[CPU] Change kvcache default type of PagedAttention to u8 for CPU plugin #1206
+12
−22
Conversation
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
github-actions
bot
added
category: continuous batching
src/cpp/src/device_config.hpp
Outdated
| // if user sets ov::kv_cache_precision hint | ||
| const auto kv_cache_precision_it = plugin_config.find(ov::hint::kv_cache_precision.name()); | ||
| if (kv_cache_precision_it != plugin_config.end()) { | ||
| const auto kv_cache_precision = kv_cache_precision_it->second.as<ov::element::Type>(); | ||
| m_kv_cache_type = kv_cache_precision; | ||
| } else { | ||
| // x86 and arm have different default kv cache type |
There was a problem hiding this comment.
what if plugin_config has accuracy execution hint or inference precision?
There was a problem hiding this comment.
Updated. For 'EXECUTION_MODE_HINT': 'ACCURACY' the plugin will force f32 precision for kvcache. If inference precision was set to f32, the kvcache is still in default u8 precision.
|
@luo-cheng2021 could you please also include reverting of #1212 ? |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
Change kvcache default type of PagedAttention to u8 for CPU plugin to aligned SDPA behaviour.