TIP: Click on subject to list as thread! ANSI
echo: clipper
to: RAYMOND PESEK
from: RAPHAEL NEVE
date: 1996-10-15 07:44:00
subject: Solution for amd/cyrix pr

Raymond Pesek wrote in a message to Raphael Neve:
 RP> Do you still have the message from the fellow who posted
 RP> what exactly AMD said the problem was? I didn't save it and
 RP> need a copy. I have someone who may want to tackle the
 RP> problem.
Subject:      CLIPPER & AMD processors & timeslices problem.
From:         taavi@ratioma.ee (Taavi Remmelg)
Date:         1996/09/28
Message-Id:   
Newsgroups:   comp.lang.clipper
Hello !
I have problem with AMD processors/ program just does't start/, maybe
somebody know solution to this problem.
I reported problem to AMD and have they'r answer. I do'nt belive I am
the first who has this problem / and not the last one/.
Can somebody comment this?
Taavi Remmelg.
Tallinn, Estonia.
Our discussion comes here:
>Hello there!
>
>We have a little problem here with with AMD 5K86 processor.
>It seems that my clipper (5.2E + Blinker 4.0 Extended mode) EXE will 
>not start (100%) on computers with this chip with INTERNAL 
>CACHE ENABLED. Two our biggest local computer-manufacturers 
>have tested they'r configurations /maybe motherboard../ with no 
>success. 
>It seems like AMD refuses to switch to DOS protected mode.
>I already contacted Blinkinc UK but they seems not to have solution 
>for this situation.
>
>Can You help us?
>
> Our local computer manufaturers here are selling a lot of 
>AMD - based workstations... 
>
>Sincerely Yours,
>Taavi Remmelg.
Dear Taavi Remmelg,
the problems with Clipper are known to us and Clipper, you
should contact Clipper for an update of their Software.
There are two reasons why it could be that Clipper programs
do not run.
First reason : Software is using Timing Loops. These are sequences of
  code to insert wait states between software events.
   It is recommended that the Time Stamp Counter be used to insert
   precise timing delays between code sequences.
   The problem with using timing loops is that the loops are solved
   faster on the K5 than on many other CPU's of the same kind.
second reason : Accessing non-existent Model Specific Registers (MSRs)
   As the "Model Specific" name implies, different processor models
   ( same vendor or not ) have different model specific registers.
   Access to not implemented MSR's or different implemented MSR's
   is causing a general protection fault ( this is correct operation ).
   It is recommended by the Software to check the CPU ID and the
Vendor String befor accessing MSR's.
At this moment you can only workaround this problem with Clipper if
you turn the cache off, or switch the brange prediction logic off.
Kind Regards
            Markus
>~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
>                                                 _______
>                                                 \____  | Advanced
>European Corporate Applications                  /|   | | Micro
>                                                | |___| | Devices
>EMail id   euro.tech@amd.com                    |____/ \|
>Phone +44 (0)1276 803299
>Fax   +44 (0)1276 803298
>Our Web site is http://www.amd.com
>~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
raph.
... A feature is a bug with seniority. 
--- FMail 0.96â
---------------
* Origin: Canada Dry BBS - France * 3 lines on 47.29.33.85 (2:321/1)

SOURCE: echomail via exec-pc

Email questions or comments to sysop@ipingthereforeiam.com
All parts of this website painstakingly hand-crafted in the U.S.A.!
IPTIA BBS/MUD/Terminal/Game Server List, © 2025 IPTIA Consulting™.