchmark/gist_1m/gist_query.fvecs +mv gist/gist_groundtruth.ivecs test/data/benchmark/gist_1m/gist_groundtruth.ivecs + +``` + +## Benchmark dependencies + +```sh +cd python + +pip install -r requirements.txt +python setup.py bdist_wheel +pip install dist/infinity_sdk-0.1.0.dev1-py3-none-any.whl +``` + +## Import the Benchmark datasets + +```sh +cd benchmark + +# options: +# -h, --help show this help message and exit +# -d DATA_SET, --data DATA_SET + +python remote_benchmark_import.py -d sift_1m +python remote_benchmark_import.py -d gist_1m +``` + +## Run Benchmark + +```sh +# options: +# -h, --help show this help message and exit +# -t THREADS, --threads THREADS +# -r ROUNDS, --rounds ROUNDS +# -d DATA_SET, --data DATA_SET + +# ROUNDS indicates the number of times Python executes the benchmark, and the result represents the average duration for each run. + +# Perform a benchmark on the SIFT1M dataset using a single thread, running it only once. +python remote_benchmark.py -t 1 -r 1 -d sift_1m +# Perform a benchmark on the GIST1M dataset using a single thread, running it only once. +python remote_benchmark.py -t 1 -r 1 -d gist_1m +``` +## A SIFT benchmark report + +- **Hardware**: Intel i5-12500H, 16C, 16GB +- **Operating system**: Ubuntu 22.04 +- **Dataset**: SIFT1M; **topk**: 100; **recall**: 97%+ +- **QPS**: 10,305 +- **P99 Latency**: 0.3 ms \ No newline at end of file diff --git a/.history/docs/benchmark_20231226142212.md b/.history/docs/benchmark_20231226142212.md new file mode 100644 index 0000000000..8e20691e8b --- /dev/null +++ b/.history/docs/benchmark_20231226142212.md @@ -0,0 +1,86 @@ +# Benchmark + +Infinity provides a Python script for benchmarking the SIFT1M and GIST1M datasets. + +## Build and start Infinity + +You have two options for building Infinity. Choose the option that best fits your needs: + +- [Build Infinity using Docker](../README.md) +- [Build from source](./build_from_source.md) + +## Download the Benchmark datasets + +To obtain the benchmark files, you have the option to download them using the wget command. + +```sh +#download sift benchmark +wget ftp://ftp.irisa.fr/local/texmex/corpus/sift.tar.gz +#download gist benchmark +wget ftp://ftp.irisa.fr/local/texmex/corpus/gist.tar.gz + +``` + +Alternatively, you can manually download the benchmark files by visiting [http://corpus-texmex.irisa.fr/](http://corpus-texmex.irisa.fr/). + +```sh +# Unzip and move the SIFT1M benchmark file. +tar -zxvf sift.tar.gz +mv sift/sift_base.fvecs test/data/benchmark/sift_1m/sift_base.fvecs +mv sift/sift_query.fvecs test/data/benchmark/sift_1m/sift_query.fvecs +mv sift/sift_groundtruth.ivecs test/data/benchmark/sift_1m/sift_groundtruth.ivecs + +# Unzip and move the GIST1M benchmark file. +tar -zxvf gist.tar.gz +mv gist/gist_base.fvecs test/data/benchmark/gist_1m/gist_base.fvecs +mv gist/gist_query.fvecs test/data/benchmark/gist_1m/gist_query.fvecs +mv gist/gist_groundtruth.ivecs test/data/benchmark/gist_1m/gist_groundtruth.ivecs + +``` + +## Benchmark dependencies + +```sh +cd python + +pip install -r requirements.txt +python setup.py bdist_wheel +pip install dist/infinity_sdk-0.1.0.dev1-py3-none-any.whl +``` + +## Import the Benchmark datasets + +```sh +cd benchmark + +# options: +# -h, --help show this help message and exit +# -d DATA_SET, --data DATA_SET + +python remote_benchmark_import.py -d sift_1m +python remote_benchmark_import.py -d gist_1m +``` + +## Run Benchmark + +```sh +# options: +# -h, --help show this help message and exit +# -t THREADS, --threads THREADS +# -r ROUNDS, --rounds ROUNDS +# -d DATA_SET, --data DATA_SET + +# ROUNDS indicates the number of times Python executes the benchmark, and the result represents the average duration for each run. + +# Perform a benchmark on the SIFT1M dataset using a single thread, running it only once. +python remote_benchmark.py -t 1 -r 1 -d sift_1m +# Perform a benchmark on the GIST1M dataset using a single thread, running it only once. +python remote_benchmark.py -t 1 -r 1 -d gist_1m +``` +## A SIFT1M Benchmark report + +- **Hardware**: Intel i5-12500H, 16C, 16GB +- **Operating system**: Ubuntu 22.04 +- **Dataset**: SIFT1M; **topk**: 100; **recall**: 97%+ +- **QPS**: 10,305 +- **P99 Latency**: 0.3 ms \ No newline at end of file diff --git a/docs/benchmark.md b/docs/benchmark.md index 0e223fbbdd..8e20691e8b 100644 --- a/docs/benchmark.md +++ b/docs/benchmark.md @@ -2,7 +2,7 @@ Infinity provides a Python script for benchmarking the SIFT1M and GIST1M datasets. -## Build Infinity +## Build and start Infinity You have two options for building Infinity. Choose the option that best fits your needs: @@ -77,7 +77,7 @@ python remote_benchmark.py -t 1 -r 1 -d sift_1m # Perform a benchmark on the GIST1M dataset using a single thread, running it only once. python remote_benchmark.py -t 1 -r 1 -d gist_1m ``` -## A SIFT benchmark report +## A SIFT1M Benchmark report - **Hardware**: Intel i5-12500H, 16C, 16GB - **Operating system**: Ubuntu 22.04 From 30938a38d7655e3d4699992a7be9c562c56899ab Mon Sep 17 00:00:00 2001 From: Vissidarte-Herman Date: Tue, 26 Dec 2023 14:23:35 +0800 Subject: [PATCH 5/7] minor updates --- .history/docs/benchmark_20231226141000.md | 86 ----------------------- .history/docs/benchmark_20231226141737.md | 86 ----------------------- .history/docs/benchmark_20231226142212.md | 86 ----------------------- 3 files changed, 258 deletions(-) delete mode 100644 .history/docs/benchmark_20231226141000.md delete mode 100644 .history/docs/benchmark_20231226141737.md delete mode 100644 .history/docs/benchmark_20231226142212.md diff --git a/.history/docs/benchmark_20231226141000.md b/.history/docs/benchmark_20231226141000.md deleted file mode 100644 index 0e223fbbdd..0000000000 --- a/.history/docs/benchmark_20231226141000.md +++ /dev/null @@ -1,86 +0,0 @@ -# Benchmark - -Infinity provides a Python script for benchmarking the SIFT1M and GIST1M datasets. - -## Build Infinity - -You have two options for building Infinity. Choose the option that best fits your needs: - -- [Build Infinity using Docker](../README.md) -- [Build from source](./build_from_source.md) - -## Download the Benchmark datasets - -To obtain the benchmark files, you have the option to download them using the wget command. - -```sh -#download sift benchmark -wget ftp://ftp.irisa.fr/local/texmex/corpus/sift.tar.gz -#download gist benchmark -wget ftp://ftp.irisa.fr/local/texmex/corpus/gist.tar.gz - -``` - -Alternatively, you can manually download the benchmark files by visiting [http://corpus-texmex.irisa.fr/](http://corpus-texmex.irisa.fr/). - -```sh -# Unzip and move the SIFT1M benchmark file. -tar -zxvf sift.tar.gz -mv sift/sift_base.fvecs test/data/benchmark/sift_1m/sift_base.fvecs -mv sift/sift_query.fvecs test/data/benchmark/sift_1m/sift_query.fvecs -mv sift/sift_groundtruth.ivecs test/data/benchmark/sift_1m/sift_groundtruth.ivecs - -# Unzip and move the GIST1M benchmark file. -tar -zxvf gist.tar.gz -mv gist/gist_base.fvecs test/data/benchmark/gist_1m/gist_base.fvecs -mv gist/gist_query.fvecs test/data/benchmark/gist_1m/gist_query.fvecs -mv gist/gist_groundtruth.ivecs test/data/benchmark/gist_1m/gist_groundtruth.ivecs - -``` - -## Benchmark dependencies - -```sh -cd python - -pip install -r requirements.txt -python setup.py bdist_wheel -pip install dist/infinity_sdk-0.1.0.dev1-py3-none-any.whl -``` - -## Import the Benchmark datasets - -```sh -cd benchmark - -# options: -# -h, --help show this help message and exit -# -d DATA_SET, --data DATA_SET - -python remote_benchmark_import.py -d sift_1m -python remote_benchmark_import.py -d gist_1m -``` - -## Run Benchmark - -```sh -# options: -# -h, --help show this help message and exit -# -t THREADS, --threads THREADS -# -r ROUNDS, --rounds ROUNDS -# -d DATA_SET, --data DATA_SET - -# ROUNDS indicates the number of times Python executes the benchmark, and the result represents the average duration for each run. - -# Perform a benchmark on the SIFT1M dataset using a single thread, running it only once. -python remote_benchmark.py -t 1 -r 1 -d sift_1m -# Perform a benchmark on the GIST1M dataset using a single thread, running it only once. -python remote_benchmark.py -t 1 -r 1 -d gist_1m -``` -## A SIFT benchmark report - -- **Hardware**: Intel i5-12500H, 16C, 16GB -- **Operating system**: Ubuntu 22.04 -- **Dataset**: SIFT1M; **topk**: 100; **recall**: 97%+ -- **QPS**: 10,305 -- **P99 Latency**: 0.3 ms \ No newline at end of file diff --git a/.history/docs/benchmark_20231226141737.md b/.history/docs/benchmark_20231226141737.md deleted file mode 100644 index 5ec5178e0e..0000000000 --- a/.history/docs/benchmark_20231226141737.md +++ /dev/null @@ -1,86 +0,0 @@ -# Benchmark - -Infinity provides a Python script for benchmarking the SIFT1M and GIST1M datasets. - -## Build and start Infinity - -You have two options for building Infinity. Choose the option that best fits your needs: - -- [Build Infinity using Docker](../README.md) -- [Build from source](./build_from_source.md) - -## Download the Benchmark datasets - -To obtain the benchmark files, you have the option to download them using the wget command. - -```sh -#download sift benchmark -wget ftp://ftp.irisa.fr/local/texmex/corpus/sift.tar.gz -#download gist benchmark -wget ftp://ftp.irisa.fr/local/texmex/corpus/gist.tar.gz - -``` - -Alternatively, you can manually download the benchmark files by visiting [http://corpus-texmex.irisa.fr/](http://corpus-texmex.irisa.fr/). - -```sh -# Unzip and move the SIFT1M benchmark file. -tar -zxvf sift.tar.gz -mv sift/sift_base.fvecs test/data/benchmark/sift_1m/sift_base.fvecs -mv sift/sift_query.fvecs test/data/benchmark/sift_1m/sift_query.fvecs -mv sift/sift_groundtruth.ivecs test/data/benchmark/sift_1m/sift_groundtruth.ivecs - -# Unzip and move the GIST1M benchmark file. -tar -zxvf gist.tar.gz -mv gist/gist_base.fvecs test/data/benchmark/gist_1m/gist_base.fvecs -mv gist/gist_query.fvecs test/data/benchmark/gist_1m/gist_query.fvecs -mv gist/gist_groundtruth.ivecs test/data/benchmark/gist_1m/gist_groundtruth.ivecs - -``` - -## Benchmark dependencies - -```sh -cd python - -pip install -r requirements.txt -python setup.py bdist_wheel -pip install dist/infinity_sdk-0.1.0.dev1-py3-none-any.whl -``` - -## Import the Benchmark datasets - -```sh -cd benchmark - -# options: -# -h, --help show this help message and exit -# -d DATA_SET, --data DATA_SET - -python remote_benchmark_import.py -d sift_1m -python remote_benchmark_import.py -d gist_1m -``` - -## Run Benchmark - -```sh -# options: -# -h, --help show this help message and exit -# -t THREADS, --threads THREADS -# -r ROUNDS, --rounds ROUNDS -# -d DATA_SET, --data DATA_SET - -# ROUNDS indicates the number of times Python executes the benchmark, and the result represents the average duration for each run. - -# Perform a benchmark on the SIFT1M dataset using a single thread, running it only once. -python remote_benchmark.py -t 1 -r 1 -d sift_1m -# Perform a benchmark on the GIST1M dataset using a single thread, running it only once. -python remote_benchmark.py -t 1 -r 1 -d gist_1m -``` -## A SIFT benchmark report - -- **Hardware**: Intel i5-12500H, 16C, 16GB -- **Operating system**: Ubuntu 22.04 -- **Dataset**: SIFT1M; **topk**: 100; **recall**: 97%+ -- **QPS**: 10,305 -- **P99 Latency**: 0.3 ms \ No newline at end of file diff --git a/.history/docs/benchmark_20231226142212.md b/.history/docs/benchmark_20231226142212.md deleted file mode 100644 index 8e20691e8b..0000000000 --- a/.history/docs/benchmark_20231226142212.md +++ /dev/null @@ -1,86 +0,0 @@ -# Benchmark - -Infinity provides a Python script for benchmarking the SIFT1M and GIST1M datasets. - -## Build and start Infinity - -You have two options for building Infinity. Choose the option that best fits your needs: - -- [Build Infinity using Docker](../README.md) -- [Build from source](./build_from_source.md) - -## Download the Benchmark datasets - -To obtain the benchmark files, you have the option to download them using the wget command. - -```sh -#download sift benchmark -wget ftp://ftp.irisa.fr/local/texmex/corpus/sift.tar.gz -#download gist benchmark -wget ftp://ftp.irisa.fr/local/texmex/corpus/gist.tar.gz - -``` - -Alternatively, you can manually download the benchmark files by visiting [http://corpus-texmex.irisa.fr/](http://corpus-texmex.irisa.fr/). - -```sh -# Unzip and move the SIFT1M benchmark file. -tar -zxvf sift.tar.gz -mv sift/sift_base.fvecs test/data/benchmark/sift_1m/sift_base.fvecs -mv sift/sift_query.fvecs test/data/benchmark/sift_1m/sift_query.fvecs -mv sift/sift_groundtruth.ivecs test/data/benchmark/sift_1m/sift_groundtruth.ivecs - -# Unzip and move the GIST1M benchmark file. -tar -zxvf gist.tar.gz -mv gist/gist_base.fvecs test/data/benchmark/gist_1m/gist_base.fvecs -mv gist/gist_query.fvecs test/data/benchmark/gist_1m/gist_query.fvecs -mv gist/gist_groundtruth.ivecs test/data/benchmark/gist_1m/gist_groundtruth.ivecs - -``` - -## Benchmark dependencies - -```sh -cd python - -pip install -r requirements.txt -python setup.py bdist_wheel -pip install dist/infinity_sdk-0.1.0.dev1-py3-none-any.whl -``` - -## Import the Benchmark datasets - -```sh -cd benchmark - -# options: -# -h, --help show this help message and exit -# -d DATA_SET, --data DATA_SET - -python remote_benchmark_import.py -d sift_1m -python remote_benchmark_import.py -d gist_1m -``` - -## Run Benchmark - -```sh -# options: -# -h, --help show this help message and exit -# -t THREADS, --threads THREADS -# -r ROUNDS, --rounds ROUNDS -# -d DATA_SET, --data DATA_SET - -# ROUNDS indicates the number of times Python executes the benchmark, and the result represents the average duration for each run. - -# Perform a benchmark on the SIFT1M dataset using a single thread, running it only once. -python remote_benchmark.py -t 1 -r 1 -d sift_1m -# Perform a benchmark on the GIST1M dataset using a single thread, running it only once. -python remote_benchmark.py -t 1 -r 1 -d gist_1m -``` -## A SIFT1M Benchmark report - -- **Hardware**: Intel i5-12500H, 16C, 16GB -- **Operating system**: Ubuntu 22.04 -- **Dataset**: SIFT1M; **topk**: 100; **recall**: 97%+ -- **QPS**: 10,305 -- **P99 Latency**: 0.3 ms \ No newline at end of file From e1905b8e2f53d94af699107d8c332b1fde31c5ca Mon Sep 17 00:00:00 2001 From: Vissidarte-Herman Date: Tue, 26 Dec 2023 14:29:10 +0800 Subject: [PATCH 6/7] Editorial updates. --- docs/benchmark.md | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) diff --git a/docs/benchmark.md b/docs/benchmark.md index 8e20691e8b..2ee2bab767 100644 --- a/docs/benchmark.md +++ b/docs/benchmark.md @@ -11,7 +11,7 @@ You have two options for building Infinity. Choose the option that best fits you ## Download the Benchmark datasets -To obtain the benchmark files, you have the option to download them using the wget command. +To obtain the benchmark datasets, you have the option to download them using the wget command. ```sh #download sift benchmark @@ -21,7 +21,7 @@ wget ftp://ftp.irisa.fr/local/texmex/corpus/gist.tar.gz ``` -Alternatively, you can manually download the benchmark files by visiting [http://corpus-texmex.irisa.fr/](http://corpus-texmex.irisa.fr/). +Alternatively, you can manually download the benchmark datasets by visiting [http://corpus-texmex.irisa.fr/](http://corpus-texmex.irisa.fr/). ```sh # Unzip and move the SIFT1M benchmark file. From d4b81a11a22dec03b2a44bb64e85bc9e25caa441 Mon Sep 17 00:00:00 2001 From: Vissidarte-Herman Date: Tue, 26 Dec 2023 19:29:26 +0800 Subject: [PATCH 7/7] Updated benchmark report --- docs/benchmark.md | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/docs/benchmark.md b/docs/benchmark.md index 2ee2bab767..42aa1e180e 100644 --- a/docs/benchmark.md +++ b/docs/benchmark.md @@ -83,4 +83,4 @@ python remote_benchmark.py -t 1 -r 1 -d gist_1m - **Operating system**: Ubuntu 22.04 - **Dataset**: SIFT1M; **topk**: 100; **recall**: 97%+ - **QPS**: 10,305 -- **P99 Latency**: 0.3 ms \ No newline at end of file +- **P99 Latency**: < 0.4 ms \ No newline at end of file