[Feat] Implement CPU and DRAM monitoring for zeusd

[wbjin]

Additions

• /cpu_id/get_index_energy route for getting CPU and DRAM Rapl energy measurements. Request body includes "cpu" and "dram" fields that indicate whether zeusd should poll cpu and/or dram. Response body contains "cpu_energy_uj" and "dram_energy_uj".

Changes

• devices/gpu/linux.rs device_count method. Changed so that returns 0 if NVML is unavailable or an error occurs. TODO: change method so that it doesn't squash NVML errors that users will have to be aware of.
• startup.rs start_cpu_device_tasks method. Create CpuManagementTasks used in main.rs. Also added CpuManagementTasks to app_data in start_server_uds and start_server_tcp methods.
• main.rs. Added a CpuManagementTasks stop_monitoring method call to stop monitoring tasks.

[jaywonchung]

Thanks a lot for your work! This is a super valuable addition to the Zeus daemon. I've left some comments and questions. Let me know what you think.

[wbjin]

I'm getting this compile error with the set_persistent method all of a sudden.

error[E0599]: no method named `set_persistent` found for struct `nvml_wrapper::Device` in the current scope
  --> src/devices/gpu/linux.rs:45:24
   |
45 |         Ok(self.device.set_persistent(enabled)?)
   |                        ^^^^^^^^^^^^^^ method not found in `Device<'static>`

For more information about this error, try `rustc --explain E0599`.
error: could not compile `zeusd` (lib) due to 1 previous error

Was there a change with Nvml?

[jaywonchung]

The error indicates a Rust NVML wrapper side issue. If actual NVML (libnvidia-ml.so) removed it, you'll get a runtime symbol not found error or alike from dynamic loading. See if stuff like nvml-wrapper or nvml-wrapper-sys were updated. I think it's very unlikely that that API will be removed from NVML.

[jaywonchung]

Look at nvml-wrapper's source code of set_persistent and see if there are any #[cfg] directives that may have removed the method during compilation. That may hint an some unintended change in your compilation environment.

[jaywonchung]

Nothing major! Thanks a lot for the great work! This will be the last round of reviewing.

[jaywonchung]

LGTM!

[jaywonchung]

Oh, there seems to be a merge conflict that needs to be resolved.

Resolved merge conflict.

[jaywonchung]

Failing test is unrelated to this PR. Merging, thanks again @wbjin!