Ticket #91 (closed bug: fixed)

Opened 3 years ago

Last modified 3 years ago

Bug in week-of-code: probable memory leak in BinaryCollapse test

Reported by: jbordner Owned by: jwise
Priority: major Milestone:
Component: enzo Version: 2.0
Keywords: Cc:

Description

Seems to allocate large amounts of memory (17GB+) for a relatively small problem (323)

Change History

  Changed 3 years ago by jbordner

  • summary changed from Bug in week-of-code: probable memory leak in `BinaryCollapse` to Bug in week-of-code: probable memory leak in BinaryCollapse test

follow-up: ↓ 3   Changed 3 years ago by jwise

After 18 outputs (interpolated level 11, cycle 52), memory usage on my Mac was reasonable at ~170MB.

in reply to: ↑ 2   Changed 3 years ago by jwise

  • owner changed from somebody to jwise

Replying to jwise:

After 18 outputs (interpolated level 11, cycle 52), memory usage on my Mac was reasonable at ~170MB.

I rescind that... I was using the WoC branch with a change that actually fixed this bug.

  Changed 3 years ago by jwise

  • status changed from new to closed
  • resolution set to fixed

The hydro_rk problem initializers didn't account for float/int casts as double/long long, i.e. FSYM, ISYM, GSYM. 1324:14218f395091 fixes this.

Note: See TracTickets for help on using tickets.