added example cpp: example_commit_with_device_memory_view()

[krakhit]

Describe the changes

This PR...
Added an example for simple commit that makes use of polynomial views.
Output attached

Example: a) commit with Polynomial views [(f1+f2)^2 + (f1-f2)^2 ]_1 = [4 (f1^2+ f_2^2)]_1
Example: b) commit with Polynomial views [(f1+f2)^2 - (f1-f2)^2 ]_1 = [4 f1 *f_2]_1
Setup: Generating mock SRS
Setup: SRS of length 1025 generated and loaded to device. Took: 19557 milliseconds
Setup: Generating polys (on device) f1,f2 of log degree 10
Setup: Gen poly done. Took: 7 milliseconds
Computing constraints..start 
Computing constraints..done. Took: 0 milliseconds
Computing Commitments with poly view
Commitments done. Took: 29 milliseconds
commitment [(f1+f2)^2 + (f1-f2)^2]_1:
[x: 0x1e35d81da10e5026dacdd907d6ed0dde673de449ff8c0137ec6acbfd6b1dfe1b, y: 0x21fc051415af35a781f84ebcf999313d489ae38ebefa561c9de2fb0b11091502]
commitment [[2 (f_1^2+f_2^2]_1:
[x: 0x1e35d81da10e5026dacdd907d6ed0dde673de449ff8c0137ec6acbfd6b1dfe1b, y: 0x21fc051415af35a781f84ebcf999313d489ae38ebefa561c9de2fb0b11091502]
commitment [(f1+f2)^2 - (f1-f2)^2]_1:
[x: 0x21e9dc012aef8d95107fbfe63f455d4345b9b21e37bcb0a49043b1066e211ffa, y: 0x2d6a3b2f1be1042a17c58ff595134b9cceb71d1af4f1c67a5696859cd4bafae3]
commitment [4 f_1*f_2]_1:
[x: 0x21e9dc012aef8d95107fbfe63f455d4345b9b21e37bcb0a49043b1066e211ffa, y: 0x2d6a3b2f1be1042a17c58ff595134b9cceb71d1af4f1c67a5696859cd4bafae3]

Linked Issues

Resolves #

[yshekel]

(1) what is this ',' syntax? no need to ';'?
(2) you'd better use the CHK_IF_RETURN macro on cuda runtime calls to make sure they do not fail. You can see it in the codebase everywehre.

[krakhit]

removed memory allocs and let bn254msm api manage memory on device

[yshekel]

you need to set the 'is_on_device' fields to true in the config since the coefffs and scalars are on device.

[krakhit]

set

config.are_points_on_device = true;
config.are_scalars_on_device = true;

[yshekel]

overall looks good but I wrote a few comments