mersenneforum.org  

Go Back   mersenneforum.org > Great Internet Mersenne Prime Search > Hardware

Reply
 
Thread Tools
Old 2003-11-14, 07:00   #1
GP2
 
GP2's Avatar
 
Sep 2003

32×7×41 Posts
Default Graph over time of what kind of machines run GIMPS (PrimeNet)

Click for graph

Thanks to Nick Glover and others for supplying old historical versions of summary.txt (I also grabbed a few from the Wayback Machine.

There's some obvious gaps in the graphs, so I'm always looking for old versions of summary.txt, status.txt and cleared.txt that anyone might have archived.
GP2 is offline   Reply With Quote
Old 2003-11-14, 15:22   #2
dsouza123
 
dsouza123's Avatar
 
Sep 2002

2×331 Posts
Default

A summary.txt file from 9/24/2002

The status.txt and cleared.txt files are too big to attach even when zipped.
Attached Files
File Type: zip summary.zip (3.5 KB, 147 views)
dsouza123 is offline   Reply With Quote
Old 2003-11-15, 01:10   #3
outlnder
 
outlnder's Avatar
 
Aug 2002

2·3·53 Posts
Default

Since a single aacount can have 3 different types of work at any given time, is there any way to calculate, fairly closely, how many GHz or THz is working on Gimps??

I guess it would need to be calculated by user rather than machines.
outlnder is offline   Reply With Quote
Old 2003-11-15, 01:55   #4
GP2
 
GP2's Avatar
 
Sep 2003

32·7·41 Posts
Default

See this thread:
http://www.mersenneforum.org/showthr...&threadid=1295

LL testing P90 years / day

Trial factoring P90 years / day

I'm not sure if total GHz or THz is an accurate measurement since not all CPUs are equal for the same clock speed.
GP2 is offline   Reply With Quote
Old 2003-11-15, 02:27   #5
GP2
 
GP2's Avatar
 
Sep 2003

A1716 Posts
Default

We can also look at the

Total machines (of all kinds) running GIMPS, as well as the total number of separate accounts.

There's a bit of a wave phenomenon, which is slightly odd.

Clearly, retaining users (and their machines) is a bit of a problem. Perhaps a fancy stats screen and screensaver-like interface is essential to compete against SETI@Home and other projects.

Last fiddled with by GP2 on 2003-11-15 at 02:31
GP2 is offline   Reply With Quote
Old 2003-11-15, 02:42   #6
GP2
 
GP2's Avatar
 
Sep 2003

1010000101112 Posts
Default

Comparing the graphs for machine count by CPU type and machine count (all CPU types combined), we can see that the old legacy machines (K6, PII, Pentium, Pentium Pro) are affected less by the waves. Mostly P3 and Athlon are involved.

However, the wave rise that begins just after the discovery of M39 on November 14 2001 involved all CPU types. We see though that a year later, levels were back to where they were before for all CPU types, except P4, which shows a continuous rise.
GP2 is offline   Reply With Quote
Old 2003-11-15, 08:48   #7
nucleon
 
nucleon's Avatar
 
Mar 2003
Melbourne

5·103 Posts
Default

I don't think there is a problem with retaining users (Athlon/P4), taking away the cyclic nature of the graphs. The machine graphs for Athon/P4 still trend up.

The graphs going down - P3/P2 etc.. would probably be more due to replacing machines with more recent hardware.

I don't think GIMPs has the inherent market power as say SETI@home or some of the medical DC projects around. There something deep in our psyche since the likes of the original Planet of the Apes to find ET.

So to make GIMPs compete with SETI@home is going to be tough battle :)

-- Craig
nucleon is offline   Reply With Quote
Old 2003-11-15, 20:57   #8
GP2
 
GP2's Avatar
 
Sep 2003

32×7×41 Posts
Default

A few new data points were added that make the graphs a little clearer.

The abruptness of the rise around November 2001 is now clearer. There seems little doubt that this was associtated with the publicity of the discovery of M39 at that time.

It's still not clear why we had an earlier slow rise and fall in the number of machines and accounts, with the rise starting around July/August 2000 and the decline continuing steadily until the sudden M39 jump in November 2001.
GP2 is offline   Reply With Quote
Old 2003-11-16, 19:27   #9
hbock
 
hbock's Avatar
 
Feb 2003

163 Posts
Default

Thanks for the very interesting graphs !
It seems that the ratio of TF to LL work is decreasing continuously. I guess the percentage of P4s working for LL is much higher than P4s doing TF. Perhaps this will change with the next CPU generation.

Unfortunately I've got only some printouts of the summary.txt (mainly 2002), so I 've just typed in the relevant values into a text file, which should easy to apply and help to fill the holes.
  • Date yr/d LL yr/d TF GFL LL GFL TF Accou P4 Athlon PIII PII Cel PPro Pent K6 486 Cyrix unspe total TF LL DC
    12 Feb 2002 06:00 258.057 14.015 3106 169 24343 1601 10408 13076 3954 2602 1762 2963 1557 112 1487 360 39882 10806 35526 10340
    19 Feb 2002 06:00 249.163 13.972 2999 168 24515 1716 10465 13203 3972 2599 1748 2953 1537 114 1547 354 40208 11332 35922 10271
    27 Feb 2002 06:00 269.201 16.782 3240 202 25116 1836 10614 13489 4004 2629 1730 2938 1571 110 1572 364 40857 11443 36500 10281
    05 Mar 2002 06:00 262.767 17.219 3163 207 24944 1947 10530 13453 3943 2603 1706 2921 1553 110 1585 361 40712 11324 36770 10248
    12 Mar 2002 06:00 265.727 16.556 3199 199 24022 2050 10118 13043 3840 2433 1673 2905 1506 113 1539 362 39682 11094 35978 10016
    20 Mar 2002 06:00 275.738 16.767 3319 202 23477 2181 9927 12812 3741 2467 1644 2856 1465 111 1548 365 39117 10856 36333 9900
    26 Mar 2002 06:00 264.162 15.317 3180 184 23559 2290 9914 12893 3712 2460 1528 2849 1459 111 1541 394 39151 10571 36888 9961
    09 Apr 2002 05:00 278.599 15.372 3354 185 22816 2445 9664 12651 3562 2401 1480 2789 1413 109 1468 358 38340 10369 36139 9786
    22 Apr 2002 05:00 292.594 13.811 3522 166 22083 2584 9398 12295 3490 2327 1427 2726 1350 99 1414 328 37467 9791 35368 9651
    30 Apr 2002 09:00 293.615 13.728 3534 165 21660 2612 9222 11977 3404 2296 1443 2692 1302 100 1403 319 36770 9583 34878 9477
    14 May 2002 06:00 303.707 14.312 3656 172 21233 2712 9148 11703 3367 2222 1419 2588 1258 107 1377 287 36188 9520 34534 9272
    29 May 2002 07:00 287.294 14.494 3458 174 20134 2709 8668 11293 3256 2132 1387 2469 1186 100 1305 274 34779 9206 32586 8950
    18 Jun 2002 06:00 295.557 13.656 3558 164 19315 2836 8350 10874 3127 2059 1425 2351 1123 94 1258 276 33773 9032 32228 8732
    18 Jul 2002 06:00 296.565 12.669 3570 152 18401 3077 8007 10410 2909 1916 1371 2136 1047 98 1261 272 32504 8546 31276 8629
    12 Sep 2002 07:00 340.251 14.144 4096 170 17342 3763 7594 9402 2519 1804 1184 1908 901 87 1178 266 30606 7976 29923 8655
    20 Oct 2002 07:00 372.285 13.296 4481 160 17545 4456 7915 9393 2440 1877 914 1846 874 72 1144 256 31187 9994 31800 8816
    15 Jan 2003 05:00 418.925 17.753 5043 214 17791 5324 9204 8929 2185 2341 784 1651 825 57 849 257 32406 9645 31340 8536
hbock is offline   Reply With Quote
Old 2003-11-17, 00:19   #10
jeff8765
 
jeff8765's Avatar
 
Aug 2002
A Dyson Sphere

32×7 Posts
Default

Looking at these machine graphs made me wonder, if it is possible to distinguish between the Athlons and the Athon 64s. This is important because the 64s use SSE2, or is it impossible to tell because the server can no longer be updated.
jeff8765 is offline   Reply With Quote
Old 2003-11-17, 00:54   #11
GP2
 
GP2's Avatar
 
Sep 2003

32×7×41 Posts
Default

The summary.txt report only mentions "AMD Athlon", nothing specifically for Athlon 64.
GP2 is offline   Reply With Quote
Reply

Thread Tools


Similar Threads
Thread Thread Starter Forum Replies Last Post
GIMPS website looks kind of weird now ixfd64 PrimeNet 1 2014-10-07 03:33
Wich kind of Gimps task's better fit this? Damian Information & Answers 19 2011-01-06 01:55
Machines Missing from Primenet Summary edron1011 PrimeNet 2 2008-11-17 19:16
Graph of leading edge of LL testing (and double-checking) over time GP2 Data 10 2003-11-17 14:55
Alienware now makes fastest GIMPS machines out there? E_tron Hardware 11 2003-09-03 03:08

All times are UTC. The time now is 19:49.

Sun Jan 24 19:49:15 UTC 2021 up 52 days, 16 hrs, 0 users, load averages: 0.90, 1.59, 2.00

Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2021, Jelsoft Enterprises Ltd.

This forum has received and complied with 0 (zero) government requests for information.

Permission is granted to copy, distribute and/or modify this document under the terms of the GNU Free Documentation License, Version 1.2 or any later version published by the Free Software Foundation.
A copy of the license is included in the FAQ.