Help with a test?

[Opry99er]

Need some help, all...

 

I have a disk-access routine that reads records from an INT/FIX file on disk, then displays the data onscreen.

 

The attached zipfile is simply a copy of my FIAD DSK1 from my Classic99 folder. It contains two files... the IF80 data file and an XB program called LOAD.

 

 

 

 

 

 

Obviously, it will autoload.

 

The test program is simple... It will ask you to press "Enter" to begin, gives you a "ready, set, GO" kind of thing so that you can start the stopwatch on time... Once the routine is complete, it displays "STOP" and you can stop the stopwatch..... (alot of "stop"s in that last sentence...)

 

 

 

 

 

 

 

You can do this 4 or 5 times, each time it will display different text. It should take EXACTLY the same amount of time for each display, even though some are shorter than others... I have padded the records and sometimes the loop/print is printing blank lines... Remember to stop the timer ONLY when the program tells you to STOP.

 

 

After the 4th or 5th time, it will give you an error that breaks the program back to the XB environment. You can run it as many times as you want by typing RUN "DSK1.LOAD"

 

WHAT I NEED:

 

I need someone with a real TI FDC running a real floppy disk (preferably 5.25, but I don't think it really matters) to test this program and time the routine. In Classic99, I am getting 1.05 seconds consistently.

 

If this takes a long time on real hardware, I have a couple other ideas on how to change the code to increase the speed... For now, please help me test this current code. I would greatly appreciate it.

DSK1.zip

Edited May 6, 2015 by Opry99er

[+mizapf]

FWIW, a test with a BwG on MESS (DSDD):

1. Iteration: 3.8 sec
2. Iteration: 4 sec
3. Iteration: 3.66 sec
4. Iteration: 3.66 sec
5. Iteration: 4 sec
6. I/O Error 25

TI controller on MESS, SSSD:

1. Iteration: 4.16 sec
2. Iteration: 4.33 sec
3. Iteration: 4.17 sec
4. Iteration: 4.02 sec
5. Iteration: 4.18 sec
6. I/O Error 25

[+mizapf]

 

FWIW, a test with a BwG on MESS (DSDD):

TI controller on MESS, SSSD:

 

I should add these times are taken by hand with a stopwatch, so there is a tolerance of about 0.2 seconds (at least).

[Opry99er]

Hey, thanks Michael! I appreciate the test run.

 

interesting results.... Did the delay seem cumbersome? In other words, were you waiting on the next line, or was it fairly smooth as far as from a reader's point of view?

[+mizapf]

I tried to create a screencast, but xvidcap does not work with the AMD driver in Linux, as it seems (only getting a still image). Then I filmed the output with my photo camera and got a 100+ MiB movie ... <sigh>.

 

So let me describe it like this: About 1.5 seconds of the four seconds of loading, you don't see any change on the screen. After that, the lines are showing up, one after the other, but too fast to keep up with reading.

 

I'll be with my real Geneve and its floppy drive on Friday again, so maybe someone else may be able test this in the meantime.

[Opry99er]

Interesting... I really appreciate the info on the MESS side.

 

1.5 seconds of load time... I wonder if that will be off-putting for users...

 

I am interested in hearing from someone on hardware too... In context, the 1.5 seconds might not be too bad... If it is much more than that with real drive and FDC however..... Might need to rethink.

 

Thank you again.

[Opry99er]

Any of you HDX guys feel like doing a quick transfer to DSK1 and rockin' this out?

 

 

I REALLY need to get my TI set back up... That means putting one of the kids out in the shed though...

 

Choices, choices!

 

[+Lee Stewart]

I should have something in an hour or so.

 

...lee

[sparkdrummer]

Uh oh, kid in the shed. Names not Karl is it?

[Opry99er]

Haven't decided which kid yet...

 

I HAVE considered building a custom desk out in the shed.... Setting up out there.

 

It is a storm shelter... Coolish in the Summer...

 

No power, BUT I have been looking at building a solar system for it... 3 marine deep cycle batteries in series out to a converter to run a power strip...

 

There are plans for them online... It would run my TI, XP machine, and my radio (for listening to baseball games)

 

 

http://m.instructables.com/id/Solar-Power-System/

 

I have WiFi, so my wireless USB adapter on my computer would hook up to the internet fine.

[+Lee Stewart]

OK...Real iron with a CorComp DSDD disk controller:

(Average of 2 runs)
   4.02 seconds
   4.16
   3.66
   3.70
   4.02

...lee

 

[Edit: Oh, yeah...and with 5.25" diskette]

Edited May 6, 2015 by Lee Stewart

[sprintcarfan]

Real Iron with TI FDC and Shugart SA-455 5.25" floppy. Had to modify line 200 to point to DSK3 because that's where my real floppy is. I get an "I/O error 25 in line 220" at the fifth or sixth go/stop cycle the program.

 

4.490 seconds

4.308

4.495

4.184

4.369

4.187

4.249

4.399

[Opry99er]

Great fellas... I REALLY appreciate the help here.

 

I do not think that 4 +/- seconds is prohibitively slow... Reckon I will keep the current code.

 

Obliged, truly.

[Opry99er]

Interestingly, the average of all of the tests shows the 4th iteration is the fastest...

 

That is odd!!!

Edited May 7, 2015 by Opry99er

[+mizapf]

And I'm really glad to see that our emulation in MESS is not too far away from the real drive behavior. I don't have a chance to verify everything by real hardware, so this was a good opportunity to check.

 

When I'll be done with the update of the hard disk emulation, I'll try to add drive sounds so that you can hear the head step and the floppy motor.

[Opry99er]

^^^ CRAZY TALK!!!!^^^

 

 

Thanks again for your help, Michael.