generated from thoth-station/template-project
-
Notifications
You must be signed in to change notification settings - Fork 10
Commit
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
Introduce prescription for deepsparse inference engine from Neural Magic
Signed-off-by: Francesco Murdaca <fmurdaca@redhat.com>
- Loading branch information
Francesco Murdaca
committed
Oct 7, 2021
1 parent
5e631e6
commit 8199b6c
Showing
2 changed files
with
66 additions
and
0 deletions.
There are no files selected for viewing
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Original file line number | Diff line number | Diff line change |
---|---|---|
@@ -0,0 +1,33 @@ | ||
units: | ||
sieves: | ||
- name: DeepSparseAVX2Sieve | ||
type: sieve | ||
should_include: | ||
adviser_pipeline: true | ||
runtime_environments: | ||
hardware: | ||
- cpu_families: [6] | ||
# See: | ||
# https://en.wikipedia.org/wiki/Advanced_Vector_Extensions#CPUs_with_AVX2 | ||
# https://en.wikichip.org/wiki/intel/cpuid | ||
# 0x5 Cascade Lake | ||
# 0x6 Broadwell, Cannon Lake | ||
# 0xA Ice Lake | ||
# 0xC Ice Lake, Tiger Lake | ||
# 0xD Ice Lake | ||
# 0xE Skylake, Kaby Lake, Coffee Lake, Ice Lake, Comet Lake, Whiskey Lake | ||
# 0xF Haswell | ||
cpu_models: | ||
not: [5, 6, 10, 12, 13, 14, 15, 16] | ||
match: | ||
package_version: | ||
name: deepsparse | ||
version: "<=0.7.0" | ||
index_url: 'https://pypi.org/simple' | ||
run: | ||
multi_package_resolution: false | ||
stack_info: | ||
- type: ERROR | ||
message: >- | ||
deepsparse requires avx2 architecture to allow deployment of ML model. | ||
link: https://github.com/neuralmagic/deepsparse#hardware-support |
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Original file line number | Diff line number | Diff line change |
---|---|---|
@@ -0,0 +1,33 @@ | ||
units: | ||
sieves: | ||
- name: DeepSparseAVX512Sieve | ||
type: sieve | ||
should_include: | ||
adviser_pipeline: true | ||
runtime_environments: | ||
hardware: | ||
- cpu_families: [6] | ||
# See: | ||
# https://en.wikipedia.org/wiki/Advanced_Vector_Extensions#CPUs_with_AVX-512 | ||
# https://en.wikipedia.org/wiki/AVX-512#VNNI | ||
# https://en.wikichip.org/wiki/intel/cpuid | ||
# 0x5 Knights Mill, Skylake (Server) SP and X, Cascade Lake SP (VNNI) | ||
# 0x6 Cannon Lake | ||
# 0x7 Knights Landing, Rocket Lake (VNNI) | ||
# 0xC Tiger Lake C (VNNI) | ||
# 0xD Ice Lake Y (VNNI), Tiger Lake H (VNNI) | ||
# 0xE Ice Lake U (VNNI) | ||
cpu_models: | ||
not: [85, 87, 102, 125, 126, 133, 140, 141, 167] | ||
match: | ||
package_version: | ||
name: deepsparse | ||
version: "<=0.7.0" | ||
index_url: 'https://pypi.org/simple' | ||
run: | ||
multi_package_resolution: false | ||
stack_info: | ||
- type: ERROR | ||
message: >- | ||
deepsparse requires avx512 architecture (and avx512 with VNNI instructions) to allow deployment of ML model. | ||
link: https://github.com/neuralmagic/deepsparse#hardware-support |