Page 1 of 1

1.x Time Passed in mxcl viewer inaccurate

Posted: Tue May 02, 2006 9:50 pm
by Frances
The Time Passed amount doesn't correspond with the Time Left amount. Time Left appears to be correct. Time Passed is about half the amount that should be displayed.

This doesn't count as a render speed increase. :lol:

Posted: Tue May 02, 2006 10:03 pm
by JDHill
...actually, the only truly *correct* number for comparing render speed would come from reading CPU Time for the mxcl.exe process in Task Manager, and then dividing by the number of threads used to render. Incidentally, this has shown less total render time than Maxwell's clock has, every time I've tested it. :wink:

~JD

Posted: Tue May 02, 2006 10:56 pm
by Frances
JDHill wrote:...actually, the only truly *correct* number for comparing render speed would come from reading CPU Time for the mxcl.exe process in Task Manager, and then dividing by the number of threads used to render. Incidentally, this has shown less total render time than Maxwell's clock has, every time I've tested it. :wink:

~JD
I'm not actually interested in comparing render speed. The point is that Time Passed is incorrect. What is the point of having a Time Passed readout if it doesn't measure actual time passed? It's misleading information. Either fix it or get rid of it.

Posted: Tue May 02, 2006 11:01 pm
by Fernando Tella
Frances wrote:It's misleading information. Either fix it or get rid of it.
I agree. I make plans based on render times.

Posted: Wed May 03, 2006 12:31 am
by JDHill
I agree, Frances...that's why I pointed out the CPU Time/Threads issue. It just depends what the user expects that readout to be telling them...in other words, what 'fix it' should mean. Personally, I've got enough clocks around me...I don't need to know how much actual time has passed, I'd rather know how much time has been devoted to the calculation without figuring it out manually.

~JD

Posted: Wed May 03, 2006 12:36 am
by Frances
JDHill wrote:It just depends what the user expects that readout to be telling them...in other words, what 'fix it' should mean.
Um...I'm pretty sure the user expects that readout to be telling them how much time has passed since the render started. Fixing it would mean having it display how much time has actually passed since the render started.

Not exactly a "gray area" issue here, is it?

Posted: Wed May 03, 2006 12:47 am
by aitraaz
and here i was convinced it was measuring my blood pressure level :?

Posted: Wed May 03, 2006 1:07 am
by JDHill
Frances wrote:Not exactly a "gray area" issue here, is it?
...if you say so. I consider the actual time passed to be a completely useless piece of information.

~JD

Posted: Wed May 03, 2006 1:13 am
by mverta
... I don't. :)

I agree that such a simple piece of feedback should represent what common sense dictates it does. Personally, this innacuracy has not had any real effect on my workflow... it's a bit annoying, but so is life. One thing I've asked for is a scrollable history of previous SL updates so I can make accurate predictions about the remaining time I'm going to need the render to run. That WOULD have a direct impact on my workflow.

In any case, I'm sure on the list of fixes, this will be well below SSS refinements and cooperative networking, and as JD suggests, there are other ways of at least getting to the info for now.

_Mike

Posted: Wed May 03, 2006 1:50 am
by Frances
JDHill wrote:
Frances wrote:Not exactly a "gray area" issue here, is it?
...if you say so. I consider the actual time passed to be a completely useless piece of information.

~JD
Well no need to consult Kierkegaard or Voltaire about it.

People are under the impression that Time Passed = render time and presume it to be accurate information. It is not. And Time Left goes away when you stop a render.

You know, this is supposed to be RELEASE, production-ready software. We should have nothing more crucial to bring up than nitpicky UI issues like this. That is unfortunately not the case.

Posted: Wed May 03, 2006 2:07 am
by mverta
You should tell that to Microsoft, Frances. :)

_Mike

Posted: Wed May 03, 2006 2:56 am
by JDHill
Frances wrote:Well no need to consult Kierkegaard or Voltaire about it.
OK, then...cancel the order on the new Rolex --> I've got Maxwell. I'll try to keep the relevance of an accurate real-time render clock in mind next time I run a machining simulation over 2 instances of mxcl.exe.

~JD

Posted: Wed May 03, 2006 3:13 am
by mverta
Warning: Toxic levels of sarcasm.


JD, I think she just means if there's going to be a clock, how about it just tell the accurate time or what's the point? I think this is hard to argue with. Your point (what's the difference, go look in the Task Manager) offers a solution, but doesn't actually answer the first question. I agree with you that this is awfully high on the nitpicky-meter. As I said before, this gets little more than a shrug from me and I get on with my work.

And this is coming from a guy who has satellite-adjusting clocks everywhere in his house because if even one of them is different from the others, I go Bakersfield Chimp.

_Mike

Posted: Wed May 03, 2006 3:37 am
by JDHill
mverta wrote:Warning: Toxic levels of sarcasm.
...you got me there. I generally prefer sincerity...but...you know, only human and all that jazz. :wink:

Really, I haven't misunderstood Frances point, but I assumed that we could have some discussion about it since it was posted here rather than in the bugs forum. And although it's a minor thing, I obviously do have a definite opinion this issue. I'd rather have a useful piece of information about the real rate of progress on a render, than to have another useless clock.

Now, having said that (again)...I exit. :)

~JD

Posted: Wed May 03, 2006 4:03 am
by Frances
JDHill wrote: OK, then...cancel the order on the new Rolex ...
Perhaps this will save you about $12k:

Image