Add apply-load mode for searching ledger limits.

[dmkozh]

Description

Add apply-load mode for searching ledger limits.

The new mode finds the ledger limits that allow applying the maximum possible TPL that still allows closing ledgers within the configured time limit, e.g. find how many Soroswap-like transactions can be applied within N ms, and what are the respective ledger limits (with some rounding to get nicer numbers).

Checklist

• Reviewed the contributing document
• Rebased on top of master (no merge commits)
• Ran clang-format v8.0.0 (via make format or the Visual Studio extension)
• Compiles
• Ran all tests
• If change impacts performance, include supporting evidence per the performance document

[Copilot]

Pull request overview

This PR adds a new "limits-for-model-tx" mode to the apply-load benchmark tool that automatically finds optimal ledger limits for a given "model" transaction type. The mode performs a binary search to determine the maximum number of model transactions that can be applied within a target close time, then calculates corresponding ledger limits.

Key Changes

• Introduces ApplyLoadMode::FIND_LIMITS_FOR_MODEL_TX enum value with corresponding findMaxLimitsForModelTransaction() implementation that uses binary search with rounding to find optimal limits
• Refactors existing apply-load functionality by extracting common code into execute(), benchmarkLimits(), and benchmarkLimitsIteration() methods
• Consolidates APPLY_LOAD_MAX_SAC_TPS_TARGET_CLOSE_TIME_MS and APPLY_LOAD_TARGET_CLOSE_TIME_MS config parameters for use across multiple modes

Reviewed changes

Copilot reviewed 10 out of 10 changed files in this pull request and generated 13 comments.

Show a summary per file

File	Description
src/simulation/ApplyLoad.h	Adds new mode enum, refactors public API with execute() method, and adds private methods for each mode (benchmarkLimits(), findMaxLimitsForModelTransaction(), findMaxSacTps())
src/simulation/ApplyLoad.cpp	Implements binary search logic for finding optimal limits, refactors benchmark code into mode-specific methods, adds validation for model transaction parameters, and updates hot archive entry calculation
src/simulation/TxGenerator.h	Adds SOROBAN_LOAD_V2_EVENT_SIZE_BYTES constant for event size calculations
src/simulation/TxGenerator.cpp	Refactors resource assignment to use target instructions directly, improving consistency with model transaction configuration
src/simulation/test/LoadGeneratorTests.cpp	Updates existing test to use execute() method and adds new test for FIND_LIMITS_FOR_MODEL_TX mode
src/main/Config.h	Renames and consolidates target close time configuration, adds documentation for ledger count parameter usage
src/main/Config.cpp	Updates config parsing for renamed parameter
src/main/CommandLine.cpp	Adds parsing for "limits-for-model-tx" mode, consolidates benchmark execution into execute() call, moves output logging into ApplyLoad class
docs/software/commands.md	Documents the three apply-load modes including the new limits-for-model-tx mode
docs/apply-load-limits-for-model-tx.cfg	Provides example configuration file with detailed comments for the new mode