facebookresearch/faiss
1
0
Fork 0
mirror of https://github.com/facebookresearch/faiss.git synced 2026-10-11 22:50:00 +00:00
Code Issues Packages Projects Releases Wiki Activity

Select OpenBLAS OMP build in conda for AArch64 + Linux (#5669)

Browse Source 
Summary:
- The current Faiss conda recipe does not specify the OpenBLAS threading variant.
- On Linux AArch64, the dependency solver selects the pthreads build.
- This causes problems when Faiss calls BLAS from within an existing OpenMP parallel region due to nested threading
- As a result, the benchmark below (with an OOB conda install of faiss) only yields 7% of the throughput reachable when an OMP build of OpenBLAS is selected (what this PR does)
- This PR fixes the issue, and as a result accelerates affected workloads by ~ 14x
```
# SPDX-FileCopyrightText: Copyright 2026 Arm Limited and/or its affiliates <[email protected]>
# SPDX-License-Identifier: MIT

import ctypes
import statistics
import time

import faiss
import numpy as np

rng = np.random.default_rng(1234)
d = 128
nlist = 4096
nq = 40_000

quantizer = faiss.IndexFlatL2(d)
quantizer.add(rng.random((nlist, d), dtype=np.float32))
index = faiss.IndexIVFPQFastScan(quantizer, d, nlist, 128, 4)
# index is empty - this only triggeres the course quantizer
index.is_trained = True
index.nprobe = 64
xq = rng.random((nq, d), dtype=np.float32)

openblas = ctypes.CDLL("libopenblas.so.0")
openblas.openblas_get_parallel.restype = ctypes.c_int
openblas.openblas_get_num_threads.restype = ctypes.c_int

print(f"nq={nq:,}, d={d}, nlist={nlist}, nprobe={index.nprobe}")
print(f"OpenMP threads={faiss.omp_get_max_threads()}")
print(
    f"OpenBLAS parallel={openblas.openblas_get_parallel()}, "
    f"threads={openblas.openblas_get_num_threads()}"
)
print(f"BLAS threshold={faiss.cvar.distance_compute_blas_threshold}")

index.search(xq, 10)
times = []
for _ in range(20):
    start = time.perf_counter()
    index.search(xq, 10)
    times.append(time.perf_counter() - start)

elapsed = statistics.median(times)
print(f"median={elapsed:.6f}s, throughput={nq / elapsed:,.0f} QPS")
```

Pull Request resolved: https://github.com/facebookresearch/faiss/pull/5669

Reviewed By: weixianghong

Differential Revision: D124380509

Pulled By: mnorris11

fbshipit-source-id: 16e1521e7aaf7d65c6a4ce8e404a47bdb26c1198
This commit is contained in:
Fadi Arafeh
2026-10-09 11:27:04 -07:00
committed by meta-codesync[bot]
parent 44afa7729d
commit 39079668a8
3 changed files with 12 additions and 6 deletions
Download Patch File Download Diff File Expand all files Collapse all files
conda/faiss-gpu-cuvs/meta.yaml
+4 -2
View File
@@ -1,4 +1,5 @@
# Copyright (c) Facebook, Inc. and its affiliates.
# Copyright 2026 Arm Limited and/or its affiliates <[email protected]>
#
# This source code is licensed under the MIT license found in the
# LICENSE file in the root directory of this source tree.
@@ -61,14 +62,14 @@ outputs:
host:
- _openmp_mutex =4.5=2_kmp_llvm # [x86_64]
- mkl >=2024.2.2,<2026 # [x86_64]
- openblas =0.3.34 # [not x86_64]
- openblas =0.3.34 openmp_* # [not x86_64]
- libcuvs =26.06
- cuda-version {{ cuda_constraints }}
- libsvs-runtime =0.4.0 # [x86_64 and linux]
run:
- _openmp_mutex =4.5=2_kmp_llvm # [x86_64]
- mkl >=2024.2.2,<2026 # [x86_64]
- openblas =0.3.34 # [not x86_64]
- openblas =0.3.34 openmp_* # [not x86_64]
- cuda-cudart {{ cuda_constraints }}
- libcublas {{ libcublas_constraints }}
- libcuvs =26.06
@@ -83,6 +84,7 @@ outputs:
- test -f $PREFIX/lib/libfaiss_avx2$SHLIB_EXT # [x86_64 and not win]
- conda inspect linkages -p $PREFIX $PKG_NAME # [not win]
- conda inspect objects -p $PREFIX $PKG_NAME # [osx]
- python -c "import ctypes; lib = ctypes.CDLL('libopenblas.so.0'); assert lib.openblas_get_parallel() == 2" # [linux and not x86_64]
- name: faiss-gpu-cuvs
script: build-pkg.sh # [x86_64 and not win and not osx]
conda/faiss-gpu/meta.yaml
+4 -2
View File
@@ -1,4 +1,5 @@
# Copyright (c) Facebook, Inc. and its affiliates.
# Copyright 2026 Arm Limited and/or its affiliates <[email protected]>
#
# This source code is licensed under the MIT license found in the
# LICENSE file in the root directory of this source tree.
@@ -60,11 +61,11 @@ outputs:
- cuda-toolkit {{ cudatoolkit }}
host:
- mkl >=2024.2.2,<2026 # [x86_64]
- openblas =0.3.34 # [not x86_64]
- openblas =0.3.34 openmp_* # [not x86_64]
- libsvs-runtime =0.4.0 # [x86_64 and linux]
run:
- mkl >=2024.2.2,<2026 # [x86_64]
- openblas =0.3.34 # [not x86_64]
- openblas =0.3.34 openmp_* # [not x86_64]
- cuda-cudart {{ cuda_constraints }}
- libcublas {{ libcublas_constraints }}
- libsvs-runtime =0.4.0 # [x86_64 and linux]
@@ -76,6 +77,7 @@ outputs:
- test -f $PREFIX/lib/libfaiss_avx2$SHLIB_EXT # [x86_64 and not win]
- conda inspect linkages -p $PREFIX $PKG_NAME # [not win]
- conda inspect objects -p $PREFIX $PKG_NAME # [osx]
- python -c "import ctypes; lib = ctypes.CDLL('libopenblas.so.0'); assert lib.openblas_get_parallel() == 2" # [linux and not x86_64]
- name: faiss-gpu
script: build-pkg.sh # [x86_64 and not win and not osx]
conda/faiss/meta.yaml
+4 -2
View File
@@ -1,4 +1,5 @@
# Copyright (c) Facebook, Inc. and its affiliates.
# Copyright 2026 Arm Limited and/or its affiliates <[email protected]>
#
# This source code is licensed under the MIT license found in the
# LICENSE file in the root directory of this source tree.
@@ -69,7 +70,7 @@ outputs:
- mkl >=2024.2.2,<2026 # [x86_64 and win]
- python_abi =3.12
{% endif %}
- openblas =0.3.34 # [linux and not x86_64]
- openblas =0.3.34 openmp_* # [linux and not x86_64]
- libopenblas =0.3.34 # [osx]
run:
- python {{ python }}
@@ -83,7 +84,7 @@ outputs:
- mkl >=2024.2.2,<2026 # [x86_64 and win]
- python_abi =3.12
{% endif %}
- openblas =0.3.34 # [linux and not x86_64]
- openblas =0.3.34 openmp_* # [linux and not x86_64]
- libopenblas =0.3.34 # [osx]
test:
requires:
@@ -95,6 +96,7 @@ outputs:
- test -f $PREFIX/lib/MetalDistance.metallib # [osx and arm64]
- conda inspect linkages -p $PREFIX $PKG_NAME # [not win]
- conda inspect objects -p $PREFIX $PKG_NAME # [osx]
- python -c "import ctypes; lib = ctypes.CDLL('libopenblas.so.0'); assert lib.openblas_get_parallel() == 2" # [linux and not x86_64]
- name: faiss-cpu
script: build-pkg.sh # [x86_64 and not win and not osx]