Skip to content
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

[ Kernel ] FP8 Dynamic-Per-Token Quant Kernel #6511

Conversation

varun-sundar-rabindranath
Copy link
Contributor

Add support for dynamic-per-token fp8 quantization.


PR Checklist (Click to Expand)

Thank you for your contribution to vLLM! Before submitting the pull request, please ensure the PR meets the following criteria. This helps vLLM maintain the code quality and improve the efficiency of the review process.

PR Title and Classification

Only specific types of PRs will be reviewed. The PR title is prefixed appropriately to indicate the type of change. Please use one of the following:

  • [Bugfix] for bug fixes.
  • [CI/Build] for build or continuous integration improvements.
  • [Doc] for documentation fixes and improvements.
  • [Model] for adding a new model or improving an existing model. Model name should appear in the title.
  • [Frontend] For changes on the vLLM frontend (e.g., OpenAI API server, LLM class, etc.)
  • [Kernel] for changes affecting CUDA kernels or other compute kernels.
  • [Core] for changes in the core vLLM logic (e.g., LLMEngine, AsyncLLMEngine, Scheduler, etc.)
  • [Hardware][Vendor] for hardware-specific changes. Vendor name should appear in the prefix (e.g., [Hardware][AMD]).
  • [Misc] for PRs that do not fit the above categories. Please use this sparingly.

Note: If the PR spans more than one category, please include all relevant prefixes.

Code Quality

The PR need to meet the following code quality standards:

  • We adhere to Google Python style guide and Google C++ style guide.
  • Pass all linter checks. Please use format.sh to format your code.
  • The code need to be well-documented to ensure future contributors can easily understand the code.
  • Include sufficient tests to ensure the project to stay correct and robust. This includes both unit tests and integration tests.
  • Please add documentation to docs/source/ if the PR modifies the user-facing behaviors of vLLM. It helps vLLM user understand and utilize the new features or changes.

Notes for Large Changes

Please keep the changes as concise as possible. For major architectural changes (>500 LOC excluding kernel/data/config/test), we would expect a GitHub issue (RFC) discussing the technical design and justification. Otherwise, we will tag it with rfc-required and might not go through the PR.

What to Expect for the Reviews

The goal of the vLLM team is to be a transparent reviewing machine. We would like to make the review process transparent and efficient and make sure no contributor feel confused or frustrated. However, the vLLM team is small, so we need to prioritize some PRs over others. Here is what you can expect from the review process:

  • After the PR is submitted, the PR will be assigned to a reviewer. Every reviewer will pick up the PRs based on their expertise and availability.
  • After the PR is assigned, the reviewer will provide status update every 2-3 days. If the PR is not reviewed within 7 days, please feel free to ping the reviewer or the vLLM team.
  • After the review, the reviewer will put an action-required label on the PR if there are changes required. The contributor should address the comments and ping the reviewer to re-review the PR.
  • Please respond to all comments within a reasonable time frame. If a comment isn't clear or you disagree with a suggestion, feel free to ask for clarification or discuss the suggestion.

Thank You

Finally, thank you for taking the time to read these guidelines and for your interest in contributing to vLLM. Your contributions make vLLM a great tool for everyone!

Copy link

👋 Hi! Thank you for contributing to the vLLM project.
Just a reminder: PRs would not trigger full CI run by default. Instead, it would only trigger fastcheck CI to run, which consists only a small and essential subset of tests to quickly catch errors with the flexibility to run extra individual tests on top (you can do this by unblocking test steps in the Buildkite run).

Full CI run is still required to merge this PR so once the PR is ready to go, please make sure to run it. If you need all test signals in between PR commits, you can trigger full CI as well.

To run full CI, you can do one of these:

  • Comment /ready on the PR
  • Add ready label to the PR
  • Enable auto-merge.

🚀

@@ -320,6 +320,17 @@ def scaled_fp8_quant(
return output, scale


def dynamic_per_token_scaled_fp8_quant(
input: torch.Tensor) -> Tuple[torch.Tensor, torch.Tensor]:
Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

To fit the existing interface, where all quantization schemes go through scaled_fp8_quant, this should probably also be merged up with the scaled_fp8_quant. How do we want to handle this ?
I believe the dynamic-per-token vs dynamic-per-tensor model quantization scheme will come from the model itself and we can plumb it down into the scaled_fp8_quant function. I'd appreciate any references into will how to plumb this down. Thanks.
@robertgshaw2-neuralmagic @comaniac @tlrmchlsmth ?

Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I think this looks good as is. I can handle the plumbing from here in a follow up with some end-to-end model tests

Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Definitely prefer to have a unified API.

@@ -3,6 +3,8 @@

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Mostly refactors in this file!

@robertgshaw2-neuralmagic robertgshaw2-neuralmagic changed the title FP8 Dynamic-Per-Token Quant [ Kernel ] FP8 Dynamic-Per-Token Quant Kernel Jul 17, 2024
@robertgshaw2-neuralmagic
Copy link
Collaborator

/ready

@github-actions github-actions bot added the ready ONLY add when PR is ready to merge/full CI is needed label Jul 17, 2024
csrc/ops.h Outdated Show resolved Hide resolved
__syncthreads();

float const inverted_scale = FP8_E4M3_MAX / block_absmax_val;
bool const vectorize_conversions = hidden_size % 4 == 0;
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Why only vectorize conversions if the hidden_size is divisible by 4? scaled_fp8_conversion_vec has a loop to "Handle the remaining elements if num_elems is not divisible by 4"

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

The input and output pointers need to be aligned at 8-byte and 4-byte addresses respectively. Each row of the input is processed by a single block.

FP8 output data-type requires that hidden-size (each token/row of the output has hidden-size elements) be a multiple of 4. When hidden-size is not a multiple of 4,

  • The first row can be processed like you mentioned.
  • But the second row's address will not be 4-byte aligned and throws a runtime error.

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Why didn't we run into this with the original scaled_fp8_quant_kernel?

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

the scaled_fp8_quant_kernel has a single scale that is applied to all of the input values and it doesn't have to process row-by-row. For quant purposes, that kernel can treat the inputs and output as one big row of values (the "first row" case from above).

tests/kernels/quant_utils.py Outdated Show resolved Hide resolved
tests/kernels/quant_utils.py Outdated Show resolved Hide resolved
Copy link
Collaborator

@tlrmchlsmth tlrmchlsmth left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I left a few in-line comments. Mostly nits, but I also had a one question on the vectorization when the hidden size isn't divisible by 4.

Copy link
Collaborator

@comaniac comaniac left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Approved to unblock.

  1. Same question @tlrmchlsmth that why not use vectorization for non-dividable problem size?
  2. Better to unify the API. Ok to move to the next PR.

Copy link
Collaborator

@tlrmchlsmth tlrmchlsmth left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Thanks!

@robertgshaw2-neuralmagic robertgshaw2-neuralmagic enabled auto-merge (squash) July 17, 2024 23:08
@robertgshaw2-neuralmagic robertgshaw2-neuralmagic merged commit b5241e4 into vllm-project:main Jul 18, 2024
72 of 73 checks passed
@robertgshaw2-neuralmagic robertgshaw2-neuralmagic deleted the varun/dynamic-per-token-fp8 branch July 18, 2024 01:38
fialhocoelho pushed a commit to opendatahub-io/vllm that referenced this pull request Jul 19, 2024
xjpang pushed a commit to xjpang/vllm that referenced this pull request Jul 24, 2024
gnpinkert pushed a commit to gnpinkert/vllm that referenced this pull request Jul 26, 2024
Alvant pushed a commit to compressa-ai/vllm that referenced this pull request Oct 26, 2024
Co-authored-by: Varun Sundar Rabindranath <[email protected]>
Signed-off-by: Alvant <[email protected]>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
ready ONLY add when PR is ready to merge/full CI is needed
Projects
None yet
Development

Successfully merging this pull request may close these issues.

5 participants