 
|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
|
|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
173.200.6.182
In Reply to: RE: Don't worry posted by steppe on December 06, 2011 at 10:29:08
on to something. Some will be jealous they did not think of it. I know I was not jealous but wondered why I had not ventured into this. You look at these files in system32 and the balloon (on a non-cMP machine) pops up and tells you what it does. In your head you know that is meaningless for the cMP application but it is easier to think, better leave well enough alone.
Thanks to you we now are looking for the new residence of "well enough" - have not found it yet though we are getting closer.
I do think I know of one who did this and seems to have come round rather strongly in favor of your initiative.
Maybe it was your playful language; your speaking of starting this journey after feeling the effects of alcohol maybe struck some as being unserious. I can assure you without the benefits of alcohol I would have hesitated doing many things that turned out well!!!
I do not think ANYONE is ridiculing you now. And if they are, luckily, they do not speak out on the forum because if they did they would meet massive resistance.
When you consider install s I have formatted the two SSDs tens of times, each. I cannot see or hear any problems. Of course, I would have to try a new one in comparison to really know.
Thanks for the link.
Follow Ups:
Rick! There was no bitterness in what I posted. It is just the funny side of it. And besides, it's not really me, it's our COMMON effort, no less. It all started with Cics, after all.
Anyway. If you have time and leasure and are willing to risk (not very much) open your registry here:
HKEY_LOCAL_MACHINE\SYSTEM\CurrentControlSet\Control\Session Manager\Memory Management
in the window to the right find NonPagedPoolSize. If it doesn't exist, just create it.
Doubleclick and set the value to 4194304 (that's in bytes. See the pick. I have just done it on my working machine I am posting from. (15-20 minutes ago) Everything seems suddenly so agile... It has to do with AWE)
Serge.
Aside from the confusion over hex vs. deci, you gave the number "4194304 (that's in bytes" which equal 4.19MB but then said you were aiming for 40MB. Likewise the larger number in parenthesis in the registry key is 68MB not 682MB, so I am at the least unsure how to proceed whichever number system. Then, of course, there is the question of what this is for or about given that the existing value for me too was 0. Are you still recommending this be tried, and what would I who has 3GB ram total put in my registry entry?
Gentlemen... please be careful to specify numbers like this as either HEX or DECIMAL values, lest we be led to wrong conclusions.
Serge - the number you quote in the body of your post ("4194304") is in hexadecimal format. It's DECIMAL (ie: base-10) equivalent is displayed in your pic to the right (in brackets, "68764420"). That is the number of bytes Windows is using for this parameter.
I mention this because Rick later used a number "halved", and it looks like he has done base-10 division on a base-16 number. The result might "work", but it's certainly not what he presumes it to be.
Just a little "gotcha" we need to avoid in these deep internal tweakings.
Hope this helps...
Grant
That's not a Toy... IT'S A TOOL !!
it more deeply. I wanted to set the nonpaged pool to 40 mb. According to You, I have set it to 687mb now ? Ok, I see. I have on my working machine that is currently working much faster and I am posting right now from it I have 2gb of memory. So, for 1 gb it should be appr 480 and for 512 - 256? No math here in my post, only a common sense, though it doesn't always work with ms programs. What can be recommended or safe numbers to set for an experiment in our cmp for 1 gb and 512 mb, do You think? What is Your opinion?
Serge.
Edits: 12/06/11 12/06/11
Hi Serge
Until I understand a little about how the NonPagedPoolQuota & Size parameters work, I have no recommendations for using them. Checking my main system here, both are set to ZERO (I don't remember setting AWE on it).
I'm not sure exactly what you're trying to achieve with these settings... how it will improve playback.
You know the math: 1024 x ?? = your mb value. My point was to be clear on exactly what we are telling Windows to do via these numbers.
Cheers,
Grant
That's not a Toy... IT'S A TOOL !!
Hi Serge
After much clicking, I came across this statement in an article by a Microsoft Critical Problem Resolution Platforms Team engineer (...with a title like that, who's to argue !):
*NonPaged pool size is not configurable other than the /3GB boot.ini switch which lowers NonPaged Pool’s maximum. 128MB with the /3GB switch, 256MB without
The very in-depth article can be found here -> http://blogs.msdn.com/b/ntdebugging/archive/2006/12/18/understanding-pool-consumption-and-event-id_3a00_--2020-or-2019.aspx
My head still hurts from trying to grasp it all... but these parameters do not get you closer to the Madman's Dream (in my opinion). But I could be wrong - it happened once before ;)
Cheers,
Grant
That's not a Toy... IT'S A TOOL !!
wondering what is the difference between the system using RAM or using the SSD, is there that much difference in speed between them?
Reading the attached article with my breathtakingly limited understanding of how WINDOWS works leads me to think this would not have anything to do with our music files in memory - I get the feeling it is concerned with OS stuff in memory.
I guess if we could find a way to limit how much of that (OS) is kept in memory would that necessarily result in more room for the music files? I am getting the feeling that is not the case.
I know the folly of coming to a conclusion after reading one article but I had spent some time chasing my tail at the MICROSOFT site before. One gets dizzy ...
I used halves since that is all I could think to do!
It seems even after dividing Serge's number by four the tracks have more segments than they did at zero.
It did not make a difference in segments (that is an easy to see difference)when I went from half to a quarter.
I was thinking that since the "full" number would not allow the track to be loaded at all that it might be working in an opposite way from what had been assumed. Then again, I was assuming what the assumption was.
I listened for a couple of hours with the quarter number and it sounded good. I cannot honestly say it sounded any better. I think my first impression was because I had the volume turned up too high!
No question we need to find out just what this does.
I thought Serge was thinking it was a way to either implement AWE or make the AWE "lock" larger. But the fact that my machine which does have AWE working also says "0" leads one to believe that this is not part of AWE,
UNLESS
I have noticed that when connecting my music computer SSD to my other computer than when I load the hive I will have entries on the music SSD's registry that have nothing to do with the "music" registry - it will show the drives in my other machine and things like that so there is not a complete divorce between the two registries.
When I change the registry on the music SSD it will not affect the other machine, so that is not a concern (unless, as I have two times now I lose track of where I am and move from the loaded hive to the "other" machines registry (very stupid)). But this could be one of those situations where the registry is reflecting the host computer's registry and not the registry's hive you have loaded. Excuse the fact that I do not know how to speak the computer vernacular. I hope this makes some sense.
This may be an instance where I am not able to truly see the setting my music computer is using when active.
Rick! I have (it seems) found the way to make awe work, so will risk these settings on cmp machine myself. The working machine works Much Faster! The fact that more segmenting occurs shows, that,
1. Cplay works as described
2. awe, even after all deletions, works
3. There is smth fishy here, that needs to be uncovered (about nonpaged pool size), because it affects the sq.Generally, there is a madman's dream in my head to achieve the following:
1. Make as much (if not all) clocks in the system work in unison as possible
2. Make awe create large pages' size so that locked pages would be few and the cpu would have to make shorter excursions to them and back and that bigger chunks of the song would fit in.
3. Exclude everything possible from the process above.
4. Create a program or a process or just a command line to constantly assign affinity to our cmp-cplay process to a DIFFERENT processor, that's why acpi multicore/multiprocessor (though it affects current stability) installation is needed (remember cics saying that even one core would enable multiple threads? Yeah, but he didn't say of affinity, and many of us tried it manually.)
AND WHEN CICS IS BACK, we'll have to try to persuade him to get back to all the cue loaded into memory, but not just into memory, but in a NonPaged Pool.
Serge.
You must understand, that I am fogwalking as I am not a computer man. I read read read read, try to make my assumptions on the level of my poor ken and try, try, try. All I ask for is more and more feedback. Your feedback is like a thread of Ariadna that must get us out of the labyrinth.
Serge.
Edits: 12/07/11
I have a similar problem with registry tweaking via a host computer. It worked fine for awhile then because of some mishap I made (exitting regedit w/o unloading hive 1st or doing what you did) I now only change my host registry and not the ssd registry with this method. This is why
I keep trying to make the Mihaylov routine work ... but unsucessfully so far.
Edits: 12/07/11
I was pointing out how one has to think about what they are seeing. Sometimes it is (on some entries) showing you what the host machine has installed. Some places it would show both the music computer AND my host computer devices connected.
When I have had a problem it was caused by my paying too little attention and deleting away like a madman!!!!
Have had to reinstall WINDOWS twice. Now to learn to make an image of the other machines OS!!!!
You can reload windows from a prior date from the windows start button on bottom left of xp screen, all programs-> system -> restore or something like that.
than I had before. (2097152)
With your number cPLAY would not load any files. Couldn't "find" my files.
Before each approx. 3 minutes was a segment. Now it is like 2 minutes.
I have halved the number again to see what happens.
So far, it certainly is not doing anything obnoxious.
Will see what 1088576 does and will let you know.
halved the number. And it works.
Haven't had a chance to listen just thought I would tell you.
How big do you think I can make the number with my half-size memory?
I will report on what I think of the sound when I can segue to EXTREME BIOS and listen for awhile.
I can tell you (who knows how influenced I am by your comments) but the music did seem to leap out of the speakers.
More to follow if you are still awake!
so I am not sure this is AWE.
Meant to say that in the first note.
writing at work so I will look as soon as I get there.
Well, after I pour some whiskey in a glass THEN I will take a look and report.
Post a Followup:
| FAQ |
Post a Message! |
Forgot Password? |
|
||||||||||||||
|
||||||||||||||
This post is made possible by the generous support of people like you and our sponsors:
Top of Page ]
[ Contact Us ]
[ Support/Wish List ]
[ Copyright © 2025, Audio Asylum, all rights reserved ]
