home
goals
about sensor
optimization
why sensor?
who's fastest?
p10 p50 p90
sensorpx
bayes and markov
drainage radius
dca
frac conductivity
tight & fractured
capillary pressure
miscible
primary_miscible
reserves
mmp
artificial intelligence
spe3
spe8
spe10
parallel?
gridding
fd vs fe
map2excel
plot2excel
third party tools
services
publications
q & a
ethics
contact us
Dr. K. H. Coats

 

 

Industry Benchmarks

Spe1 through spe10 are problems from the SPE Comparative Solution Projects37-43*.   The formulation options shown are those found to run best in Sensor.

Machine 1 is a 2004 Dell Precision 470 workstation with 2.8 GHz Intel Xeon processors (Nocona, 1MB L2, 800 MHz FSB) running Sensor6k (32-bit) and Windows XP Professional with 2 GB RAM.

Machine 2 is a mid-2007 Dell Precision 690 workstation with 3.0 GHz Intel Xeon processors (Woodcrest 5160, 4 MB L2, 1333 MHz FSB) running Sensor64 and Windows XP Professional x64 with 16 GB RAM.

Machine 3 is a 2012 Dell 8500 Workstation with dual 3.4 GHz Intel Core I7 3770 quad core processors and 16 GB RAM.

The Sensor data and (Machine 1) output files are available in Downloads.  Our Sensor6k evaluation version will run all these cases to completion.  All CPU times are in seconds.

Case

Formulation

Blocks

Memory MB

Newtons

Steps

Solver CPU (1)

Total CPU(1)

Total CPU (2)

Total CPU (3)

spe1

Impes, black oil, CFL

300

0.810

156

145

 

0.23

0.094

0.062

spe2

Implicit, black oil

150

0.889

34

10

 

0.09

0.047

0.015

spe3

Impes, nc=9, CFL

324

1.96

106

105

0.1

0.70

0.30

0.203

spe5

Impes, nc=6, CFL

147

0.784

718

716

0.1

1.1

0.47

0.296

spe7_1a

Implicit, black oil

270

1.28

17

16

 

0.09

0.03

0.031

spe7_4a

Implicit, black oil

270

1.28

45

36

 

0.29

0.09

0.062

spe9

Implicit, black oil, NF

9000

31.7

56

39

4.9

8.2

4.4

1.45

spe10_case1

Implicit, black oil, D4

2000

6.9

835

725

22.3

29.6

12.2

5.13

spe10_case2_75

Implicit, black oil

75

.463

23

22

   

0.06

0.02

spe10_case2_2000

Implicit, black oil

2000

4.1

112

111

   

1.5

0.3

spe10_case2 30x55x85

Implicit, black oil

140158

551

225

214

2452

2634

1803

490

The following case is from Reference 1 of our Miscible page:

spe79692x3

Impes, nc=7, CFL

1000

4.06

1011

1004

1.9

13.8

5.3

3.0

 

CFL - Impes stable step logic used.  NF - Nested Factorization solver option (default is RBILU(0)). D4 - Direct solver option.

Sensor's performance advantage of course varies by problem, but is not in general a function of problem size or reservoir complexity.  Also compare using your cases or their examples.  Run time comparisons of different models should be made on the same machine.

 

* Reference numbers are those of the Sensor Manual


2000 - 2021 Coats Engineering, Inc.