Here's a comparison between different hardware and binaries which were built as described in http://verahill.blogspot.com/2012/01/debian-testing-64-wheezy-compiling.html
The simulation in question is a 100,000 step 100 ps simulation of a carbonate ion in water. Check back later for more details.
Conclusions: 1) Double precision is slower by 25-35 % than single precision. 2) On a single machine there's no gain in using mpi. 3) Sadly, it appears that intel i5-2400 X4 3.1GHz is more expensive than AMD Phenom II X6 3.1GHz for a reason.
Machines:
Be -- Phenom II X6, 8Gb RAM.
B -- Athlon X3 3.1 GHz 3 core, 4Gb RAM
Ta -- Optiplex 990 (i5 2400 3.1 GHz 4 core, 8Gb RAM). This machine was running a full gnome3/gnome-shell desktop at the same time as the tests were carried out. Take the results with a grain of salt.
Using mdrun (no mpi, single precision)
-------------------------------
Run Be (6) Ta (4) B ( 3)
-------------------------------
1 1m27s 1m1s 1m48s
2 1m28s 1m1s 1m46s
3 1m35s 1m1s 1m47s
4 1m32s 1m1s 1m47s
5 1m33s 1m1s 1m47s
Using mdrun_dd (no mpi, double precision)
-------------------------------
Run Be (6) Ta (4) B ( 3)
-------------------------------
1 1m49s 1m15s 2m25s
2 1m47s 1m15s 2m26s
3 1m51s 1m15s 2m26s
4 1m59s 1m15s 2m24s
5 1m58s 1m15s 2m26s
Using mdrun_mpi (mpi, single precision)
Machine: Be (Phenom II X6, 8Gb RAM).
(5 cores doesn't work)
---------------------------------
Cores/Run: 1 2
---------------------------------
1 4m11s 4m13s
2 2m12s 2m15s
3 1m46s 1m43s
4 1m31s 1m31s
5 ----------------
6 1m28s 1m35s
Machine: Ta (Optiplex 990 (i5 3.1 GHz 4 core, 8Gb RAM).
---------------------------------
Cores/Run: 1 2 3
--------------------------------
1 3m20s 3m20s 3m20s
2 1m39s 1m38s 1m40s
3 1m12s 1m13s 1m12s
4 1m01s 1m01s 1m00s
Machine: Athlon X3 3.1 GHz 3 core, 4Gb RAM.
---------------------------------
Cores/Run: 1 2 3
--------------------------------
1 4m32s 4m33s 4m36s
2 2m28s 2m28s 2m27s
3 1m49s 1m50s 1m49s
Using mdrun_ddmpi (mpi, double precision):
Machine: Phenom II X6, 8Gb RAM.
---------------------------------
Cores/Run: 1 2
---------------------------------
1 5m23s 5m25s
2 2m56s 2m54s
3 2m11s 2m11s
4 1m56s 1m57s
5 -----------------
6 1m51s 1m52s
Machine: Optiplex 990 (i5 3.1 GHz 4 core, 8Gb RAM).
---------------------------------
Cores/Run: 1 2 3
--------------------------------
1 4m14s 4m13s 4m13s
2 2m09s 2m09s 2m10s
3 1m33s 1m33s 1m33s
4 1m16s 1m16s 1m16s
Machine: Athlon X3 3.1 GHz 3 core, 4Gb RAM.
---------------------------------
Cores/Run: 1 2 3
--------------------------------
1 5m01s 5m52s 5m50s
2 3m17s 3m17s 3m18s
3 2m31s 2m31s 2m31s
The simulation in question is a 100,000 step 100 ps simulation of a carbonate ion in water. Check back later for more details.
grompp -f carbonate.mdp -c carbonate.pdb -p carbonate.top -o carbonate.tpr
time mdrun -v -deffnm carbonate
Conclusions: 1) Double precision is slower by 25-35 % than single precision. 2) On a single machine there's no gain in using mpi. 3) Sadly, it appears that intel i5-2400 X4 3.1GHz is more expensive than AMD Phenom II X6 3.1GHz for a reason.
Machines:
Be -- Phenom II X6, 8Gb RAM.
B -- Athlon X3 3.1 GHz 3 core, 4Gb RAM
Ta -- Optiplex 990 (i5 2400 3.1 GHz 4 core, 8Gb RAM). This machine was running a full gnome3/gnome-shell desktop at the same time as the tests were carried out. Take the results with a grain of salt.
Using mdrun (no mpi, single precision)
-------------------------------
Run Be (6) Ta (4) B ( 3)
-------------------------------
1 1m27s 1m1s 1m48s
2 1m28s 1m1s 1m46s
3 1m35s 1m1s 1m47s
4 1m32s 1m1s 1m47s
5 1m33s 1m1s 1m47s
Using mdrun_dd (no mpi, double precision)
-------------------------------
Run Be (6) Ta (4) B ( 3)
-------------------------------
1 1m49s 1m15s 2m25s
2 1m47s 1m15s 2m26s
3 1m51s 1m15s 2m26s
4 1m59s 1m15s 2m24s
5 1m58s 1m15s 2m26s
Using mdrun_mpi (mpi, single precision)
Machine: Be (Phenom II X6, 8Gb RAM).
(5 cores doesn't work)
---------------------------------
Cores/Run: 1 2
---------------------------------
1 4m11s 4m13s
2 2m12s 2m15s
3 1m46s 1m43s
4 1m31s 1m31s
5 ----------------
6 1m28s 1m35s
Machine: Ta (Optiplex 990 (i5 3.1 GHz 4 core, 8Gb RAM).
---------------------------------
Cores/Run: 1 2 3
--------------------------------
1 3m20s 3m20s 3m20s
2 1m39s 1m38s 1m40s
3 1m12s 1m13s 1m12s
4 1m01s 1m01s 1m00s
Machine: Athlon X3 3.1 GHz 3 core, 4Gb RAM.
---------------------------------
Cores/Run: 1 2 3
--------------------------------
1 4m32s 4m33s 4m36s
2 2m28s 2m28s 2m27s
3 1m49s 1m50s 1m49s
Using mdrun_ddmpi (mpi, double precision):
Machine: Phenom II X6, 8Gb RAM.
---------------------------------
Cores/Run: 1 2
---------------------------------
1 5m23s 5m25s
2 2m56s 2m54s
3 2m11s 2m11s
4 1m56s 1m57s
5 -----------------
6 1m51s 1m52s
Machine: Optiplex 990 (i5 3.1 GHz 4 core, 8Gb RAM).
---------------------------------
Cores/Run: 1 2 3
--------------------------------
1 4m14s 4m13s 4m13s
2 2m09s 2m09s 2m10s
3 1m33s 1m33s 1m33s
4 1m16s 1m16s 1m16s
Machine: Athlon X3 3.1 GHz 3 core, 4Gb RAM.
---------------------------------
Cores/Run: 1 2 3
--------------------------------
1 5m01s 5m52s 5m50s
2 3m17s 3m17s 3m18s
3 2m31s 2m31s 2m31s
This is dummy text. It is not meant to be read. Accordingly, it is difficult to figure out when to end it. But then, this is dummy text. It is not meant to be read. Period.
ConversionConversion EmoticonEmoticon