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

Does Amazon EC2 c5.xarge support simd for cloud hosting #114

Open
wycst opened this issue Feb 14, 2025 · 2 comments
Open

Does Amazon EC2 c5.xarge support simd for cloud hosting #114

wycst opened this issue Feb 14, 2025 · 2 comments

Comments

@wycst
Copy link
Contributor

wycst commented Feb 14, 2025

Hello, I would like to inquire whether the machine used for testing is still Amazon EC2 c5.xarge, or as mentioned in the question, whether the machine can support vector operations well. For example, the new feature added to JDK17 can be enabled by adding -- add modules=JDK.incubator. vector. However, if the hardware does not support it, it can actually cause the performance of the code that uses vector acceleration to decrease significantly (possibly by about 10 times) at runtime. If the hardware does not support vectors, can we consider removing the -- add modules=JDK.incubator. vector startup parameter.
The latest version of the wast library will enable partial vector acceleration based on whether -- add modules=jdk. incubator. vector has been added (which can normally accelerate by about 6%). If possible, could you please help run the results separately? Thank you very much!

./run deser --apis databind --libs wast --size 1000 --datatype users

@fabienrenaud
Copy link
Owner

I don't know. But it sounds like something you can find out and report on (you actually have more info than me on this at this point since I wasn't even aware of this).

I'll run benchmarks on a single type of instance. Please let me know which you think is most capable/appropraite. Otherwise, I'll likely just pick the latest cX instance type next time.

@wycst
Copy link
Contributor Author

wycst commented Feb 14, 2025

I don't know. But it sounds like something you can find out and report on (you actually have more info than me on this at this point since I wasn't even aware of this).

I'll run benchmarks on a single type of instance. Please let me know which you think is most capable/appropraite. Otherwise, I'll likely just pick the latest cX instance type next time.

Can we run the above single test commands separately, adding --add-modules=jdk.incubator.vector to the startup parameter for the first time and not adding it for the second time? Let's see the results of the two runs. Generally, if the hardware supports vector acceleration, the first result will be better than the second one.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

2 participants