Bug 2552 - x32: identify workload & collect benchmark data to show value of x32
Summary: x32: identify workload & collect benchmark data to show value of x32
Status: RESOLVED DUPLICATE of bug 4578
Alias: None
Product: OE-Core
Classification: Build System, Metadata & Runtime
Component: core (show other bugs)
Version: unspecified
Hardware: x86 Multiple
: Low normal
Target Milestone: Future
Assignee: Anuj Mittal
QA Contact:
URL:
Whiteboard:
Depends on:
Blocks:
 
Reported: 2012-06-06 22:52 UTC by Nitin Kamble
Modified: 2018-07-09 01:39 UTC (History)
5 users (show)

See Also:
OS type for building Yocto: ---
Type of Regression: ---
Verified:
Documentation change: Don't know

Attachments
Add an attachment (proposed patch, testcase, etc.)

Note You need to log in before you can comment on or make changes to this bug.
Description Nitin Kamble 2012-06-06 22:52:34 UTC 
We need to identify a workload which can show value of x32 to customers. It will help customers to switch to x32 ABI for production use.
Comment 1 Nitin Kamble 2012-06-06 23:25:20 UTC 
spec CPU is coined as a good candidate for this purpose


So need to collect data on a real hardware for spec CPU benchmark for x32, i686 & x86-64 ABIs
Comment 2 Nitin Kamble 2012-06-13 14:27:03 UTC 
Bogdan,
  You can use a core i5 system with sugarbay bsp for this work.
Comment 3 Radu Moisan 2012-12-03 09:46:47 UTC 
I will be taking care of this one.
Nitin can you give me a hint about the workload for this bug, I need to fill in some estimate to help Song plan stuff for 1.4
I'll go for now with 5days.
Comment 4 Nitin Kamble 2012-12-03 17:10:47 UTC 
Radu,
 The workload is already identified for this purpose. It is Spec CPU as seen in the comment 1 bellow.
 Now next part is running it in 3 different configurations. 1st with a x86 image, 2nd with x86_64 iamge & 3rd with x32 image. And all this on the same hardware, so that other than the ABI all other parameters are the same. Also you need to collect 3 sets of the same data, to understand any variation in the collected data.
 I am changing the title of this bug to reflect the current work involved.
And I think 5 days will be highly optimistic estimate. If you don't encounter any build or runtime issues, then 5 days may be right. But I am not sure the current state of x32 support. If the current x32 support is broken then that need to be fixed 1st. Any way, you can open bugs for such blocking issues and mark this bug dependent on those. so you can go with the 5 days or add 1 or 2 more days to be more safe.
  I am also sending you instructions I received from HJ Lu for obtaining the benchmark files.

Nitin
Comment 5 Radu Moisan 2012-12-04 07:47:24 UTC 
thanks Nitin, I'm going with 6 days and update appropriately once I get into the problem.
Comment 6 Saul Wold 2017-03-20 17:08:20 UTC 
This could be similar to phoronix work also
Comment 7 Anuj Mittal 2018-07-09 01:39:10 UTC 
Duplicate of 4578.

*** This bug has been marked as a duplicate of bug 4578 ***