
==============================================================================
MERA BENCHMARK REPORT
==============================================================================
Date         : 2026-09-07
Host         : op4pgn
CPU          : INTEL(R) XEON(R) GOLD 6534 (32 threads)
RAM          : 1511.41 GB
Filesystem   : btrfs   mount /data1
Julia        : 1.12.7, 24 compute + 24 GC threads
Mera         : 1.8.0

Dataset      : L13_SN5_CD_only output 390
  ncpu       : 5120
  levelmax   : 13   (read capped at lmax=6)
  files      : 20489
  on disk    : 53.18 GB

Reading the RAMSES output (16 threads):
  hydro      :    23.67 s +- 0.23 s
  particles  :     4.63 s +- 1.70 s
  gravity    :    23.58 s +- 0.88 s
  TOTAL      :    51.88 s

Conversion (hydro, gravity, particles, 16 threads):
  read from RAMSES   :    51.08 s
  savedata write     :     3.30 s
  one-off conversion :    54.37 s
  MERA re-read warm  :  84.8 ms   (602.0x faster)
  size on disk       : 53.18 GB -> 74.7 MB   (levels up to 6 only, not comparable)
  break-even         : 1.1 re-reads
  memory to load the same data:
    allocated RAMSES :   61.89 GB
    allocated MERA   :   207.4 MB   (305.6x less churn)
    GC RAMSES / MERA : 15.03 s / 0.0 ms
    peak RSS RAMSES  :    2.02 GB
    peak RSS MERA    :    2.14 GB   (0.9x lower)

Storage by component:
  hydro        24.94 GB   46.9%
  gravity      17.52 GB   32.9%
  amr          10.66 GB   20.1%
  particle      65.0 MB    0.1%

Reading thread sweep (:hydro):
    1 threads :   128.87 s  1.00x
    2 threads :    67.35 s  1.91x
    4 threads :    38.89 s  3.31x
    8 threads :    28.92 s  4.46x
   16 threads :    23.69 s  5.44x
   24 threads :     23.2 s  5.56x
  fastest 24 threads, sweet spot 16 threads

Storage, IOPS by thread count:
    1 threads :     105455 IOPS
    2 threads :     165791 IOPS
    4 threads :     199558 IOPS
    8 threads :     244197 IOPS
   16 threads :     257663 IOPS
   24 threads :     306679 IOPS
==============================================================================
