Closed yakra closed 6 years ago
single-crawl vs multi-crawl RAM usage test on a fresh-rebooted system
Baseline
Mem: 695.5M 17.6% 3.8 G Swap: 0
Road_Inventory.dbf (115.6 MB, "both2" version)
• Before, multi: Mem: 703.8M 17.9% 3.8G Swap: 0
• Before, single: Mem: 673.5M 17.1% 3.8G Swap: 0
• After, single: Mem: 663.4M 16.8% 3.8G Swap: 0
Road_Inventory.Ytrim.dbf (264.5 MB)
• Before, multi: Mem: 711.0M 18.0% 3.8G Swap: 0
• Before, single: Mem: 665.0M 16.9% 3.8G Swap: 0
• After, single: Mem: 703.6M 17.8% 3.8G Swap: 0
Road_Inventory.Ycull.dbf (890.8 MB)
• Before, multi: Mem: 697.0M 17.7% 3.8G Swap: 0
• Before, single: Mem: 699.4M 17.7% 3.8G Swap: 0
• After, single: Mem: 702.2M 17.8% 3.8G Swap: 0
Road_Inventory.Yorig.dbf (2.6 GB)
• Before, multi: Mem: 425.3M 10.8% 3.8G Swap: 326.3M 8.0% 4.0G
• Before, single: Mem: 446.1M 14.1% 3.8G Swap: 247.8M 6.1% 4.0G
• After, single: Mem: 704.2M 17.9% 3.8G Swap: 0
The figures above don't tell the full story, particularly how for Yorig, memory usage steadily climbed to almost 100% during the Before tests, then stayed practically flat in the After tests. With the exception of "Before, multi", the amount of memory (mem+swap) used is about the same across the board. Alt+Tabbing to Pluma after completion seems to immediately add ~15 MB, so there's some more noise added to the usual fluctuations...
In any case: Fixed in eb65347. Further deletions of unused variables can get a separate issue. Or not.
15b14c14e39088ae5f395b7508a8d8c2d6f8b52e also relevant
• delete unused values that don't get assigned to MaxVal • check RAM usage. Fixed in eb65347b6c90d2b956b56518753fd0e74d6ea0d8