Skip to content

Env flag affects the peak memory in e2e benchmark. #4937

@qjia7

Description

@qjia7
Contributor

Steps to reproduce:

  1. Access https://tensorflow.github.io/tfjs/e2e/benchmarks/local-benchmark/index.html
  2. Use the default backend wasm and click Run benchmark.
  3. See the left table
Peak memory 19.63 MB
  1. Switch to webgl backend. Change any flags, for example webgl pack. And then switch back to wasm.
  2. Click Run benchmark again. The peak memory increases as below:
Peak memory 33.51 MB
  1. If we repeat 4 and 5, you will find the peak memory gets bigger and bigger. However, if you don't change the flag, the peak memory won't change. It happens for all backends that the flag will affect the peak memory. It seems like a bug.

Activity

qjia7

qjia7 commented on Apr 15, 2021

@qjia7
ContributorAuthor

fyi @lina128. This is the issue that I mentioned in yesterday's meeting.

assigned and unassigned on Apr 15, 2021
assigned and unassigned on Apr 16, 2021
gaikwadrahul8

gaikwadrahul8 commented on May 15, 2023

@gaikwadrahul8
Contributor

Hi, @qjia7

Apologize for the delayed response and we're re-visiting our older issues and checking whether those issues got resolved or not as of now and I tried from my end and still this issue exists so May I know are you still looking for the solution ? Thank you!

For your reference I have added screenshot below:

image

qjia7

qjia7 commented on May 16, 2023

@qjia7
ContributorAuthor

I expect it can be resolved since it's a bug.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Type

No type

Projects

No projects

Milestone

No milestone

Relationships

None yet

    Development

    No branches or pull requests

      Participants

      @lina128@qjia7@Linchenn@rthadur@gbaned

      Issue actions

        Env flag affects the peak memory in e2e benchmark. · Issue #4937 · tensorflow/tfjs