Hetero support continuous batching #30371
Open
+325
−4
Conversation
This file contains hidden or 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: Core
WeldonWangwang
changed the title
WeldonWangwang
changed the title
songbell
reviewed
May 7, 2025
|
|
||
| jit.AddConstant(MakeJitConstant("REMAINDER_K", !k_full)); | ||
| jit.AddConstant(MakeJitConstant("KV_GROUP_SIZE", Q_num_heads_dim.v / K_num_heads_dim.v)); | ||
| jit.AddConstant(MakeJitConstant("KV_GROUP_SIZE", 1)); |
There was a problem hiding this comment.
does it mean hetero do not support grouped attention?
There was a problem hiding this comment.
This is still under improvement and will be modified from the graph structure
songbell
reviewed
May 7, 2025
| tensor->copy_to(dst, src_offset, dst_offset + i * get_strides()[0], roi_shape); | ||
| i++; | ||
| } | ||
| } |
There was a problem hiding this comment.
have you tried beam search? can it work with hetero? it had problem with tensor parallel, not sure hetero
There was a problem hiding this comment.
Hi bell, I haven't tried beam search yet, will try it.
zhaixuejun1993
force-pushed
the
wangwang/hetero_support_cb
branch
from
5bb2915 to
1b6f2ee
Compare
| @@ -433,6 +433,9 @@ ov::Any Plugin::get_property(const std::string& name, const ov::AnyMap& options) | |||
| if (options.find(ov::device::id.name()) != options.end()) { | |||
| device_id = options.find(ov::device::id.name())->second.as<std::string>(); | |||
| } | |||
| if (!(m_configs_map.find(device_id) != m_configs_map.end())) { | |||
There was a problem hiding this comment.
Remove the debug log.
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Outdated suggestions cannot be applied.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.
Details:
Tickets: