Re: States 2D on MSL?
G. Pearson (gpearson@blue.weeg.uiowa.edu)
Fri, 2 Feb 1996 19:28:11 -0600
I asked last week for ideas of how to GRACEFULLY do hypercomplex 2D on an MSL.
Doug Burum responded with a suggested change of syntax in one of my ideas, and
asked me to post what I find out here. I tried it out, and it works. But
there's a "gotcha" -- it's very slow.
What Doug said was:
> I think your idea of using a .AUM program is close to the right method.
> The only problem is that you apparrently specified the pulse program as
> FILE.PC or FILE.PPG. In an AU program, the RP command should be followed
> by the name of the pulse program with NO EXTENSION.
>
> For example, if the file is called SINE2D.PC, you should first compile it
> using PPE SINE2D.PC to read it into the editor, and then exit with (esc-Q)
> to cause PCCOM to compile it.
>
> After running the PPE editor, you will have created a file called
> SINE2D.PPG. In the .AUM program, you will then need the following
> statement:
>
> RP SINE2D
>
> This statement will cause the system to look for a file called SINE2D.PPG,
> and to read it in as the current pulse program. I would expect you to get
> errors in compiling the .AUM program if you use either of the following
> statements:
>
> RP SINE2D.PC
>
> or
>
> RP SINE2D.PPG
>
> Let me know if this suggestion fixes the problem, and also please be so
> kind as to post the information on the Bruker User's mail at Berkeley.
So I wrote & compiled COSTHING.PC and SINTHING.PC for acquiring the cosine and
sine fids. The key parts of the AUM program are:
10 RP COSTHING ; read in COSTHING.PPG
ZE ; zero the data
GO ; acquire a fid with the COSTHING pulse sequence
WR #1 ; save fid to disk
IF #1 ; increment file pointer
RP SINTHING
ZE
GO
WR #2
IF #2
I4 ; increment D4, the "dwell time" for the 2nd dimension
IN=10
EXIT
Now for the bad news. With a stopwatch, I measured the time between
successive disk saves. This was 14.8 seconds longer than what I had in the
pulse program itself. This means that the 2 commands RP and WR take about
15 seconds of overhead -- with an old 96 MB CDC disk drive.
On an AC system running DISRxx, the overhead is only 3 seconds for the disk
write, and there are no disk reads in the AUR program.
This MSL overhead time might improve somewhat after we replace the CDC with a
Winchester disk, but I expect that it'll still be slow.
BOTTOM LINE: You _can_ do hypercomplex data acquisition on an MSL, but the
long time overhead -- about 30 seconds per evolution time -- is probably
_much_ too high a price to pay in at least most cases. In all probability,
you're better off doing absolute-magnitude 2D.
---------------------------------------------------------------
Gerald A. Pearson INTERNET: gerald-pearson@uiowa.edu
Chem. Dept., Univ. of Iowa VOICE: 319-335-1336
Iowa City, IA 52242-1219, USA FAX: 319-335-1270
---------------------------------------------------------------