Fluent benchmark history on Pleiades cluster

Mark L. Sawley

Description

A comparison of FLUENT benchmarksPleiades – Pleiades 2 – Pleiades 2bMark L. Sawley16 January 2007Fluent benchmark casesSmall class (< 100k cells) FL5S1 FL5S2 FL5S3 Medium class (between 100k and 500k cells) FL5M1 FL5M2 FL5M3 Large class (> 500k cells) FL5L1 FL5L2 FL5L3 Fluent benchmark casesClass Benchmark Cells Mesh Models Solver Description segregated ...

Informations

• Publié par: Moshug
• Nombre de lectures: 56
• Langue: Norwegian

Extrait

A comparison of FLUENT benchmarks

Pleiades – Pleiades 2 – Pleiades 2b

Mark L. Sawley

16 January 2007

 

FL5S1  

FL5M1  

FL5L1  

Fluent benchmark cases

 

 

 

FL5S2  

FL5M2  

FL5L2  

 

 

 

FL5S3  

FL5M3  

FL5L3  

Small class ( 100k cells) <  

Medium class (between 100k and 500k cells)  

Large class ( 500k cells) >  

small   FL5S2  

FL5S3  

FL5M1  

Fluent benchmark cases

Class   Benchmark Cells   Mesh   Models Solver   Description     e   FL5S1   32,000hexahedralk se im re pli a c t i e t   d  turbulent flow in a bend   32,000 hexahedral   ke   c i o m u p p li l c e i d t     turbulent flow in a bend   89,856 hexahedra k   coupled l   e implicit   f 3 l 7 o     w in a compressor, rotor ke coal combustion in a r 155,188 tetrahedral   r 6 e s ac e t     se i g mp e l g i at it e   d  boiler, with particle  c DPM tracking   P1     242,782 han h gi y n b g ri - d n , ode   ke   se i g m r p e l g i a c t it e   d t v u a r l b v u e l p e o n r t t     flow in an engine ke  segregated combn  352,800 hexahedral   r 6 e s ac e t     implicit   veloci u t s y t   i b o u n r  i ner a   high led transonic flow around a FL5L1   847,746 hexahedral   ke   c e o x u pl p icit   fighter   large   FL5L2   3,618,080 hybrid   R k N e   G se i g m re pl g i a c t i e t   d e ar x o te u r n n d a   l a     a c e a r r o  b d o y d namics  y   FL5L3   9,792,512 hexahedral   RSM   segregated turbulent flow in a implicit   transition duct  

medium   FL5M2  

FL5M3  

Fluent benchmark cases

Machines considered

Pleiades : Intel Pentium 4 mono-processors, 2.8 GHz, 2 GB/node Fast Ethernet interconnect

Pleiades 2 : Intel Xeon mono-processors, 2.8 GHz, 4 GB/node Gigabit Ethernet interconnect

Pleiades 2b : Intel Xeon dual-processor dual-core, 2.66 GHz, 8 GB/node Gigabit Ethernet interconnect

Methodology

Fluent version 6.2.16 was used for all results For Pleiades 2b, 4 Fluent processes were run on each node (except possibly last node)

8000

6000

4000

2000

0

0

4

fl5s1 benchmark

8 12 16 Number of processors

20

M.L. Sawley 16 January 2007

pleiades (FLUENT 6.2.16) pleiades2 (FLUENT 6.2.16) pleiades2b (FLUENT 6.2.16)

12000

10000

8000

6000

4000

2000

0

0

4

fl5s2 benchmark

8 12 16 Number of processors

20

M.L. Sawley 16 January 2007

pleiades (FLUENT 6.2.16) pleiades2 (FLUENT 6.2.16) pleiades2 (FLUENT 6.2.16)

8000

7000

6000

5000

4000

3000

2000

1000

0

0

4

fl5s3 benchmark

8 12 16 Number of processors

20

M.L. Sawley 16 January 2007

pleiades (FLUENT 6.2.16) pleiades2 (FLUENT 6.2.16) pleiades2b (FLUENT 6.2.16)

3000

2500

2000

1500

1000

500

0

0

4

fl5m1 benchmark

8 12 16 Number of processors

20

M.L. Sawley 16 January 2007

pleiades (FLUENT 6.2.16) pleiades2 (FLUENT 6.2.16) pleiades2b (FLUENT 6.2.16)

7000

6000

5000

4000

3000

2000

1000

0

0

4

fl5m2 benchmark

8 12 16 Number of processors

20

M.L. Sawley 16 January 2007

pleiades (FLUENT 6.2.16) pleiades2 (FLUENT 6.2.16) pleiades2b (FLUENT 6.2.16)

1000

800

600

400

200

0

0

4

fl5m3 benchmark

8 12 16 Number of processors

20

M.L. Sawley 16 January 2007

pleiades (FLUENT 6.2.16) pleiades2 (FLUENT 6.2.16) pleiades2b (FLUENT 6.2.16)

1200

1000

800

600

400

200

0

0

4

fl5l1 benchmark

8 12 16 Number of processors

20

M.L. Sawley 16 January 2007

pleiades (FLUENT 6.2.16) pleiades2 (FLUENT 6.2.16) pleiades2b (FLUENT 6.2.16)