using mmap instead of malloc

[Lorenzobattistela]

Simply replaces malloc allocation with mmap.

Performance reports after:

MacOS

MIPS diff:
Bench_cnots: +1.24%
Bench_count: +1.5%
Bench_sum_range: +1.28%
Enum_bin: +0.9%
Fuse_inc: +13.6%
Enum_lam_smart: +1.23%

X86

MIPs diff
Performances on x86 did not improve / inconclusive, e.g from -1% to +1% on the runs.

X86

MacOs individual reports

bench_cnots

• malloc:
  WORK: 117440614 interactions
  TIME: 0.9676010 seconds
  SIZE: 335544641 nodes
  PERF: 121.373 MIPS

• mmap:
  WORK: 117440614 interactions
  TIME: 0.9557490 seconds
  SIZE: 335544641 nodes
  PERF: 122.878 MIPS

bench count:

• malloc:
  WORK: 12000000004 interactions
  TIME: 1.0456890 seconds
  SIZE: 3 nodes
  PERF: 11475.687 MIPS

• mmap:
  WORK: 12000000004 interactions
  TIME: 1.0303480 seconds
  SIZE: 3 nodes
  PERF: 11646.550 MIPS

bench sum range:

• malloc:
  WORK: 600000007 interactions
  TIME: 0.1483380 seconds
  SIZE: 100000005 nodes
  PERF: 4044.817 MIPS

• mmap:
  WORK: 600000007 interactions
  TIME: 0.1464510 seconds
  SIZE: 100000005 nodes
  PERF: 4096.933 MIPS

enum bin:

• malloc:
  WORK: 390749326 interactions
  TIME: 3.1301000 seconds
  SIZE: 1742192467 nodes
  PERF: 124.836 MIPS

• mmap:
  WORK: 390749326 interactions
  TIME: 3.1006360 seconds
  SIZE: 1742192467 nodes
  PERF: 126.022 MIPS

fuse inc

• maloc:
  WORK: 11472 interactions
  TIME: 0.0002010 seconds
  SIZE: 37339 nodes
  PERF: 57.075 MIPS

• mmap:
  WORK: 11472 interactions
  TIME: 0.0001770 seconds
  SIZE: 37339 nodes
  PERF: 64.814 MIPS

enum lam smart:

• malloc
  WORK: 189534 interactions
  TIME: 0.0055890 seconds
  SIZE: 839736 nodes
  PERF: 33.912 MIPS

• mmap:
  WORK: 189534 interactions
  TIME: 0.0055210 seconds
  SIZE: 839736 nodes
  PERF: 34.330 MIPS

X86 individual reports (on 16 allocation on runtime)

bench count:

• malloc:
  ! a = 4000
  a
  WORK: 12004 interactions
  TIME: 0.0000747 seconds
  SIZE: 3 nodes
  PERF: 160.591 MIPS

• mmap:
  ! a = 4000
  a
  WORK: 12004 interactions
  TIME: 0.0000747 seconds
  SIZE: 3 nodes
  PERF: 160.705 MIPS

bench sum range

• malloc:
  ! a = 12497500
  a
  WORK: 60007 interactions
  TIME: 0.0002372 seconds
  SIZE: 10005 nodes
  PERF: 252.992 MIPS

• mmap:
  ! a = 12497500
  a
  WORK: 60007 interactions
  TIME: 0.0002528 seconds
  SIZE: 10005 nodes
  PERF: 237.388 MIPS

enum bin

(4 bit enum)

• malloc:

! a = &2{* &1{* &1{&2{* *} &2{* λb ((b 3) 3)}}}}
a

WORK: 1918 interactions
TIME: 0.0001612 seconds
SIZE: 8043 nodes
PERF: 11.901 MIPS

• mmap:
  ! a = &2{* &1{* &1{&2{* } &2{ λb ((b 3) 3)}}}}
  a
  WORK: 1918 interactions
  TIME: 0.0001714 seconds
  SIZE: 8043 nodes
  PERF: 11.192 MIPS

enum lam smart

(with nested lambdas depth=3)

! a = &1{&1{&1{* &1{&1{* &1{* &2{* &2{* &2{&2{* &2{&2{* &2{* &2{* *}}} &4{* &4{* &4{* *}}}}} *}}}}} &1{&1{* &1{&1{* &1{&1{* &1{* *}} &1{* *}}} &2{* &2{&2{* &2{* *}} &2{* *}}}}} *}}} &1{&1{* &1{&1{&1{* &1{&1{* &1{* *}} &1{* *}}} &1{* *}} &2{&2{* &2{&2{* &2{* *}} &2{* *}}} &2{* *}}}} *}} *}
a

• malloc:
  WORK: 12236 interactions
  TIME: 0.0008063 seconds
  SIZE: 60357 nodes
  PERF: 15.176 MIPS

• mmap:
  WORK: 12236 interactions
  TIME: 0.0007667 seconds
  SIZE: 60357 nodes
  PERF: 15.958 MIPS

fuse inc

• malloc:
  ! a = 1234567
  a
  WORK: 11472 interactions
  TIME: 0.0005718 seconds
  SIZE: 37339 nodes
  PERF: 20.062 MIPS

• mmap:
  WORK: 11472 interactions
  TIME: 0.0005820 seconds
  SIZE: 37339 nodes
  PERF: 19.713 MIPS