Intro  
  Common LISP  
     LISP benchmarks  
  Birthdays in AWK  
  IPC Theory  


  Lithuanian version Lithuanian


The benchmarks of some Common LISP implementations

1. Preface

   PRODATA recently had performed benchmarking for a number of different Common LISP implementations. You can see the results below.    Most benchmarks are based on the well-known Gabriel benchmarks. Some additional benchmarks were taken from the CMU CL source code tree (src/benchmarks), which include Richard's benchmark (an operating system simulation) and modified version of cascor1.lisp (cascade correlation learning algorithm). You can find this modified version here.

   All benchmarks have been performed using following hardware, having all three OS'es (Windows, Linux and FreeBSD) installed with multiboot.

  • Model - HP Omnibook XE3 laptop
  • CPU - Intel Celeron (Coppermine) 650 MHz
  • RAM - 128 MB
  • ACPI - disabled
  • HDD - internal IBM - 10 GB
  • Video - S3 Savage IX / TFT panel (isn't very important for tests)

   Some notes and conclusions about these benchmarking are placed at the end of the page. Please see the results:

2. Benchmarking results

Linux 2.6.3   (interpreted / compiled)
BenchmarkAllegro CL
6.2
CMU CL
18e
CLISP
2.30
GCL
2.5.3
triangle487.240
3.160
310.582
3.301
105.324
11.393
92.370
0.920
boyer42.080
0.240
19.374
0.079
2.925
0.536
11.730
0.150
browse54.990
0.160
22.996
0.194
3.875
0.768
6.400
1.080
ctak37.550
0.140
18.455
0.057
5.522
0.501
5.540
0.150
dderiv5.340
0.038
2.230
0.036
0.808
0.229
0.640
0.080
deriv4.340
0.032
1.993
0.107
0.596
0.256
0.510
0.080
destructive3.330
0.020
3.262
0.011
1.791
0.074
0.680
0.030
iterative-div22.660
0.011
2.667
0.014
0.606
0.145
0.400
0.040
recursive-div27.770
0.017
2.354
0.018
0.600
0.156
0.610
0.050
fft24.500
0.471
6.668
0.013
2.269
0.426
1.600
0.300
frpoly-fixnum35.650
0.042
4.759
0.017
0.910
0.135
3.430
0.080
frpoly-bignum36.000
0.120
5.065
0.077
1.112
0.317
3.510
0.240
frpoly-float35.770
0.056
4.819
0.028
0.932
0.234
3.430
0.040
puzzle15.240
0.317
19.408
0.027
4.275
0.637
3.470
0.030
puzzle
(C-code)
0.011
tak34.940
0.033
13.156
0.028
4.345
0.332
5.110
0.040
rtak34.460
0.032
13.124
0.027
4.396
0.332
5.100
0.050
takl252.430
0.285
109.679
0.151
16.676
1.745
31.310
0.120
takr3.620
0.007
2.336
0.005
0.411
0.039
0.560
0.010
stak45.870
0.470
23.035
0.089
12.393
0.616
4.770
0.090
init-traverse25.760
0.073
28.061
0.026
15.363
0.558
6.110
0.250
run-traverse201.170
0.706
110.582
0.508
43.565
3.413
32.140
0.610
richards.lisp59.940
0.295
29.551
0.021
12.184
0.873
9.920
0.160
cascor1.lisp
sec./epoch (*)
29.721
0.0478
1.401
0.0215
2.005
0.1843
3.236
0.0310
Total score:
(sum of places)
91
52
68
30
35
90
36
58
Final ranking: 4
2
3
1
1
4
2
3


FreeBSD 5.2.1   (interpreted / compiled)
BenchmarkAllegro CL
6.2
CMU CL
18e
CLISP
2.30
GCL
2.5.3
triangle513.523
3.039
297.250
3.286
99.696
10.450
207.690
1.050
boyer51.266
0.171
17.975
0.056
2.982
0.688
16.520
0.290
browse58.593
0.180
23.590
0.169
3.837
0.793
14.120
6.590
ctak40.133
0.141
17.346
0.054
5.300
0.382
13.100
0.430
dderiv5.797
0.023
2.177
0.027
0.743
0.233
1.870
0.200
deriv4.602
0.023
1.937
0.026
0.618
0.227
1.590
0.120
destructive3.320
0.007
3.212
0.010
1.762
0.055
1.830
0.160
iterative-div22.771
0.008
2.607
0.014
0.618
0.107
1.160
0.020
recursive-div26.172
0.016
2.315
0.018
0.618
0.116
1.640
0.100
fft24.109
0.476
6.628
0.012
2.161
0.420
2.790
0.590
frpoly-fixnum37.352
0.023
4.348
0.003
0.933
0.211
6.830
0.190
frpoly-bignum37.742
0.110
4.763
0.072
1.136
0.341
7.230
0.740
frpoly-float37.508
0.047
4.563
0.025
0.951
0.222
6.830
0.040
puzzle15.461
0.289
18.029
0.023
4.335
0.564
8.460
0.030
puzzle
(C-code)
0.017
tak37.218
0.023
12.872
0.027
4.314
0.245
11.520
0.050
rtak37.469
0.023
12.982
0.027
4.243
0.245
11.370
0.050
takl273.594
0.234
106.377
0.146
16.537
1.235
64.630
0.140
takr3.898
0.008
2.391
0.005
0.483
0.032
1.170
0.010
stak48.422
0.484
22.102
0.087
11.742
0.598
11.740
0.100
init-traverse25.742
0.063
27.579
0.027
14.589
0.418
14.580
0.850
run-traverse214.172
0.680
105.640
0.483
40.482
2.642
69.290
0.770
richards.lisp59.969
0.273
27.519
0.020
11.064
0.759
27.560
0.450
cascor1.lisp
sec./epoch (*)
30.055
0.0504
1.184
0.0238
1.561
0.1744
10.933
0.0883
Total score:
(sum of places)
91
44
64
33
26
85
49
68
Final ranking: 4
2
3
1
1
4
2
3


MS Windows 98 SE   (interpreted / compiled)
BenchmarkAllegro CL
6.2
LispWorks
4.3
CLISP
2.30 (**)
Corman Lisp
2.5 (***)
triangle504.820
3.080
111.720
1.650
948.020
81.620
 
6.643
boyer43.770
0.440
4.830
0.980
34.320
5.490
 
0.135
browse56.740
0.220
3.960
0.550
23.730
2.360
 
0.832
ctak38.620
0.160
3.510
0.060
43.500
3.630
 
0.851
dderiv5.600
0.060
0.660
0.170
3.730
1.260
 
0.080
deriv4.450
0.010
0.550
0.160
2.530
1.040
 
0.063
destructive 3.290
0.050
1.100
0.060
10.710
0.280
 
0.059
iterative-div22.750
0.010
0.600
0.050
2.740
0.390
 
0.014
recursive-div27.300
0.060
0.490
0.110
5.330
1.700
 
0.040
fft24.610
0.490
2.530
0.550
8.900
2.300
 
1.040
frpoly-fixnum36.090
0.060
0.990
0.060
5.600
0.880
 
0.041
frpoly-bignum36.690
0.110
1.210
0.160
6.310
1.380
 
0.206
frpoly-float36.360
0.050
1.040
0.060
5.660
1.050
 
0.081
puzzle15.330
0.330
4.010
0.220
49.490
3.350
 
0.459
puzzle
(C-code)
not caclulated
tak35.920
0.010
3.180
0.010
32.630
2.420
 
0.040
rtak35.970
0.060
3.190
0.060
33.610
2.970
 
0.041
takl258.980
0.220
18.730
0.210
116.060
6.640
 
0.188
takr3.730
0.010
0.440
0.060
3.290
0.550
 
0.008
stak 46.520
0.440
5.600
0.110
131.880
4.830
 
Failed (***)
init-traverse 25.050
0.110
12.580
0.160
101.000
2.740
 
0.238
run-traverse206.250
0.660
32.460
1.210
511.910
15.370
 
1.097
richards.lisp59.590
0.160
5.050
0.110
74.590
3.730
 
0.593
cascor1.lisp
sec./epoch (*)
24.752
0.0596
2.026
0.0501
14.824
0.4821
 
0.1788
Total score:
(sum of places)
61
37
23
49
54
91
-
53
Final ranking: 3
1
1
2
2
4
-
3


3. Notes on benchmarking process

   All benchmarks where performed in a 'quiet' environment, which means that almost all daemons on Linux and FreeBSD had been stopped and no graphical environment was running. Also the amount of physical memory was big enough for running LISP processes without page faults and swapping.

   All open-source Lisp distributions (all, except Allegro CL) either had been found preinstalled, or installed from legal source (RPM's and 'ports') in both Linux and FreeBSD distributions. The version numbers of all LISPs on both platforms are chosen the same. Due to strange GCL test results on FreeBSD (much slower, than on Linux) it had been recompiled from the sources, but nothing significantly had changed.

   (*) The cascor1.lisp benchmark was finishing every time with different 'epoch' counts on various LISPs, so the benchmark result was calculated as a mean value for one epoch. Even more, it has to be pointed out, every latest epoch calculation took slightly longer time to complete, so this benchmark has not to be treated as very strict. Also, this benchmark was slightly modified in order to run on all LISPs by working around with floating point arithmetic underflows and setting up all arithmetic strictly to float type (slowing benchmark results by this, though).

   (**) The Windows version of CLISP was extremely slow. We think that it is because CLISP uses the virtual machine concept. The other possible reason - it runs in MSDOS window, which is scheduling rather slowly under MS Windows.

   (***) The Corman Lisp 'stak' benchmark took about 5 minutes and then hanged. There is nothing we can tell more, except that there was intensive garbage collecting during the test. We can suppose, that this probably had happened because of unregistered version being used.
Also, the Corman Lisp benchmarks were performed in compiled form only, because all LISP source files were compiled on-the-fly by it, - so there was no such possibility to test non-compiled interpretation.

   Each benchmark was run in non-compiled (interpreted) and compiled variants, because we also like to inspect the LISP behaviour as a dynamic language (even when the code is compiled, you still can use the constructs like (eval) and lambda expressions, which are dynamically interpreted). These test results show, that not all LISP implementations are very good in that.

   Very important note: These benchmarks are NOT Linux vs. FreeBSD performance tests. They are only cross-LISPs tests. In this particular case Linux performed some better. For those who want to know where FreeBSD performs better - please read more about FreeBSD and the cases when FreeBSD really outperforms Linux. It's only to say here - one of them is when memory is tight, and/or load is high, and this is not the case we had here when we've tested LISPs.

4. Conclusion

   Despite some important and unique features of other LISP inplementations we can state here, that the clear winner between LISPs on UNIX platforms is CMU CL. Even its speed of interpretation is very close to winners in this category. But CMU CL has one serious disadvantage - it has no MS Windows implementation, nor it has a good graphics support (there are GTK bindings for Linux, though). So if we chose to write a cross-platform graphical application, the Allegro CL with its Common Graphic Interface would be a better alternative. CLISP and GCL have their own benefits in some cases, especially for writing applications, which are designed to run on UNIX-like platforms only.

   Disclaimer: We are aware, that there are lies, damn lies, statistics and benchmarks. The benchmarks published here, are proposed for your own use (and risk) only.

   Comments and corrections are welcome. Please visit the PRODATA contacts page for contact information.



Powered by PROGRESS

| ProData | PROGRESS | Linux | Programming | TROLIS.NET |

Last updated: 2014-07-30