RAMDisk

[manisandher]

As I understand, before playing XX will first perform conversion to WAV (if needed) and then load resulting file into a pre-locked RAM - Only then, it starts playing i.e. reading music bits from there.

Obviously Peter is the guy to clear this up. But my understanding is that this is exactly where the SFS comes into play - XX will load into RAM as much as is dictated by the SFS. If this covers a whole track, then that's what will be loaded. If it doesn't, then XX will transfer data to RAM when it needs to, during playback. This will be the case whether 'Copy to XX-drive' is selected or not.

But there's still a very curious thing going on here. Take a small, say, 15MB track and don't apply any additional processing to it in XX (e.g. upsampling). Now, you would think that there'd be no difference in SQ if the SFS was set to, say, 50 or 100. I haven't tried this (I'm away from home right now) but I bet the sound still changes... Also, IIRC it still takes longer to load a single 15MB when the SFS is set to 100 vs 50. Why this should be, I have no idea...

Mani.

[Josef]

Hmmm - I think I've found a method which clearly demonstrates that XX indeed behaves _very_ differently when run from RAM disk!

Experiment is simple: make a playlist with 1 track (preferably short so you don't have to wait long for results but it works with any length). I used Pink Floyd's 'Goodbye Cruel World' (The Wall) as it lasts only 45 seconds.
It does not really matter that much what settings you have but, to simplify, disable 'Copy to XX drive' and make sure Split File Size is big enough for your test track to fit in RAM in 1 piece (I just set it to 100MB). And, oh yeah, I used a WAV file as input: no FLAC conversions etc and no upsampling mumbo-jumbo: just straight 44/16.
Open Process Explorer, open XX from RAM disk and play the track in Unattended mode.

You should see something like this on XXEngine3.exe properties dialog: (Capture1.PNG)

XX has 9 threads of which only 2 are active 99% of time. This makes sense: one thread for XX and other for .Net main loop: Nice & clean.

Now try running XX from OS disk (does not matter if it's HD or SSD) and look at same dialog: (Capture2.PNG).

Ouch. Instead of 9 there are now 39(!) threads and about 32(!) of those are active 99% of time!
And all these damn extra threads keep 'context switching' all the time so at end of song there are in total about 10,600 context switches (~240/sec): (CaptureOS.PNG)

In contrast, when XX is run from RAM disk there were in total about 5,600 context switches (~120/sec): (CaptureRAM.PNG)

Don't ask me what 10x more threads are doing and what 2x more context switches do or mean but my gut tells me that more threading & switching probably is not good for music playback....

And a small shocker for the end: Try the same thing but start XX from _any_ disk (HD or SSD) as long as that is not OS disk and you should get pretty much same numbers as with RAM disk

So, it would at least seem that XX is best run from a non-OS disk.

An exciting speculation is that perhaps less context-switching has beneficial impact on sound quality?

Can somebody run this test on Vista and let us know if Vista does indeed make less context switches? (I used W7)

If it does, then perhaps we can push Peter to investigate how to make XX do 0 context switches in next version?

[Per]

So, it would at least seem that XX is best run from a non-OS disk.

That would explain why Mani experienced better SQ running XXHE from a 2nd SSD, compared to running on only one SSD, wouldn't it?

Per

[Josef]

>That would explain why Mani experienced better SQ running XXHE from a 2nd SSD, compared to running on only one SSD, wouldn't it?

Yep, sure seems that way
Still, I speculate that RAM disk is probably always a better option as everything is contained within MB i.e. less internal/external hardware gets powered up/down around and more stable power is always preferable for audio....

Fun part is that implications of using #context-switches as measure may go far beyond: For example, what if it can be used to explain (at least to a degree) why different players sound different even though they are all bit-perfect?
And why, e.g. Foobar set to pre-buffer music into RAM before playing and also using KS or WASAPI still sounds worse (eh, different than XX?

So, for fun I tried same experiment with Foobar. One thing immediately becomes painfully obvious: Foobar is context-switching like crazy all the time - even when not playing at all Interestingly, this does not seem to change during playback very much and is very similar to what XX does when run from OS disk. However, #threads is drastically better (=smaller) with Foobar and does not seem to be affected with where Foobar was started from...
So, there seems to be some code in XX which, for whatever reason, is highly susceptible to where XX is run from.....

[PeterSt]

All,

... Great !

Already last Monday I started a post about this, but never got round to finish it. But :

1. Mani is right;
2. Josef is right.

#2 is the most important.
I's a bit late to explain it all in detail now, but indeed open Process Explorer edit : Resource Monitor, and look at the CPU graph THE BLUE LINE at various SFS settings, and file sizes; This blue line is about the processor speed ...

Whatever it is you see, it is not related to anything I do, *except* for the CPU getting more active because a track part has to be loaded. So, at SFS=12 you understand this easlily. But now look at SFS=70 vs. SFS=200 etc. ... '
Does it make any sense to you ?

The reason I didn't post it yet, is because I found that this is very different in W7 vs. Vista. Also, it is different per MOBO "implementation". So, in the end I couldn't see the common denominator, hence the sense of it all. However, it looks like the processor speed is doing it to us all.
Think of something like "when the amount of memory to have under control is larger, it takes me (the CPU) more time to keep that control". But also : "when it is time to load a track part, well, I have to do something for that".
With this in mind, you will start to see how SQ can be so different, and the graph (blue line) just shows it.

But also have the Task Manager's cpu graph in the same screen. Why ? so you can see that the processor speed is independend from the "load" needed. So, CPU cycles needed stay zero all the time (especially when the track fits into memoy), but the speed of the CPU is ony larger.

This is only a small part ofthe whole story, I'm sure, and please keep in mind that this looks totally different for Vista vs W7, will be dependent on energy saving settings, plus it will be motherboard impeeded.
All 'n all we're all on the right track, but we don't know where the train is coming from, nor do we know where it is heading to. Not yet.

Peter


PS: The RAMDisk and saving of the image all works 100%, and after a reboot everything will be in there automatically, but takes 3 reboots to get there. Maybe it can be less than 3, but then it depedens on other (though related) sections.

[PeterSt]

Allright, my last post is a bit "loose" from what Josef discovered, and maybe it looks like I want to show something like "I already knew". But this is far from the case of course, and I only was at "some" track;

Josef, thank you for this really GREAT finding. I hope we all realize how important this is. Also, I hope I can do something with it, and I will do my ever best for that.

I also want to thank Marcin for his trial with the RAMDisk in the first place. It was not new to me, however, when I heard it first (well over a year ago I think) I could not believe it could ever matter for us ... also see Nick's post about the view how the RAMDIsk can ever make a difference when all is played from RAM already. But it so much matters where one reads things, and there is no doubt that we we read things here on our forum, it should be listened to and be further explored. Such a great value ...

Although I should be explicitly thanking everybody for his contribution to this small evolution, I also want to mention the huge amount of time Mani must have been putting into the SFS comparisons, as well as his motivation to proceed on this phenomenon. Boleary may be the starter of this all, although I am not sure at this moment whether Marcin mentioned it first.

It may start to be boring to ever hear me say that it is thanks to you out there that we are where we are and that I can never do this alone. But I'll just say it again ...

THANKS

Peter

[PeterSt]

Can somebody run this test on Vista and let us know if Vista does indeed make less context switches? (I used W7)

I will do this later today.

So, it would at least seem that XX is best run from a non-OS disk.

For Mani : Which coincidentally *is* what I did when I had the two SSDs in my system, a few days back. I forgot about your situation, but most probably in one of your scenario's (with the 2 SSDs) you had it the same.


Peter

[Calibrator]

Reading through the manual for RamDisk I noted that if you are using a 32bit version of Vista/Win7 , that RamDisk will utilize any memory above the 4GB limit normally restricted to. Great I thought, stick in another 4GB in the spare mem slots of my m/b and I can then have my cake and eat it too !

Turned out this wasn't such a joyous endeavour. I have just spend several hours phutzing around trying to get RamDisk working with that additional RAM but have been unsuccessful.

I'm currently playing under 32 bit Vista with SP2 installed.

The issue arises when asking RamDisk to load the saved image at a reboot. I could never find a combination of settings that allowed this. It would create a RamDisk from scratch without issue, but as soon as you requested a reload the system would start to load but then die and kick you back to POST. A jump into SAFE MODE to delete the image file and change settings for RamDisk were the only way of continuing.

In the end I've gone back to working with a 1.5GB Ramdisk using mem under Vista's immediate control. This is sufficient to load 2 albums at a time, which is usually the most I ever do.

I would be interested to know if anyone running 32bit versions of Vista/Win7 with MORE THAN 4GB MEMORY, has been successful in getting RamDisk working correctly using that additional memory.

Also, those running 64 bits versions with more than 4GB ram installed, are you able to create RamDisks up to 4GB OK that survive a reboot intact?

Cheers,

Russ

[Calibrator]

Nick,
your theory definitely holds water.

If the idea is to keep the transfer rates high and the paths short... perhaps it will help to have XXHE on Ramdisk and the music on a SSD connected through PCI slot. Then South bridge is out of the picture.

Is that possible???

Raj

One of these would be nice methinks

http://www.guru3d.com/article/ocz-revodrive-120gb-review/1

Cheers,

Russ

[AUDIODIDAKT]

Hello ALL,

I just want to thank everybody on forum for working together (Marcin, Nick, Mani, Peter, Telstar and more)
I still have to read all posts again.

I know that many people will be eager to hear my findings about the NOS1.
As you could imagine i spend this week over 40-50 hours on the xx project.
I think i will come up with a very nice post this weekend, so stay tuned.
Please give me some time.

For now I only say NOS1 and XX are TOP NOTCH, I will eleborate this, wait for my post !

Peter, Ciska and Paul, you are TOP NOTCH people as well, but I think you'll already know that.


The xx project evolved so much this week, but only due to the fact of combined forces on forum.

I am proud to be a member on this forum, working together with so much smart people.

SO to all who made it happen this week:

Roy

[PeterSt]

[Marcin_gps]

I'm very happy that my work is appreciated I keep tweaking OS, XXHE settings, because sth is always wrong and I can't sleep at night, haha. To be honest, the reason behind this is probably the same with everyone - to achieve the best SQ. Now I must admint I enjoy music like I've never have before, but it's not just because of RAMdisk only. I applied series of tweaks and those that Nick http://www.phasure.com/index.php?topic=1398.msg13228#msg13228 are amazing, I'm speechless. This is the new dimension of sound. I could only imagine what would happen if I turned network services off, etc... but I don't want another OS just for XXHE playback, at least not now.

SFS still matters, unfortunatelly, and I haven't found optimum for me yet.

Greets,
Marcin

[Gerard]

I'm very happy that my work is appreciated I keep tweaking OS, XXHE settings, because sth is always wrong and I can't sleep at night, haha. To be honest, the reason behind this is probably the same with everyone - to achieve the best SQ. Now I must admint I enjoy music like I've never have before, but it's not just because of RAMdisk only. I applied series of tweaks and those that Nick http://www.phasure.com/index.php?topic=1398.msg13228#msg13228 are amazing, I'm speechless. This is the new dimension of sound. I could only imagine what would happen if I turned network services off, etc... but I don't want another OS just for XXHE playback, at least not now.

SFS still matters, unfortunatelly, and I haven't found optimum for me yet.

Greets,
Marcin

There was someone that made a small program to do this automatically. Is there someone that can do it again?

 

[AUDIODIDAKT]

There was someone that made a small program to do this automatically. Is there someone that can do it again?

 

We will work on that too, if there are people with experience on this, please come forward.

We need to make .reg or .bat files to apply reg and os settings automaticly (maybe sound-card and xx settings too),
Also be patience with the vLite subject, we'll go to the bottum of this, but as you could imagine it will take time and lots of experimenting.

Grtz

[Telstar]

This is the new dimension of sound. I could only imagine what would happen if I turned network services off, etc... but I don't want another OS just for XXHE playback, at least not now.

SFS still matters, unfortunatelly, and I haven't found optimum for me yet.

Greets,
Marcin

When i was using the network card to take some stuff from my other pc, i experienced severe SQ loss, so I'm sure that killing all network services will improve the sound a lot, and since I have a dedicated pc for audio, it's the very first thing that I will try in the next optimizations (read back to Vista).