search

van-smith / OPBM

Open Productivity Benchmark
1 stars 0 forks source link

Reboot and settle down atoms should behave like other atoms #95

Closed van-smith closed 13 years ago

van-smith commented 13 years ago

Running the reboot and settle down atoms do not do anything when clicked.

Running the reboot atom should reboot the system and record the reboot time.

I think the best solution for the second issue is to combine the settle down atom with the reboot atom where it will return the reboot time and the settle down time.

Therefore, the two atoms should be changed:

  1. Reboot time: reboot the system and record reboot time.
  2. Reboot and settle down time: reboot the system and record reboot time and settle down time.

These will likely become two of the most important atoms in OPBM.

ghost commented 13 years ago

There are aspects of the OPBM harness design and the way it handles and processes data internally which need to be discussed. There's a misunderstanding of how things operate within OPBM, resulting in this issue report.

van-smith commented 13 years ago

Rick, this is the way I want OPBM to work.

I am open to suggestions, but this seems to be the most logical way to approach these needs.

ghost commented 13 years ago

How did you measure reboot time at Centaur? Is there a system call? Or a
Windows log to examine? What generated the number?

Cossatot Analytics Laboratories

-----Original message----- From: Van Smith
reply@reply.github.com To: "Rick C. Hodgin" rick@canalabs.com Sent: Thu, Oct 13, 2011 20:44:54 GMT+00:00 Subject: Re: [OPBM] Reboot and settle down atoms should behave like other
atoms (#95)

Rick, this is the way I want it OPBM to work.

I am open to suggestions, but this seems to be the most logical way to
approach these needs.

Reply to this email directly or view it on GitHub: https://github.com/van-smith/OPBM/issues/95#issuecomment-2399916

van-smith commented 13 years ago

It was a while ago when I did this, so I don't remember the details (and the needs for doing this were not always the same), but it involved writing out the time to a log file and as soon as the program woke up on reboot, comparing the last time in the log file with the current time.

The program was invoked at startup through either the runonce registry key or by putting an appropriate link in the Startup folder.

ghost commented 13 years ago

The reboot, record reboot time, and record settle down time are each separate operations. They can be employed as molecules with the required parts (reboot + record reboot time for one, or reboot + record settle down time for another) which do that which you desire. In the course of an official run, there will be one reboot, and it will record settle down time and reboot time in separate steps.

van-smith commented 13 years ago

Until we have two atoms explicitly what I requested, this is an open bug.

ghost commented 13 years ago

This is working now with two new atoms: Standalone Reboot Time and Standalone Reboot Settle Down. The fix that's applied uses the current timing, which will be revamped today.

Revamped timing: The new reboot timing system will continuously on shutdown record the time every second to an xml file called preboot.xml, which is the last time before the OS closed processes and forced the reboot. Whenever the harness executes a reboot, it will create a new RunOnce registry entry called "opbmpostboot" which will execute postboot.exe, a new executable that will record the earliest startup time to postboot.xml, a file which will then be read as the actual startup time, used for computing reboot time, and as the baseline for the startup settle down time, rather than the current method of using the restart time of the benchmark within the harness.

ghost commented 13 years ago

Have created the preboot.exe and postboot.exe projects, which take a command line parameter as to their output filename, and write an xml structure to later be parsed by OPBM's harness.

ghost commented 13 years ago

Executing preboot during reboot sequence.

kathy-smith commented 13 years ago

I ran the Standalone Reboot Time atom through the GUI. It executed. The results viewer showed no values. Infinite mark and ? in the results viewer. I ran the Standalone Reboot Stettle Down atom through the GUI. It executed. The results viewer displayed values.

ghost commented 13 years ago

Have you upgraded to the latest version? There was an error running the preboot.exe file due to a space in the path name on the machine you're running. It was corrected in a fix this morning.

kathy-smith commented 13 years ago

Van pulled the latest version for me about an hour ago.

I documented what I ran. Didn't/don't know if the results are normal or
not. On Sat, 29 Oct 2011 16:45:20 -0500, Rick C. Hodgin
reply@reply.github.com
wrote:

Have you upgraded to the latest version? There was an error running the
preboot.exe file due to a space in the path name on the machine you're
running. It was corrected in a fix this morning.

Using Opera's revolutionary email client: http://www.opera.com/mail/

van-smith commented 13 years ago

Behavior is less than ideal. For instance, there should not be different forms of these atoms.

However, given the time constraints, I am closing this issue.