Page 3 of 4

Re: LBB much slower than LB in this pgm

Posted: Mon Dec 30, 2019 4:41 pm
by guest
xtal wrote: Mon Dec 30, 2019 3:41 pm I think AMD / windows drivers may be issue....
Probably, but I still wonder why a driver would impose a speed cap of 1000 characters-per-second, a suspiciously 'round' number. It seems to be deliberate rather than an 'accident'.

Re: LBB much slower than LB in this pgm

Posted: Mon Dec 30, 2019 5:51 pm
by xtal
I just dropped baud to 57600 and still get 220+ seconds

Re: LBB much slower than LB in this pgm

Posted: Mon Dec 30, 2019 6:01 pm
by xtal
19200 baud still get 220088 ms 220+ seconds
this makes little since to me......

Re: LBB much slower than LB in this pgm

Posted: Mon Dec 30, 2019 7:20 pm
by xtal
Tried different program prioritys, high and realtime , made no difference

Re: LBB much slower than LB in this pgm

Posted: Mon Dec 30, 2019 8:14 pm
by guest
xtal wrote: Mon Dec 30, 2019 6:01 pm 19200 baud still get 220088 ms 220+ seconds
this makes little since to me......
On the contrary it makes perfect sense if you assume my theory is correct: that the speed is being deliberately capped to 1000 characters per second. If that's true all rates up to and including 9600 baud will run at the expected speed, and everything above 9600 baud will be capped at the equivalent of 10000 baud irrespective of the selected speed.

Try rates of 9600 baud and below. If you find they run at the expected speed it would support my theory.

Re: LBB much slower than LB in this pgm

Posted: Wed Jan 01, 2020 4:44 pm
by xtal

Code: Select all

crlf$ = chr$(10);Chr$(13)
baud = 2400
passes=10000 
open DefaultDir$;"\baudtst.txt" for output as #out
for xx = 1 to 7  
  bx = baud/10000
  ct=(1000000/baud)*10 
  print #out, "Baud = ";baud;"  ";ct/1000; " ms/char "; ct; " us/char   bx/10k = ";bx  
  print "Baud = ";baud;"  ";ct/1000; " ms/char "; ct; " us/char   bx/10k = ";bx 
  ''open "com12:115200,n,8,1,ds0,cs0,rs" for random as #serial
  open "com12:"+str$(baud)+",n,8,1" for random as #serial
    start = time$("ms")
    for n = 1 to passes   
      print #serial, "ABCDEFGHIJKLMNOPQRST"
    next n
    finish = time$("ms")
    nchars = passes * (20 + 2)
    tt = finish - start
    ''print "Baud = ";baud;"  ";ct/1000; " ms/char "; ct; " us/char   bx/10k = ";bx;chr$(10);Chr$(13)
    print #out, "Actual time taken = ";tt ; " milliseconds  ";tt/1000; "seconds"
    print "Actual time taken = ";tt ; " milliseconds  ";tt/1000; "seconds"
    print #out,"Expected time for ";nchars;" chars is  = "; (nchars *ct)/1000 ; " milliseconds  "; (nchars *ct)/1000000;" seconds"
    print "Expected time for ";nchars;" chars is  = "; (nchars *ct)/1000 ; " milliseconds  "; (nchars *ct)/1000000;" seconds"
    print #out,"Expected time = "; nchars /bx ; " milliseconds  "; (nchars /bx)/1000; " seconds" 
    print "Expected time = "; nchars /bx ; " milliseconds  "; (nchars /bx)/1000; " seconds"
    print #out," "
    print
  close #serial
if baud = 38400 then baud = 28800  
baud = baud * 2  
next xx 
  close #out
print DefaultDir$;"\baudtst.txt   -----DONE----"
    end


Looks like what ever your theory is? May be correct.........


Baud = 2400 4.16666667 ms/char 4166.66667 us/char bx/10k = 0.24
Actual time taken = 914703 milliseconds 914.703seconds
Expected time for 220000 chars is = 916666.667 milliseconds 916.666667 seconds
Expected time = 916666.667 milliseconds 916.666667 seconds

Baud = 4800 2.08333333 ms/char 2083.33333 us/char bx/10k = 0.48
Actual time taken = 457368 milliseconds 457.368seconds
Expected time for 220000 chars is = 458333.333 milliseconds 458.333333 seconds
Expected time = 458333.333 milliseconds 458.333333 seconds

Baud = 9600 1.04166667 ms/char 1041.66667 us/char bx/10k = 0.96
Actual time taken = 228679 milliseconds 228.679seconds
Expected time for 220000 chars is = 229166.667 milliseconds 229.166667 seconds
Expected time = 229166.667 milliseconds 229.166667 seconds

Baud = 19200 0.520833333 ms/char 520.833333 us/char bx/10k = 1.92
Actual time taken = 213589 milliseconds 213.589seconds
Expected time for 220000 chars is = 114583.333 milliseconds 114.583333 seconds
Expected time = 114583.333 milliseconds 114.583333 seconds

Baud = 38400 0.260416667 ms/char 260.416667 us/char bx/10k = 3.84
Actual time taken = 220032 milliseconds 220.032seconds
Expected time for 220000 chars is = 57291.6667 milliseconds 57.2916667 seconds
Expected time = 57291.6667 milliseconds 57.2916667 seconds

Baud = 57600 0.173611111 ms/char 173.611111 us/char bx/10k = 5.76
Actual time taken = 220030 milliseconds 220.03seconds
Expected time for 220000 chars is = 38194.4444 milliseconds 38.1944444 seconds
Expected time = 38194.4444 milliseconds 38.1944444 seconds

Baud = 115200 0.0868055556 ms/char 86.8055556 us/char bx/10k = 11.52
Actual time taken = 220042 milliseconds 220.042seconds
Expected time for 220000 chars is = 19097.2222 milliseconds 19.0972222 seconds
Expected time = 19097.2222 milliseconds 19.0972222 seconds

Re: LBB much slower than LB in this pgm

Posted: Wed Jan 01, 2020 4:55 pm
by xtal
Why does you I7 laptop perform correctly ,,, is it windows or a driver that is doing the capping?

Re: LBB much slower than LB in this pgm

Posted: Wed Jan 01, 2020 5:56 pm
by guest
xtal wrote: Wed Jan 01, 2020 4:55 pm Why does you I7 laptop perform correctly ,,, is it windows or a driver that is doing the capping?
I believe the capping is the driver's reaction to detecting a fake chip. I assume I don't see any capping since my adapter has a genuine Prolific device. I can't think of any other reason why speed capping would be applied, since speeds up to 115,200 baud (and sometimes higher) are common and standard. As far as I know, all drivers, CPUs and versions of Windows support those higher speeds, so long as the chip is genuine.

Re: LBB much slower than LB in this pgm

Posted: Wed Jan 01, 2020 7:08 pm
by xtal
Interesting I will attempt to aquire a Genuine Adapter... Thx

In case you are interested here is LB using same adapter --- The timings don't make a lot of since to me .
Possibly LB is internally buffering????

Baud = 2400 4.16666667 ms/char 4166.66667 us/char bx/10k = 0.24
Actual time taken = 35424 milliseconds 35.424seconds
Expected time for 220000 chars is = 916666.667 milliseconds 916.666667 seconds
Expected time = 916666.667 milliseconds 916.666667 seconds

Baud = 4800 2.08333333 ms/char 2083.33333 us/char bx/10k = 0.48
Actual time taken = 30062 milliseconds 30.062seconds
Expected time for 220000 chars is = 458333.333 milliseconds 458.333333 seconds
Expected time = 458333.333 milliseconds 458.333333 seconds

Baud = 9600 1.04166667 ms/char 1041.66667 us/char bx/10k = 0.96
Actual time taken = 30053 milliseconds 30.053seconds
Expected time for 220000 chars is = 229166.667 milliseconds 229.166667 seconds
Expected time = 229166.667 milliseconds 229.166667 seconds

Baud = 19200 0.52083333 ms/char 520.833333 us/char bx/10k = 1.92
Actual time taken = 30003 milliseconds 30.003seconds
Expected time for 220000 chars is = 114583.333 milliseconds 114.583333 seconds
Expected time = 114583.333 milliseconds 114.583333 seconds

Baud = 38400 0.26041667 ms/char 260.416667 us/char bx/10k = 3.84
Actual time taken = 30006 milliseconds 30.006seconds
Expected time for 220000 chars is = 57291.6667 milliseconds 57.2916667 seconds
Expected time = 57291.6667 milliseconds 57.2916667 seconds

Baud = 57600 0.17361111 ms/char 173.611111 us/char bx/10k = 5.76
Actual time taken = 30004 milliseconds 30.004seconds
Expected time for 220000 chars is = 38194.4444 milliseconds 38.1944444 seconds
Expected time = 38194.4444 milliseconds 38.1944444 seconds

Baud = 115200 0.86805556e-1 ms/char 86.8055556 us/char bx/10k = 11.52
Actual time taken = 30006 milliseconds 30.006seconds
Expected time for 220000 chars is = 19097.2222 milliseconds 19.0972222 seconds
Expected time = 19097.2222 milliseconds 19.0972222 seconds

C:\Users\msw10\Downloads\baudtst.txt -----DONE----

Re: LBB much slower than LB in this pgm

Posted: Fri Jan 03, 2020 10:14 pm
by xtal
Got the same Plugable usb serial that you have..
Driver is Prolific 7/30/2019 ver 3.8.31.0
Testing on my main unit is samo samo.
The Adapter made no difference .
Must be AMD MB ???? or Intel i7

Try your AMD w7 at 57600 to see if it is correct.