[HECnet] Re: [HECnet] [HECnet] Re: [HECnet] Issues bringing TOPS-20 UP (again )

[Jordi Guillaumes i Pons]

El 04/02/2013, a les 22:56, Cory Smelosky <b4 at gewt.net> va escriure:

Nevermind, Everything's fine after a rebuild of klh10   

Or not   still lossy   This is bizarre.   It was fine until the VM host moved to vmware...

I have three news for you:

1: It's a known bug (good)
2: It has a workaround (good)
3: You won't like it (baad!)

The bug is related with the async timer implementation in KLH10. It has been discussed in the TOPS-20 mailing list and someone (IIRC was the late Mark Crispin himself) said it was not fixable. The relevant details (and the workaround) are here:

https://groups.google.com/forum/?hl=en&fromgroups=#!msg/alt.sys.pdp10/xndbqJ6Qmwg/rg4pcD923AgJ

Now the bad news. The KLH10 idling "device" is based on the asynch timer implementation, and DOES NOT WORK with synch timers. That means KLH10 will use all your available CPU, all the time. 

There is ANOTHER workaroud for that (I'm using it with my KHL10 simulator running in a rasp-pi). If you are going to use Linux search for a nifty utility called cpulimit. It does just what you think it does :). In a crude way, but it works.


Jordi Guillaumes i Pons
jg at jordi.guillaumes.name
HECnet: BITXOV::JGUILLAUMES