Hi ARM folks!
I have some work to do with EAS + IPA. I have a platform with arm64 4 big + 4 LITTLE that will hopefully get EAS and IPA (when I solve some thermal issues). Unfortunately, in this SoC and code we don't have idle states (new SoCs will just get normal psci support). I run different kernel versions, also your latest integration-next (after dropping whole /drivers/usb because the changes in that branch broke it).
Have you ever tried to evaluate EAS on a platform which has only WFI idle? Do you know what could be the impact on the EAS? I have some other thought and issues to discuss regarding EAS (migrations, load balancing, etc) but this question regarding idle is probably fundamental.
Cheers, Lukasz Luba
Hi Lukasz!
On 10/09/18 23:27, Lukasz Luba wrote:
Hi ARM folks!
I have some work to do with EAS + IPA. I have a platform with arm64 4 big + 4 LITTLE that will hopefully get EAS and IPA (when I solve some thermal issues). Unfortunately, in this SoC and code we don't have idle states (new SoCs will just get normal psci support). I run different kernel versions, also your latest integration-next (after dropping whole /drivers/usb because the changes in that branch broke it).
Those USB changes are from the HiKey960 topic branch that I cobble together from John Stultz's mainline WIP branch ([1]). We need those because the HiKey960 is a really nice test platform, but it's not entirely supported in mainline (yet).
The topic branches are not exposed in the public integration branch, but if that helps the HiKey960 changes on integration_20180831 are between
5a4a0bcc5c5f arm64: hikey960: update idle-states up to 0eb8cbafffcb arm64: defconfig: enable f2fs and squashfs
You probably want to completely remove those changes from your local branch if you don't plan to play with HiKey960.
Have you ever tried to evaluate EAS on a platform which has only WFI idle? Do you know what could be the impact on the EAS? I have some other thought and issues to discuss regarding EAS (migrations, load balancing, etc) but this question regarding idle is probably fundamental.
Quentin's new Energy Model ([2]) doesn't take into account idle states, it just uses IPA's good old P = C * v * f², so I'd say EAS (at least the latest expression of it) would work fine on such platforms.
Cheers, Lukasz Luba
Cheers, Valentin
[1]: https://git.linaro.org/people/john.stultz/android-dev.git/log/?h=dev/hikey96... [2]: https://lkml.org/lkml/2018/8/20/150
Hi Valentin,
Thanks for your instant response. I am still waiting for hikey960 board to test the default EAS behavior and compare with my SoC. I hope I will be able to share some results soon.
Cheers, Lukasz
On 09/11/2018 08:47 AM, Valentin Schneider wrote:
Hi Lukasz!
On 10/09/18 23:27, Lukasz Luba wrote:
Hi ARM folks!
I have some work to do with EAS + IPA. I have a platform with arm64 4 big + 4 LITTLE that will hopefully get EAS and IPA (when I solve some thermal issues). Unfortunately, in this SoC and code we don't have idle states (new SoCs will just get normal psci support). I run different kernel versions, also your latest integration-next (after dropping whole /drivers/usb because the changes in that branch broke it).
Those USB changes are from the HiKey960 topic branch that I cobble together from John Stultz's mainline WIP branch ([1]). We need those because the HiKey960 is a really nice test platform, but it's not entirely supported in mainline (yet).
The topic branches are not exposed in the public integration branch, but if that helps the HiKey960 changes on integration_20180831 are between
5a4a0bcc5c5f arm64: hikey960: update idle-states up to 0eb8cbafffcb arm64: defconfig: enable f2fs and squashfs
You probably want to completely remove those changes from your local branch if you don't plan to play with HiKey960.
Have you ever tried to evaluate EAS on a platform which has only WFI idle? Do you know what could be the impact on the EAS? I have some other thought and issues to discuss regarding EAS (migrations, load balancing, etc) but this question regarding idle is probably fundamental.
Quentin's new Energy Model ([2]) doesn't take into account idle states, it just uses IPA's good old P = C * v * f², so I'd say EAS (at least the latest expression of it) would work fine on such platforms.
Cheers, Lukasz Luba
Cheers, Valentin
-
Lukasz Luba -
Valentin Schneider