Zkcuda new API and compilation refactor

[siq1]

The goal of this PR is to allow (arbitrary) reshape and transpose in Zkcuda. It will solve #133, and will utilize #161.

For overall mechanism, please refer to #133.

---

For circuit developers, this overall experience only differs a little. Here's an example:

Previous code:

let kernel_add_2: Kernel<M31Config> = compile_add_2_macro().unwrap();
let kernel_add_16: Kernel<M31Config> = compile_add_16_macro().unwrap();

let mut ctx: Context<M31Config> = Context::default();
let a = ctx.copy_to_device(&a, false);
let mut b = None;
call_kernel!(ctx, kernel_add_2, a, mut b);
let b = b.reshape(&[1, 16]);
let mut c = None;
call_kernel!(ctx, kernel_add_16, b, mut c);
let c = c.reshape(&[]);
let result: M31 = ctx.copy_to_host(c);
assert_eq!(result, M31::from(32 * 33 / 2));

let computation_graph = ctx.to_computation_graph();
let (prover_setup, verifier_setup) =
    ExpanderGKRProvingSystem::<M31Config>::setup(&computation_graph);
let proof = ExpanderGKRProvingSystem::<M31Config>::prove(
    &prover_setup,
    &computation_graph,
    &ctx.device_memories,
);
assert!(ExpanderGKRProvingSystem::<M31Config>::verify(
    &verifier_setup,
    &computation_graph,
    &proof
));

Current code:

// Kernel -> KernelPrimitive
let kernel_add_2_tmp: KernelPrimitive<M31Config> = compile_add_2_macro().unwrap();
let kernel_add_16: KernelPrimitive<M31Config> = compile_add_16_macro().unwrap();

let mut ctx: Context<M31Config> = Context::default();
let a = ctx.copy_to_device(&a); // Removed argument is_broadcast 
let mut b = None;
call_kernel!(ctx, kernel_add_2, 16, a, mut b).unwrap(); // Additional argument num_parallel
let b = b.reshape(&[1, 16]);
let mut c = None;
call_kernel!(ctx, kernel_add_16, 1, b, mut c).unwrap();
let c = c.reshape(&[]);
let result: M31 = ctx.copy_to_host(c);
assert_eq!(result, M31::from(32 * 33 / 2));

let computation_graph = ctx.compile_computation_graph().unwrap(); // Function name and behavior changed
ctx.solve_witness().unwrap(); // Additional step
let (prover_setup, verifier_setup) =
    ExpanderGKRProvingSystem::<M31Config>::setup(&computation_graph);
let proof = ExpanderGKRProvingSystem::<M31Config>::prove(
    &prover_setup,
    &computation_graph,
    &ctx.export_device_memories(), // Need to export now
);
assert!(ExpanderGKRProvingSystem::<M31Config>::verify(
    &verifier_setup,
    &computation_graph,
    &proof
));

Execution times also differs:

Particular Step	Previous function	Current function
Compile Func -> IR	compile_xx_kernel	compile_xx_kernel
Compile IR -> layered circuit	compile_xx_kernel	compile_computation_graph
Eval circuit for output	call_kernel	call_kernel
Eval circuit for hint witness	call_kernel	solve_witness