Batch Processing with Multiple tasks in a task-graph

[jjfumero]

Description

This fixes and adds support for batch processing when having a TaskGraph with multiple tasks. Besides, it removes old/unreachable methods that are not needed anymore.

However, this PR still has the issue of incorrect results when the task-graph does not perform a copy-in. I added two more unit-tests, one with copy in and the other without. It seems that without copy-in, the batch processing works (compiler and internals of the runtime), but the results are not correct.

For task-graphs with copy-in/out (which was the original intended use), multiple tasks in a taskgraph with batch processing now works. We can go ahead this is partial fix, and add a new extension on top of this to process also for only outputs.

Problem description

The issue was that, when having multiple tasks in a task-graph, the TornadoVM interpreter was not invalidating the code cache (to force recompilation).

Backend/s tested

Mark the backends affected by this PR.

• [X] OpenCL
• [ ] PTX
• [ ] SPIRV

OS tested

Mark the OS where this PR is tested.

• [X] Linux
• [ ] OSx
• [ ] Windows

Did you check on FPGAs?

If it is applicable, check your changes on FPGAs.

• [ ] Yes
• [X] No

How to test the new patch?

$ make
$ tornado-test -V --fast --debug --threadInfo --printKernel --printBytecode uk.ac.manchester.tornado.unittests.batches.TestBatches#testBatchNotEven
$ tornado-test -V --fast --debug --threadInfo --printKernel --printBytecode uk.ac.manchester.tornado.unittests.batches.TestBatches#testBatchNotEven2