Skip to content

[Continuous Batching] Introducing new env variables #67

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.

Sign up for GitHub

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

Merged
merged 2 commits into from
Apr 1, 2025
Merged

[Continuous Batching] Introducing new env variables #67

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
2 commits
Select commit Hold shift + click to select a range
256041f
introducing env variables for AIU Spyre KV cache dimensions
yannicks1 Mar 31, 2025
109e91e
removing prints
yannicks1 Mar 31, 2025
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
2 changes: 2 additions & 0 deletions examples/offline_inference_spyre_cb.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -15,6 +15,8 @@
os.environ['MASTER_ADDR'] = 'localhost'
os.environ['MASTER_PORT'] = '12355'
os.environ['VLLM_SPYRE_USE_CB'] = '1'
os.environ['VLLM_SPYRE_MAX_BATCH_SIZE'] = '4'
os.environ['VLLM_SPYRE_MAX_CONTEXT_LENGTH'] = '2048'

# Sample prompts.
template = (
Expand Down
10 changes: 10 additions & 0 deletions vllm_spyre/envs.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -7,6 +7,8 @@
VLLM_SPYRE_WARMUP_NEW_TOKENS: Optional[List[int]] = None
VLLM_SPYRE_WARMUP_BATCH_SIZES: Optional[List[int]] = None
VLLM_SPYRE_USE_CB: bool = False
VLLM_SPYRE_MAX_BATCH_SIZE: int = 0
VLLM_SPYRE_MAX_CONTEXT_LENGTH: int = 0

environment_variables: Dict[str, Callable[[], Any]] = {
# Defines the prompt lengths the Spyre accelerator should be prepared
Expand Down Expand Up @@ -45,6 +47,14 @@
# If set, use the V1 continuous batching implementation
"VLLM_SPYRE_USE_CB":
lambda: bool(int(os.getenv("VLLM_SPYRE_USE_CB", "0"))),

# Maximal supported batch size
"VLLM_SPYRE_MAX_BATCH_SIZE":
lambda: int(os.getenv("VLLM_SPYRE_MAX_BATCH_SIZE", "0")),

# Maximal supported context length
"VLLM_SPYRE_MAX_CONTEXT_LENGTH":
lambda: int(os.getenv("VLLM_SPYRE_MAX_CONTEXT_LENGTH", "0")),
}


Expand Down
12 changes: 2 additions & 10 deletions vllm_spyre/model_executor/model_loader/spyre.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -15,7 +15,6 @@
from vllm.model_executor.model_loader.weight_utils import (
download_weights_from_hf)
from vllm.model_executor.sampling_metadata import SamplingMetadata
from vllm.platforms import current_platform

import vllm_spyre.envs as envs_spyre

Expand Down Expand Up @@ -346,15 +345,8 @@ def __init__(
max_decode_length)

# physical KV cache (fms wrapper/ AIU Spyre)
# lives in SpyreCausalLM only for convenient model access
warmup_shapes = current_platform.get_warmup_shapes()
max_batch = max(shape["batch_size"] for shape in warmup_shapes)
max_prompt_length = max(shape["prompt_length"]
for shape in warmup_shapes)
max_new_tokens = max(shape["new_tokens"] for shape in warmup_shapes)
# Eventually max_model_len = self.config.max_position_embeddings,
# but saving some memory here to only allocate the max in practise
max_model_len = max_prompt_length + max_new_tokens
max_batch = envs_spyre.VLLM_SPYRE_MAX_BATCH_SIZE
max_model_len = envs_spyre.VLLM_SPYRE_MAX_CONTEXT_LENGTH

if self.config.model_type == 'llama':
num_layers = self.config.num_hidden_layers
Expand Down