Feature/vecadd sharding

[dchenTT]

Ticket

Link to Github Issue

Problem description

A new programming example of vecadd using sharding

What's changed

1. Provide examples for sharding type of HEIGHT, WIDTH, and BLOCK.
2. Host create sharded buffer on L1 directly.
3. Host write vector a and b to L1 sharded buffer, read c from L1 to host. No read and write kernel is needed. Only compute kernel.
4. cb buffer points to L1 buffer address with global address.

Checklist

• Post commit CI passes
• Blackhole Post commit (if applicable)
• Model regression CI testing passes (if applicable)
• Device performance regression CI testing passes (if applicable)
• (For models and ops writers) Full new models tests passes
• New/Existing tests provide coverage for changes