Dense iter

[sgauthamr2001]

No description provided.

[Copilot]

Pull Request Overview

This PR introduces support for dense iterations and integrates a new data_format mechanism to handle both sparse and dense tensor representations. Key changes include:

• Enhancements in pre_process.py to add a new tensor dumper class and a tiled tensor parser.
• Addition of mem_op_gen.py for generating C++ memory operation code and adjustments in main.py and gold_cgen.py to integrate the data_format workflow.
• Extensive modifications in codegen.py to propagate the new data_format and data_format_dict parameters in various code‐generation routines.

Reviewed Changes

Copilot reviewed 5 out of 7 changed files in this pull request and generated 1 comment.

Show a summary per file

File	Description
pre_process.py	Added PydataSparseTensorDumper class and parse_tiled_tensor function; updated process_coo logic to integrate data_format.
mem_op_gen.py	New file for generating C++ struct definitions and memory operation functions.
main.py	Updated parsing functions and tensor declarations to support data_format.
gold_cgen.py	Minor modifications to integrate the data_format dict in the generated gold code.
codegen.py	Extensive changes to include data_format parameters in code generation functions and update control‐flow statements accordingly.

Files not reviewed (2)

• input/program.txt: Language not supported
• input/tensor.txt: Language not supported

Comments suppressed due to low confidence (2)

pre_process.py:215

• Consider using 'if keys is not None:' instead of 'if keys != None:' to ensure proper type-safe comparison.

if keys != None:

codegen.py:501

• Using a fallback condition 'if(1){' when no sparse sub-points are found may cause unintended execution of the if-block. Please verify that this behavior is intentional.

return ["if(1){"]

[Copilot]

There is commented-out code within the cp_op_stmt function that could be removed to reduce confusion and improve code clarity if it is no longer necessary.

Suggested change

	# stmt += " " * (level + 2)
	# stmt += "else if (mode == \"reduce\")\n"
	# stmt += " " * (level + 3)
	# stmt += "partial = read_subtile_output(subtile_path);\n"