MatSimKit-HT V2.3.0 restructures result storage and browsing with automatic result saving, paginated browsing for large-scale high-throughput calculations, pause and resume, storage space management, and streaming CSV export, stably supporting millions of parameter combinations.

MatSimKit-HT V2.3.0 Released

Release date: July 23, 2026 Release type: Major feature expansion

Version Overview

This release systematically restructures how calculation results are stored and browsed, moving result data management from the interface layer down to the backend storage layer, so the software can stably support high-throughput calculations over large-scale parameter combinations. Calculation results are saved automatically upon success and the interface loads data on demand page by page; together with resumable calculation, storage space management, and streaming data export, the stability and usability of the software in long-running, large-batch calculation scenarios are significantly improved.

This release also includes all updates from the previous V2.2.1 and V2.2.2 patch releases: CSV export of result data was added, and a related issue in the high-temperature strength calculation for Ni-based superalloys was fixed. See "Additional Version Notes" at the end of this document.

New Features

Automatic Result Saving

Overview data and complete curves from successful calculations are saved automatically during the calculation process, without manually selecting results and clicking "Save to Knowledge Base".

After each parameter combination finishes, its overview and curves are written to local storage immediately, with calculation and saving proceeding in sync
Only successfully saved data enters the result list for browsing, filtering, export, and deletion, ensuring that all visible data has been persisted to disk
On restart after an abnormal exit, saved data is automatically validated and repaired to avoid corruption or duplication

Large-Scale High-Throughput Calculation and Paginated Browsing

Pagination, sorting, filtering, and statistics of results are all completed by the backend. The interface only loads the data needed for the current page, so memory usage does not grow with calculation scale, and the software can comfortably handle calculation and browsing of millions of parameter combinations.

The overview table loads page by page; paging, sorting, filtering, and ranking are all handled by the backend
Parameter combinations are generated and calculated one group at a time on demand, instead of being expanded all at once
Results can be filtered by curve status (complete data, partial data, cleaned, not generated)
Browsing settings such as page size and curve display count can be saved and reused next time
As the calculation scale grows, interface responsiveness remains stable and does not lag with increasing result counts

On-Demand Curve Loading and Envelope Display

Curve data is read from storage on demand, and the interface controls how many curves are drawn on screen at once through the "curve display count" parameter, balancing detail inspection with overall trend viewing.

Curves for the current page load on demand, with continuous cross-page curve loading starting from the current page
When the data volume exceeds the curve display count, the interface shows a continuous subset of real curves and displays the curve envelope of the current data range as a background, with a clear indication of the displayed range
When zooming curves, the envelope background is clipped in sync, keeping the view clean
During calculation, the interface automatically follows the latest result page by default, turning to the next page once the current page is full; actively viewing history pages pauses following, which resumes automatically upon returning to the latest page

Pause and Resumable Calculation

Calculation tasks support safe pause and resume, and can continue from the interruption point after the software is closed and restarted, so long-running tasks no longer need to finish in one run.

Pausing stops safely after the current combination completes, protecting calculation engine stability
Closing the software normally saves task progress; on the next launch, the unfinished task can be restored and calculation continued
After resuming, the accumulated saved count, average time per combination, and remaining time are calculated based on actual progress in the current run

Storage Location Settings and Space Management

New settings and management capabilities for the calculation data storage location allow users to review space usage, reclaim redundant space, and migrate the data directory.

During installation, after selecting the software installation path, a "Calculation Data Storage Location" page is provided where users can accept the recommended path or specify their own, with a "Restore Recommended Path" option
A storage management entry inside the software shows actual usage, valid data, reclaimable space, and remaining disk space, with usage listed per session
The data storage location can be changed with existing data migrated to the new location; switching happens only after migration completes, keeping data consistent
Data can be deleted per record, per selection, per filter result, or per calculation session, with a confirmation step that clearly states "once deleted, data cannot be recovered"
Tiered reminders are shown when disk space runs low, and saving is paused when space is insufficient to protect data

Result Data Export

Overview and curve data can be exported to CSV and packaged into a single ZIP file. Export is completed by the backend in streaming fashion, independent of the data loaded in the interface, and can cover the complete result set.

Selected results or all results can be exported, with automatic deduplication and exact totals within the export scope
Curves are output separately by result type, all selected by default, and only result types with actual data are exported
CSV files use UTF-8 BOM and stable English headers, with overview and curve files linked by a unified index column for easy opening and further processing in Excel and other tools
JSON export is also retained, staying compatible with existing usage habits

Layered Calculation Progress Feedback

The calculation interface provides layered feedback including total elapsed time, average time per combination, current combination time, and success/failure counts, making the running status clear at a glance, with each type of metric displayed in its own area without interference.

Improvements

Result Display

In the knowledge base overview table, the "Curves" column is moved forward, headers are aligned with content columns, and bottom rows of long lists are no longer covered by the hint bar

Major Architecture Changes

Result data ownership moved to the backend: calculation results are no longer held in full by the interface layer, but managed uniformly by the backend storage layer and queried on demand by the interface, fundamentally removing the memory and transfer bottlenecks of large-scale calculations
Storage format upgrade: structured data such as overviews and status is stored in a local database, while complete curves are stored independently in compressed shards, supporting random access and large-scale retrieval; the storage format carries a version identifier for smooth future upgrades

Compatibility Notes

The upgrade does not affect existing knowledge-base data or saved material configurations, and historical calculation records remain viewable; existing historical data is automatically migrated to the new storage structure on first use
Calculation results are saved automatically by default, and long-running large-batch calculations consume considerable disk space; it is recommended to reserve sufficient free space for the calculation data storage location, and the data directory can be set to a larger disk as needed
The calculation data storage location is set during installation; upgrade installations keep the existing data path

System Requirements

Operating system: Windows 10+ (64-bit)
Memory: 4 GB or above
Storage: about 500 MB for software installation; the space required for the calculation data storage location grows with calculation scale, and tens of GB or more of free space is recommended for large-scale high-throughput calculations
JMatPro: must be installed with the corresponding license configured

Additional Version Notes

The following content comes from the previous V2.2.1 and V2.2.2 releases and is included in this version.

V2.2.1

Added CSV export of calculation results: both the calculation workspace and the knowledge base can export results to CSV, split by result type and packaged into a single ZIP file for download; the existing JSON export remains compatible. V2.3.0 has further upgraded this capability to backend streaming export (see "Result Data Export").

V2.2.2

Fixed an issue in the "high-temperature strength" calculation for Ni-based superalloys where the γ′ precipitate size was not correctly taken into account. This issue caused the contribution of γ′ to yield strength to be ignored, resulting in lower calculated high-temperature yield strength; after the fix, high-temperature strength results correspond correctly to the configured γ′ precipitate size.