EFS Drives CPU to 100%

[cpatte7372]

Hi,

Can someone please take a quick look at the attached efs and give me some idea why each time I load it on over 5 charts it drives my CPU crazy.

Cheers

Carlton

P.S.

I think the root of the problem is where I have added "Sym" everywhere, but not quite sure.

[gspiker]

Carlton,

There is nothing that really leaps out as being a CPU hog here, unless you are constantly hitting your alerts. That would cause some problems due to the overhead of those calls.

There are certainly more efficient ways of doing some things (from the POV of the CPU), but I don't think these will result in any major resource improvement.

Garth

[cpatte7372]

Thats great Garth,

Thanks for checking it out mate.

Cheers

Carlton

[cpatte7372]

Actually Garth,

I do believe there is a problem with the efs causing the cpu to run constantly at 100%. Immediately after loading 12 charts the cpu is stuck at 100% and esignal grinds to a halt.

So, if there is anybody that can lend a hand to troubleshoot this problem I would be most appreciative as I plan on making this efs an integral part of my trading.

Cheers

Carlton

[cpatte7372]

What I have noticed is that the efs only appears to hog the cpu when they are maximised. However, when they're minimised everything appears norma. Can someone let me know if that is common. And how many charts do you open at any even time.

[DuaneG]

Hi Carlton,

I'm getting the same thing here. I think that while the chart may appear to be ok when minimized, the load time when applying the study to the Advanced Chart window is quite significant. Just wanted to confirm what you are experiencing.

[cpatte7372]

Can someone please help me out. Its driving me crazy. I'm having to continully restart esignal everytime I maximise as few as 3 charts with this efs. I will be getting in touch with a support representative, however I tend to get more detailed assistance from this forum than I do with a esignal support representative.

Cheers

Carlton

[cpatte7372]

Duane,

Its good to know that you also identified the same issues I'm currently experiencing.

I called the British esignal support and I was informed that I would have to be patient and wait for a volunteer to help me from this forum. Oh well.


Carlton

[gspiker]

Carlton,

Attached is a performance moded version of the DI EFS.
I fixed a few bugs (both from the original script and from your mod to the original script) along the way - so the values may not line up exactly with what you had.

It seems to take up roughly 1/2 the amount of CPU resources.

I think the killer (which I missed when I first looked at it) was the call to volume() in the loop statement. Still...I was able to run your version of the script on my system with 4 charts and had no problems...

Anyway here it is.

Garth

[cpatte7372]

Hi Garth,

Thanks ever-so-much mate for taking a look at this. It now appears to be functioning fine. However, as you mentioned, the values are out of line with the original script. The original script was written by a chap called Alex. As I didn't write the original script can you let me know what changes you made that make the values differ? Alternatively, I suppose I could familarise myself with Steve's trace function and try and work it out myself.

Cheers

Carlton

[gspiker]

Carlton.

Unless the coder at TSSUPPORT was named Alex, I think you have the wrong name.

re: bugs.

In your version of the EFS, along with the bugs generic to the original EFS you had the volume() statement in the loop not referencing your alternate symbol.

In the original version of the EFS there were conditions where you could get NaN's from some of the math logic. This in turn would screw up the DMI result.

There was also what I presume was a bug in a loop, where they were taking a array of length 5 and trying to reference 6 paramters from it (they used length instead of length - 1). This could cause some odd results.

As you can see from the comments in the original you posted I have already tweaked this code before to allow it to run in realtime...so this isn't the first time I have touched this.

Garth

[cpatte7372]

Garth,

What I should have said was Alex tweaked it for me in the past - sorry about that.

So, my only other question is, are the values more likely to be in tune with what they should be, now that you've made the changes? I will still run Steve's trace function on it and carry out a few tests myself but I just thought I would ask. If I get the same values after running a few tests I will feel more confident.

Cheers

Carlton

[gspiker]

Carlton,

They should be. With the original EFS you psoted (pre your mods) the problems may not always cause errors in the result. So it is possible that the results of mine and that EFS would line up in some cases.

Since you mods sometimes used the volume from "sym", and sometimes used the volume from the symbol on the chart, it should never line up exactly.

G

[cpatte7372]

Steve,

I couldn't thank you enough mate. It hasn't crashed once.

I'm going to search for the original D.I. you posted sometime ago and compare the values with this one. Hopefully, they'll be similar if not identical.

Cheers

Carlton

[cpatte7372]

Steve,

Sorry to be a pain. I was comparing the values of your original demand index and the one I posted(before and after I made modifications) and the tally exactly.

I have been going over and over the script you modified and I can't see why I keep on getting different values from your original and my modified one.

Carlton