-
Notifications
You must be signed in to change notification settings - Fork 106
Onnx runtime set allocator #619
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Conversation
…me backend. We register RedisModule Alloc and Free functions as onnx allocator, so that onnx is now using them in model create, delete and run. - The memory usage of onnx (via redis allocator) and the number of accesses to the allocator is saved as part of the module info (defined in redisai.c), under a new section named "ai_memory_info". These are exported to RedisAI as new fields in the RAI_LoadedBackend struct.
…t since onnx memory info is saved in redis' INFO report, that can be produced by calling the INFO MODULES command)
3b775bb to
67c40f8
Compare
2d6cde9 to
6626672
Compare
6626672 to
e404178
Compare
rafie
left a comment
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
tests/module/Makefile should be replaced by a corresponding CMakeLists.txt.
2285c68 to
5e4af41
Compare
Codecov Report
@@ Coverage Diff @@
## master #619 +/- ##
==========================================
+ Coverage 76.35% 76.71% +0.36%
==========================================
Files 41 41
Lines 6517 6519 +2
==========================================
+ Hits 4976 5001 +25
+ Misses 1541 1518 -23
Continue to review full report at Codecov.
|
3db8a1b to
f9e45ff
Compare
…I/RedisAI into OnnxRuntime_set_allocator
89fd4a3 to
049417e
Compare
…I/RedisAI into OnnxRuntime_set_allocator
…I/RedisAI into OnnxRuntime_set_allocator
cd7c9d1 to
5972370
Compare
Extend tests to make a more precise check of the allocator usage.
5972370 to
f27405b
Compare
Introduce an API that allows setting a custom allocator for onnxruntime backend. We register RedisModule Alloc and Free functions as onnx allocator, so that onnx is now using them in model create, delete and run.
The memory usage of onnx (via redis allocator) and the number of accesses to the allocator is saved as part of the module info (defined in redisai.c), under a new section named "ai_memory_info". These are exported to RedisAI as new fields in the RAI_LoadedBackend struct.