| T.R | Title | User | Personal
Name | Date | Lines |
|---|
| 123.1 | Will it be changed? | ALFSS2::OLSEN_G | Hey y'all, how y'all doin'? | Mon Mar 17 1997 10:09 | 18 |
| John,
Your note implies that you "may" change this but doesn't say you will.
I have a customer complaining about it - not because it makes anything
fail and not because it reduces performance, but because he wants to
see his idle time when he monitors it. I gave him the explanation in
.0 but he is still whining about it.
Not that I'm suggesting you change it, but I need to tell this customer
if you are going to leave it like that.
BTW - thanks for the note - saved me some time on this call Saturday
night!
Gary Olsen
Atlanta CSC
|
| 123.2 | | JAMIN::WASSER | John A. Wasser | Mon Mar 17 1997 10:13 | 6 |
| > Your note implies that you "may" change this but doesn't say you will.
The work is piled so high here that unless a CLD arrives I don't
think the problem will be addressed. Remember: the fault is in
the DECnet stack.
|
| 123.3 | There is a Performance degradation | ALFSS2::OLSEN_G | Hey y'all, how y'all doin'? | Mon Mar 17 1997 13:16 | 12 |
| OK - I'll consider a CLD.
Just FYI, the customer now claims there is a
performance degradation - typing sometimes takes a few seconds to echo
to the screen, mouse clicks become so slow they re-click because they
don't think it took - all of which started happening after PW32 was
installed.
Thanks,
Gary
|
| 123.4 | Another PW32 Customer bites the dust! | ALFSS2::OLSEN_G | Hey y'all, how y'all doin'? | Fri Mar 21 1997 15:01 | 28 |
| After a rebuffed CLD on this and Vivian telling me Engineering doesn't
have any plans to change this, and unsuccessfully trying to convert the
customer to Server based licenses (doesn't want to spend more money),
this customer is not going to install PW32 on his network and will ask
for a product and support contract refund.
He is upset that we "hog all the cpu". I know the answer - we do but we
give it up - but that didn't convince him. He is still convinced PW32
causes his delays in echoing keystrokes and recording mouse clicks.
He has agreed to install pw32 on another machine to try to reproduce
the problem, so we'll see.
Just wanted to register his complaints with engineering regarding:
1. The cpu utilization issue
2. Our "premature" support of winsock 2 - ahead of Microsoft-
causing problems (Exceed).
3. Our unforgiveable decision to not include a de-install program.
4. The crashing exiting from windows which seems to be going away.
The two biggies are #1 and #3. #4 is being worked on. If you are
trying to prioritize tasks, consider this customer.
Thanks,
Gary
|
| 123.5 | delay in general, or w/eXcursion | EDSCLU::NICHOLS | | Tue Mar 25 1997 10:38 | 0 |
| 123.6 | Another slow one ! | NETRIX::"[email protected]" | Stephen Langstaff | Thu Mar 27 1997 06:04 | 16 |
| Hi,
I have just had a customer complaining about how his CPU is being used up 98%
by the PNLT32.exe process. I gave him the explanation in note .0 but he is
saying his PC is much slower than when using v1.0a under Win95. He is probably
going to switch back to V1.0a. Just thought I would register my customers
comments as well !!!
Is anything going to be done about this ???
Cheers
Stephen
PCI Support Basingstoke
[Posted by WWW Notes gateway]
|
| 123.7 | | JAMIN::WASSER | John A. Wasser | Thu Mar 27 1997 09:34 | 3 |
| > Is anything going to be done about this ???
Yes.
|
| 123.8 | and again.... | NETRIX::"[email protected]" | Stephen | Wed Apr 09 1997 11:28 | 11 |
| Just had another cust who installed PW32 on Win95 system, complaining heavinly
about the speed his keyboard was now working. he has just informed me he will
be reverting back to PW v1.0a which he thought was brilliant until he gets a
new
faster machine.
What is happening about this and when >>> anybody know ?
Cheers
Steve
[Posted by WWW Notes gateway]
|
| 123.9 | | JAMIN::MAJEWSKI | | Wed Apr 09 1997 13:34 | 3 |
| we (PW engineering) are working on 7.0a and expect to make a change in
licensing to work-around this issue.
|
| 123.10 | Update to 100% cpu stuff | SPELNK::curless | | Thu Apr 10 1997 12:10 | 7 |
|
The change as been made, and does indeed lower the CPU utilization to something
more normal.
This will be available on v7.0a
Jeff
|
| 123.11 | | VMSNET::N_WILKINSON | 16 bit? We don't need no stinkin' 16 bit! | Wed Apr 16 1997 10:33 | 21 |
| Well, I have been able to reproduce the transponder causing the system to slow
down. Due to constant reconfiguring of my pc for testing, I had decided to
delete win95 and totally reinstall. Hardware is a Digital Celebris 5100. I
installed WIN95 then Plus! then PW32 and went home. In the morning I noticed
that my screen saver's animation was VERY slow. After removing the license
agent, normal screen refresh and screen saver motion was restored. I did some
further testing. I removed DECnet and licensing and everything was fine. Then
I ran PNLT32.EXE and System Monitor shows 100% CPU and my screen saver is slow
again.
Keep in mind, I don't have anything else installed so I am using the screen
saver, which is very CPU intensive, to test response. Also notice I am seeing a
problem without DECnet installed which is contrary to the base note.
Also, this system ran PW32 fine before I deleted everything and reinstalled! I
have had many beta versions on this machine as well as all kinds of apps, so
there is no telling what might have been running. I am installing from the
PATHWORKS cd, all versions are SSB.
Nelson
ALF CSC
|
| 123.12 | | JAMIN::WASSER | John A. Wasser | Thu Apr 17 1997 11:43 | 12 |
| > I removed DECnet and licensing and everything was fine. Then I ran
> PNLT32.EXE and System Monitor shows 100% CPU and my screen saver is slow
> again.
It may be that you have a faulty transport and PNLT32.EXE is
getting stuck in a retry loop. Run the debug version in
\w95\debug\service\license\lanman and check the resulting
PNLT.LOG file to see if it is growing.
In order to support dial-up transports, PNLT32.EXE will
re-try a transport that fails in case it was a dial-up
transport that just wasn't dialed yet.
|
| 123.13 | | VMSNET::N_WILKINSON | 16 bit? We don't need no stinkin' 16 bit! | Mon Apr 21 1997 16:53 | 23 |
| I fired up the debug version of PNLT32.EXE and got 100% CPU utilization and my
screen saver goes into molassas mode. PNLT.LOG follows. BTW, all I did was
take my machine with no licensing installed and run debug PNLT32.EXE. After 5
minutes, this is all that is in the log file:
******************************************************************
Log started at 1997-04-21 15:41:13.510
Copyright (c) 1993-1997 Digital Equipment Corporation
Service name is: "WORF R"
LANA 0x00: Starting first NCBADDNAME
LANA 0x00: Starting thread
LANA 0x07: Starting first NCBADDNAME
LANA 0x00: Starting NCBLISTEN
LANA 0x07: Starting thread
Ready on 2 LANAs.
PWCLR_running() == FALSE
LANA 0x07: Starting NCBLISTEN
GetLastError()=
"The filename, directory name, or volume label syntax is incorrect."
PWCLR_start...
Entering message loop
Received LIC_GET_TABLE IPC
Received LIC_LOAD_TABLE IPC
|
| 123.14 | | JAMIN::WASSER | John A. Wasser | Tue Apr 22 1997 10:20 | 13 |
| > I fired up the debug version of PNLT32.EXE and got 100% CPU utilization and my
> screen saver goes into molassas mode.
Perhaps the screen saver is running at a lower than deault
priority. What was the "normal screen refresh" that you
said was restored when you removed licensing?
> PNLT.LOG follows. BTW, all I did was take my machine with no licensing
> installed and run debug PNLT32.EXE. After 5 minutes, this is all that
> is in the log file:
My fears were not realized... The problem in the DECnet
stack is probably the cause of the screen-saver slow down.
|
| 123.15 | NO DECnet loaded | VMSNET::N_WILKINSON | 16 bit? We don't need no stinkin' 16 bit! | Tue Apr 22 1997 15:43 | 42 |
|
> Perhaps the screen saver is running at a lower than deault
> priority. What was the "normal screen refresh" that you
> said was restored when you removed licensing?
I am using the MS "Science" screen saver from MS Plus! pack. Normally when the
screen saver kicks in the first thing that happens is that the wallpaper is
rewritten to the screen. This normally happens in a split-second. With
PNLT32.EXE running WITH OR WITHOUT DECnet, the wallpaper is rewritten in chunks
of about 1/6th of the screen over a period of 10 seconds and the "ball" normally
moves across the screen at about 1 inch per second now moves 1/4 inch per 3
seconds. This is only the benchmark I have been using to test system
performance.
So, I have now run WINTUNE to benchmark the system. The system is a pentium 100
mhz with 40 mb of memory. LAT, Netbeui and TCP/IP only. No DECnet!
Video = 5.447 Mpixels/Sec
Disk = 12.71 MBytes/Sec
CPU = 102.894 MIPS
Math = 20.833 MFLOPS
With PNLT32 running:
Video = 5.221 Mpixels/Sec
Disk = 12.122 MBytes/Sec
CPU = 90.141 MIPS
Math = 18.519 MFLOPS
Essentially unchanged. And of course the numbers do have some "play" in them in
that they will change from one running to the next.
Conclusions:
1) The screen saver is running at a low priority that affects the way it runs
with PATHWORKS licensing.
2)DECnet is not installed, so that is not causing the 100% utilization nor the
slow screen saver action.
3) This all makes us look very bad to customers who see how much PATHWORKS is
__preceived__ to slow down thier system.
|