create seperate timers for cpu and gpu

[emre-safa]

gpu timer and cpu timer are now 2 seperate timers

[yhmtsai]

we also test the benchmark output.
you need to modify the *solver*.stdout in benchmark/test/reference.
it should be mostly "apply": { "components": { * } } to "apply": { "components": { "cpu": { * }, "gpu": {} } }

[yhmtsai]

it seems to has order effect in the test.
swapping "gpu" and "cpu" might solve it

[yhmtsai]

also do the same thing for the generate

[yhmtsai]

just notice that you only create cpu/gpu timer for apply but not for generation

[yhmtsai]

I think the current pr did what we discuss for two timer separately although it might need #1896 and #1897 such that the cpu timer get submission overhead only. I approve it but this pr should wait for another reviewer's approval because I involve it quite a lot.

[ginkgo-bot]

Error: The following files need to be formatted:

core/base/extended_float.hpp

You can find a formatting patch under Artifacts here or run format! if you have write access to Ginkgo