Hi, thanks again for fixing the 32-bit issue!
While testing on larger volumes, I noticed that memory usage can get quite high.
For example, reading a ~500GB NTFS volume can use close to ~1GB of RAM.
This could become problematic for multi-terabyte drives.
Suggestion:
Instead of eagerly loading/parsing the entire MFT into memory, a lazy/streaming
approach (reading records on demand) might significantly reduce memory usage.
This could:
- Lower peak RAM usage
- Improve scalability for large disks
Not sure how this fits with the current design, but thought it might be a useful improvement.
Happy to help test or experiment if needed.
If this approach sounds reasonable, I’d also be happy to work on a PR/MR—just let me know or assign the issue to me.
Hi, thanks again for fixing the 32-bit issue!
While testing on larger volumes, I noticed that memory usage can get quite high.
For example, reading a ~500GB NTFS volume can use close to ~1GB of RAM.
This could become problematic for multi-terabyte drives.
Suggestion:
Instead of eagerly loading/parsing the entire MFT into memory, a lazy/streaming
approach (reading records on demand) might significantly reduce memory usage.
This could:
Not sure how this fits with the current design, but thought it might be a useful improvement.
Happy to help test or experiment if needed.
If this approach sounds reasonable, I’d also be happy to work on a PR/MR—just let me know or assign the issue to me.