Simplified Test File Naming Scheme

This document explains the simplified naming scheme for test files in the zkEVM benchmark workload project.

Overview

The original test file names follow a complex pattern that includes redundant information:

test_file.py::test_function[fork_Prague-benchmark-gas-value_1M-blockchain_test-{additional_params}].json

The simplified naming scheme extracts only the meaningful additional parameters:

{category}_{additional_params}.json

Naming Rules

Category Prefixes

Original Test File	Category Prefix
`test_worst_blocks.py`	`blocks`
`test_worst_bytecode.py`	`bytecode`
`test_worst_compute.py`	`compute`
`test_worst_memory.py`	`memory`
`test_worst_opcode.py`	`opcode`
`test_worst_stateful_opcodes.py`	`stateful`

Parameter Simplifications

Common Parameters (Removed)

fork_Prague - All tests use Prague fork
benchmark-gas-value_1M - All tests use 1M gas limit
blockchain_test - Test source type
blockchain_test_from_state_test - Test source type

Memory Parameters

big_memory_expansion_True → big_mem
big_memory_expansion_False → small_mem
offset_initialized_True → init_offset
offset_initialized_False → uninit_offset
offset_N → off_N

Data Parameters

non_zero_data_True → non_zero
non_zero_data_False → zero_data
fixed_offset_True → fixed
fixed_offset_False → dynamic
zeros_topic → zero_topic
non_zero_topic → non_zero_topic

Value Parameters

value_bearing_True → with_value
value_bearing_False → no_value
absent_target_True → absent_target
absent_target_False → present_target

Size Parameters

0 bytes → 0bytes
100 bytes → 100bytes
1 MiB → 1MiB
0.25x max code size → 0.25x_max_code
max code size → max_code

Opcode Parameters

opcode_OPCODE_NAME → OPCODE_NAME

Case Parameters

case_id_CASE_NAME → CASE_NAME

Examples

Original vs Simplified Names

Original Name	Simplified Name
`test_worst_compute.py::test_worst_memory_access[fork_Prague-benchmark-gas-value_1M-blockchain_test_from_state_test-big_memory_expansion_True-offset_initialized_False-offset_0-opcode_MSTORE8].json`	`compute_big_mem_uninit_offset_off_0_MSTORE8.json`
`test_worst_compute.py::test_worst_memory_access[fork_Prague-benchmark-gas-value_1M-blockchain_test_from_state_test-big_memory_expansion_False-offset_initialized_False-offset_31-opcode_MSTORE].json`	`compute_small_mem_uninit_offset_off_31_MSTORE.json`
`test_worst_stateful_opcodes.py::test_worst_selfdestruct_created[fork_Prague-benchmark-gas-value_1M-blockchain_test_from_state_test-value_bearing_False].json`	`stateful_no_value.json`
`test_worst_opcode.py::test_worst_log_opcodes[fork_Prague-benchmark-gas-value_1M-blockchain_test_from_state_test-fixed_offset_False-zeros_topic-1_MiB_non_zero_data-log3].json`	`opcode_dynamic_zero_topic_1_MiB_non_zero_data_log3.json`

Usage

Using the Simplification Script

The simplify_test_names.py script can be used to rename files:

# Show what would be renamed (dry run)
python scripts/simplify_test_names.py /path/to/test/files --dry-run
 
# Actually perform the renames
python scripts/simplify_test_names.py /path/to/test/files --execute
 
# Create a shell script for renaming
python scripts/simplify_test_names.py /path/to/test/files --create-script

Generated Shell Script

The script can generate a shell script (rename_tests.sh) that contains the actual mv commands:

./rename_tests.sh

Benefits

Readability: Much shorter and more readable file names
Focus: Only includes the parameters that actually vary between tests
Consistency: Standardized naming across all test categories
Maintainability: Easier to understand what each test does at a glance
File System Friendly: Shorter names work better with file system limitations

Implementation Notes

The script handles parameter parsing carefully to avoid splitting parameters that contain hyphens
Common parameters that are the same across all tests are automatically filtered out
The script preserves the original files and only shows proposed renames in dry-run mode
Generated shell scripts can be reviewed before execution

Integration with Results Generation

The simplified naming scheme is integrated with the results generation scripts:

# Use simplified names in markdown tables
python3 scripts/generate_markdown_tables.py --name-format simplified zkevm-metrics-1M
 
# Use display names (which incorporate simplified naming)
python3 scripts/generate_markdown_tables.py --name-format display zkevm-metrics-1M

This makes the generated reports much more readable and easier to understand.