|
2019-05-19, 22:26
|
#1 |
"Ed Hall"
Dec 2009
Adirondack Mtns
11×479 Posts |
CADO-NFS Timing Data For Many Factorizations
I'm becoming overwhelmed with the data gathered from CADO-NFS runs. I am creating this thread to provide a place to upload this data in a form I can keep updated without adding endless posts. Any additions to the data will be made in this and the following five posts. Each file will consist of all previous runs plus any new ones of the appropriate size.
The format for the files posted will be as follows: The .zip files will be based on which size params file was used and whether it was the default or a modified version, for example, params140modified.zip. Modified files are from the CADO threads in the Factoring sub-forum. Note: I might be off by a day for the changeover from default to modified. For <133 decimal digits (dd), that changeover happened around 13Apr2019 - for >138 around 11May2019. Nothing below 88 digits is modified at this time. Within the .zip files, the individual names will be based on the date/time started (for uniqueness) and the actual size of the composite, for example, 20190515132119c138. The c138 means the actual size of the composite is 138 dd. Someone referred to my many machines as a farm and I liked the term, so when reviewing the timing data, you might keep in mind that my "farm" typically has from 20-30 machines doing chores. This does have an effect on overall performance, other than the simple speeding up of a factorization. Although many machines are doing polyselect and sieving, only one is performing Linear Algebra at this time. Note: Due to the limit of five attachments per post and the absence of any modified params files below c90, all smaller runs (<88 dd) will be combined in the first attachment of this post. Update info: params30-85default - 24Jun2019 params90default - 19May2019 params90modified - 24Jun2019 params95default - 19May2019 params95modified - 24Jun2019 Last fiddled with by EdH on 2019-06-24 at 20:01 |
|
|
|
2019-05-19, 22:29
|
#2 |
|
"Ed Hall"
Dec 2009
Adirondack Mtns
526910 Posts |
Files for c100 and c105
Update info:
params100default - 19May2019 params100modified - 24Jun2019 params105default - 19May2019 params105modified - 24Jun2019 Last fiddled with by EdH on 2019-06-24 at 20:02 |
|
|
|
|
2019-05-19, 22:43
|
#3 |
|
"Ed Hall"
Dec 2009
Adirondack Mtns
11·479 Posts |
Files for c110 and c115
Update info:
params110default - 19May2019 params110modified - 24Jun2019 params115default - 19May2019 params115modified - 24Jun2019 Last fiddled with by EdH on 2019-06-24 at 20:03 |
|
|
|
|
2019-05-19, 22:44
|
#4 |
|
"Ed Hall"
Dec 2009
Adirondack Mtns
11×479 Posts |
Files for c120 and c125
Update info:
params120default - 19May2019 params120modified - 24Jun2019 params125default - 24Jun2019 Last fiddled with by EdH on 2019-06-24 at 20:04 |
|
|
|
|
2019-05-19, 22:44
|
#5 |
|
"Ed Hall"
Dec 2009
Adirondack Mtns
11×479 Posts |
Files for c130 and c135
Update info:
params130default - 24May2019 params130modified - 24Jun2019 params135default - 24Jun2019 Last fiddled with by EdH on 2019-06-24 at 20:05 |
|
|
|
|
2019-05-19, 22:44
|
#6 |
|
"Ed Hall"
Dec 2009
Adirondack Mtns
11·479 Posts |
Files for c140 and c145
Update Info:
params140default - 19May2019 params140modified - 24Jun2019 Last fiddled with by EdH on 2019-06-24 at 20:06 |
|
|
|
|
2019-05-19, 22:55
|
#7 |
|
"Ed Hall"
Dec 2009
Adirondack Mtns
11×479 Posts |
CADO-NFS Timing Comparison
I've compiled over 2100 CADO-NFS runs with default and modified params files into a single speadsheet. Do let me know if this is of any value to anyone, other than curiosity.
The file is an Excel .xlsx document that has been compressed. Here is a more readable version of the highlights (8Jun2019): Code:
CPU (seconds) Wall (seconds)
default modified default modified
88 digits
5078.94 4394.5 138.64 118.82
89 digits
0 4698.46 0 122.69
90 digits
6684.08 5383.1 144.84 122.46
91 digits
6737.56 5841.09 137.22 130.2
92 digits
0 6161.96 0 137.58
93 digits
0 6710.28 0 177.01
94 digits
9363.64 7037.12 241.56 178.14
95 digits
11464.4 7735.96 230.58 186.24
96 digits
11738.68 8134.72 248.29 196.8
97 digits
13343.6 8976.57 248.42 207.14
98 digits
15093.93 10313.52 324.01 288.47
99 digits
15890.3 11127.17 375.64 301.87
100 digits
18859.4 12080.51 378.5 320.03
101 digits
0 13337.22 0 334.03
102 digits
22752.78 14964.76 464.01 362.88
103 digits
19836.05 16892.17 579.47 419.98
104 digits
22606.8 18560.1 618.27 417.89
105 digits
25752.2 19671.76 657.46 446.22
106 digits
0 21681.29 0 468.56
107 digits
32252.9 24150.98 641.67 479.16
108 digits
31295.9 27350.1 667.01 602.86
109 digits
0 29937.09 0 621.77
110 digits
38115.8 34537.93 889.07 712.36
111 digits
39385.7 37874.05 985.47 775.71
112 digits
0 41966.6 0 893.12
113 digits
59437.2 44976.24 1206.31 1102.61
114 digits
63621.9 51938.79 1721.12 1271.64
115 digits
63498.4 57239.52 1682.17 1410.68
116 digits
0 63450.81 0 1562.44
117 digits
84944.73 71926.45 1933.54 1689.1
118 digits
101399 74250.71 2625.73 1807.62
119 digits
0 83247.14 0 1953.14
120 digits
115829.67 91275.12 2470.88 2155.61
121 digits
140780 101974.92 2998.51 2368.41
122 digits
146063 112347 2690.93 2671
123 digits
178112.29 0 4158.55 0
124 digits
192393.17 0 4313.99 0
125 digits
221596.73 0 4512.86 0
126 digits
245606.25 0 4800.09 0
127 digits
269015.14 0 5650.05 0
128 digits
292603.5 219488.22 11008.8 5150
129 digits
306136.5 246015.5 10804.3 5826.52
130 digits
374315 295734.11 9931.01 6741.06
131 digits
418847 331289.5 14039.8 7705.62
132 digits
0 389333.17 0 8401.95
133 digits
574194 0 12115.37 0
134 digits
654055.67 0 13606.73 0
135 digits
692236.22 0 15072.51 0
136 digits
804202 0 15840.32 0
137 digits
867536.25 0 18243.2 0
138 digits
709733 737290.6 16633.3 15789.62
139 digits
966460 863787 22582.6 18532.7
140 digits
0 963200 0 19386.1
141 digits
1240150 1149020 30295.2 23818.6
142 digits
1374580 0 34461.4 0
143 digits
0 0 0 0
144 digits
0 0 0 0
145 digits
0 0 0 0
146 digits
0 0 0 0
147 digits
0 0 0 0
148 digits
0 0 0 0
149 digits
0 0 0 0
150 digits
0 0 0 0
Last fiddled with by EdH on 2020-12-31 at 23:42 |
|
|
|
|
2019-05-20, 01:42
|
#8 |
"Curtis"
Feb 2005
Riverside, CA
33·11·19 Posts |
Thanks Ed!
When I get more time (after finals and once the Cunningham c207 CADO project is in production rather than testing) I'll go through this data and see if I can find parameter edits to speed any of the particular sizes. RichD is doing testing at 145/150 digit sizes; we'll have files for those in June, still editing parameters nearly every job. |
|
|
|
|
2019-05-20, 15:07
|
#9 | |
|
"Ed Hall"
Dec 2009
Adirondack Mtns
122258 Posts |
Quote:
I've made some scripts that will hopefully allow me to upload more easily and in a more timely fashion. Looking forward to the 145/150 files. I'm not really working there ATM, but might with one of my reserved Aliquot Sequences. Last fiddled with by EdH on 2019-05-20 at 15:42 |
|
|
|
|
 |
| Thread Tools | |
Similar Threads
|
||||
| Thread | Thread Starter | Forum | Replies | Last Post |
| Timing for different B1 values? | CRGreathouse | GMP-ECM | 8 | 2018-05-12 05:57 |
| mprime #7 timing summary request | aurashift | Software | 5 | 2017-11-26 21:28 |
| Timing for large candidate | carpetpool | Conjectures 'R Us | 6 | 2016-12-31 06:02 |
| Strange timing for GMP-ECM 6.2.3 | jyb | GMP-ECM | 5 | 2010-02-10 14:01 |
| Timing Options | Kevin | Software | 3 | 2002-09-12 14:03 |
