Added more detailed timers #27
Conversation
There was a problem hiding this comment.
@johnmauff, just a small change to be made in two places and I think this will be ready to go.
There was a problem hiding this comment.
Thanks @johnmauff. These changes look good and will be merged soon.
|
I think I will make the changes to the timer that was mentioned during the call. I will send an update shortly. |
|
@gdicker1 Dylan, I meant to update the form of the timers in the ew-develop-timers branch, but looking at the history I am concerned that I committed them to ew-develop instead. Sorry if I did this accidentally. |
|
@johnmauff, everything looks good on GitHub for the branch and in the PR. Thanks for the heads up! |
|
@johnmauff, do you think you could rebase this based on cam6_4_019? I got a merge conflict I wasn't sure how to resolve when I tried layering this on top of #31 |
|
Dylan,
Not quite sure how to rebase the PR on cam6_4_019. Could you tell me how
John
…On Fri, Aug 30, 2024 at 4:04 PM G. Dylan Dickerson ***@***.***> wrote:
@johnmauff <https://github.com/johnmauff>, do you think you could rebase
this based on cam6_4_019? I got a merge conflict I wasn't sure how to
resolve when I tried layering this on top of #31
<#31>
—
Reply to this email directly, view it on GitHub
<#27 (comment)>,
or unsubscribe
<https://github.com/notifications/unsubscribe-auth/ADH7NUXYJZWZMQRZWQ6GBQTZUDT6NAVCNFSM6AAAAABMI3TDPKVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDGMRSGQYTGMZWGA>
.
You are receiving this because you were mentioned.Message ID:
***@***.***>
|
|
This PR has been replaced by #32 |
This PR adds more detailed timers to CAM to help identify which areas of CAM have been GPU-enabled. These changes can be grouped into two different types:
(2). Additional detail was encoded into some of the timers' names to indicate whether it was executed on the CPU or GPU.
(3) Host-to-device and device-to-host data movement was also indicated through the use of the H2D or D2H strings