diff options
| author | Ali Saidi <Ali.Saidi@ARM.com> | 2012-06-05 01:23:08 -0400 |
|---|---|---|
| committer | Ali Saidi <Ali.Saidi@ARM.com> | 2012-06-05 01:23:08 -0400 |
| commit | 1b370431d0ac51eb54bfbf17247f935d48995a34 (patch) | |
| tree | e9b90a8f9bdf914f29a20b3c287dae5b5a16db4c /src/cpu/ozone | |
| parent | d6997777bee827c89578cf730b186991485c647c (diff) | |
| download | gem5-1b370431d0ac51eb54bfbf17247f935d48995a34.tar.xz | |
sim: Remove FastAlloc
While FastAlloc provides a small performance increase (~1.5%) over regular malloc it isn't thread safe.
After removing FastAlloc and using tcmalloc I've seen a performance increase of 12% over libc malloc
when running twolf for ARM.
Diffstat (limited to 'src/cpu/ozone')
| -rw-r--r-- | src/cpu/ozone/lw_lsq.hh | 3 |
1 files changed, 1 insertions, 2 deletions
diff --git a/src/cpu/ozone/lw_lsq.hh b/src/cpu/ozone/lw_lsq.hh index db8e53b43..855e5a97c 100644 --- a/src/cpu/ozone/lw_lsq.hh +++ b/src/cpu/ozone/lw_lsq.hh @@ -37,7 +37,6 @@ #include <queue> #include "arch/types.hh" -#include "base/fast_alloc.hh" #include "base/hashmap.hh" #include "config/the_isa.hh" #include "cpu/inst_seq.hh" @@ -301,7 +300,7 @@ class OzoneLWLSQ { }; /** Derived class to hold any sender state the LSQ needs. */ - class LSQSenderState : public Packet::SenderState, public FastAlloc + class LSQSenderState : public Packet::SenderState { public: /** Default constructor. */ |