wiki:XmlStats

Project statistics data

Projects export statistics data in XML-format files. Most projects regenerate the files every 24 hours. These files are contained in a directory DOWNLOAD_URL/stats/. For most projects, DOWNLOAD_URL is the project's master URL. However, some projects (notably IBM WCG) put them elsewhere; in these cases, the URL is given by the web_rpc_url_base element returned by the project's get_project_config.php web RPC.

The download directory contains the following files:

tables.xml
This gives the total number of teams, users, and hosts. It also includes the UNIX time when the files were last generated, a list of the project's applications, and the list of badges defined by the project. For example:
<tables>
    <update_time>1046220857</update_time>
    [ <credit_by_app/> ]
    <nusers>127</nusers>
    <nteams>14</nteams>
    <nhosts>681</nhosts>
    <total_credit>1234.234</total_credit>
    <applications>
        <application>
            <name>setiathome</name>
            <results_unsent>100</results_unsent>
            <results_in_progress>1000</results_in_progress>
            <results_over>10000</results_over>
        </application>
        ...
    </applications>
    <badges>
       <badge>
           <id>10</id>
           <name>user_pct_2</name>
           <title>Top 25% in average credit</name>
           <image_url>img/pct_25.png</image_url>
       </badge>
	 ...
    </badges>
</tables>
host.gz
List of hosts
team.gz
List of teams
user.gz
List of users
badge_user.gz
list of user badge assignments
badge_team.gz
list of team badge assignments

If the <credit_by_app/> flag is present in tables.xml, there will also be

user_work.gz
user credit, broken down by app
team_work.gz
team credit, broken down by app

The format of the various XML elements in the output files is as follows.

Notes:

  • <cpid> ('cross-project identifier') is a unique identifier across multiple projects. Accounts with the same email address on different projects will have the same cross-project identifier (as long as at least one computer is attached to both accounts).
  • All 'expavg_credit' values were computed at some point in the past (given by 'expavg_time'). To compute their current values, they must be scaled according to the RAC decay formula.
  • Text fields may contain XML entities. If you aren't using a real XML parser, you need to decode the entities yourself.

Team summary

<team>
 <id>2</id>
 <type>1</type>
 <name>DistributedComputing@P&#197;&#130;ock</name>
 <userid>9</userid>
 <total_credit>52263.564346</total_credit>
 <expavg_credit>0.120263</expavg_credit>
 <expavg_time>1218301227.351580</expavg_time>
 <founder_name>Piotr Zalewski</founder_name>
 <create_time>1094061560</create_time>
 [ <url>www.example.com</url> ]
 [ <name_html>&lt;b>DistributedComputing@P&#197;&#130;ock&lt;/b></name_html> ]
 [ <description>We are crunching...&#13;&#10;to make LHC better :)</description> ]
 [ <country>Poland</country> ]
</team>

User summary

<user>
 <id>12</id>
 <name>John Keck</name>
 <country>Uganda</country>
 <create_time>1178797286</create_time>
 <total_credit>42698.813543</total_credit>
 <expavg_credit>117348.653646</expavg_credit>
 <expavg_time>1087542007.701900</expavg_time>
 <cpid>283472938743489759837498347</cpid>
 [ <url>ww.boinc.us</url> ]
 [ <teamid>5</teamid> ]
 [ <has_profile/> ]
</user>

Host summary

<host>
  <id>102</id>
  <userid>3</userid>
  <total_credit>0.000000</total_credit>
  <expavg_credit>0.000000</expavg_credit>
  <expavg_time>1087542007.701900</expavg_time>
  <p_vendor>GenuineIntel</p_vendor>
  <p_model>Pentium</p_model>
  <os_name>Windows XP</os_name>
  <os_version>5.1</os_version>
  <boinc_version>[BOINC|7.4.12|GridRepublic]</boinc_version>
  <vbox_version>[vbox|4.3.14]</vbox_version>
  <coprocs>[CUDA|GeForce 9500 GT|1|256MB]</coprocs>
  <create_time>1040170006</create_time>
  <rpc_time>1217799877</rpc_time>
  <timezone>28800</timezone>
  <ncpus>2</ncpus>
  <p_fpops>45724737.082762</p_fpops>
  <p_iops>43233895.373973</p_iops>
  <p_membw>4032258.064516</p_membw>
  <m_nbytes>670478336.000000</m_nbytes>
  <m_cache>1000000.000000</m_cache>
  <m_swap>1638260736.000000</m_swap>
  <d_total>9088008192.000000</d_total>
  <d_free>3788505088.000000</d_free>
  <n_bwup>24109.794088</n_bwup>
  <n_bwdown>57037.049858</n_bwdown>
  <avg_turnaround>465609.562145</avg_turnaround>
  <credit_per_cpu_sec>0.000526</credit_per_cpu_sec>
  <host_cpid>e129b5fa44ed8ba58e41c472822f2807</host_cpid>
</host>

The following fields are of the form [x1|x2|...|xn]. Depending on the client version, only an initial subset of items may be present.

boinc_version
The BOINC client version currently running on the host. The third item is the client "brand", such as "HTC Power to Give" or "Samsung Power Sleep".
vbox_version
The version of VirtualBox installed on the client host.
coprocs
A list of GPU descriptors. The items in each one are:
  • The manufacturer. "CUDA" means NVIDIA, "CAL" means AMD, and "INTEL" means Intel.
  • The model name.
  • The number of GPUs.
  • The amount of video RAM.
  • The driver version.
  • The OpenCL version; 102 means 1.2.
Last modified 10 years ago Last modified on Sep 11, 2014, 5:53:17 PM