Add __quantum__rt__device_init(int8_t *, int8_t *, int8_t *)

[maliasadi]

We currently lower the device name, kwargs, and library with 3 __quantum__rt__device(key, spec) calls. The runtime initializes the targeted device after getting the device name and kwargs at the third call:

__quantum__rt__device("rtd_name", "lightning.qubit");
__quantum__rt__device("rtd_kwargs", "{...}");
__quantum__rt__device("rtd_lib", "rtd_lightning.so"); <--- init device

This PR reduces the number of these device calls into 1 single __quantum__rt__device_init("rtd_lib", "rtd_name", "rtd_kwargs") call and deprecates the __quantum__rt__device instruction. The runtime is able to initialize the device without necessarily storing partial device info or implying any specific order to define these specifications in the IR and the runtime. This PR is related to #381 and will help better scope the lifetime of device instances. We also add __quantum__rt__device_release() to release the initialized device by request (this'll be used in the async support).

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Comparison is base (88bd0e2) 99.48% compared to head (a6efe74) 99.49%.

Additional details and impacted files

@@           Coverage Diff           @@
##             main     #396   +/-   ##
=======================================
  Coverage   99.48%   99.49%           
=======================================
  Files          42       42           
  Lines        7053     7134   +81     
  Branches      477      477           
=======================================
+ Hits         7017     7098   +81     
  Misses         20       20           
  Partials       16       16           

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[josh146]

@maliasadi what is the context behind this change?

[maliasadi]

@maliasadi what is the context behind this change?

@josh146 First, it removes 2 instructions calls from the generated LLVM IR with hardcoding the device triple (name, library, kwargs) from the frontend to the backend in one single rt__device_init call instead of 3 rt__device calls.

// Before this change, (calls should be in some order: name->kwargs->lib)
__quantum__rt__device("rtd_name", "LightningQubit");
__quantum__rt__device("rtd_kwargs", "{...}");
__quantum__rt__device("rtd_lib", "rtd_lightning.so"); <--- init device

// Now,
__quantum__rt__device_init("rtd_lightning.so", LightningQubit", "{...}");

Second, it lets runtime to initialize device instances without keeping record of any partial information about that device. Before this change, the runtime had to store partial information about name and kwargs of the device before initializing it. But now, the runtime is able to initialize the device without necessarily storing partial device info or implying any specific order to define these specifications in the IR and the runtime. This PR is related to #381 and will help better scope the lifetime of device instances.

[josh146]

thanks :)

[dime10]

Looks great, I think it just needs a changelog :)