[MUD-Dev] Re: Ethernet NICS, maximum connections..mud testing.

Chris Gray cg at ami-cg.GraySage.Edmonton.AB.CA
Sun Aug 9 15:58:33 CEST 1998


[Ben Greear:]

 >Check my web page (see .sig at bottom).  Its on the Downloads link.

Got it, thanks.

[My complaining about unknown delays in sockets on Linux.]

 >Not that I know of.  You do know that select sleeps if there is
 >nothing available right?  Other than that, I don't know enough
 >about your code to make a better guess.

Yes. I use that delay as the pacer in the whole system. I also check to
see if the actual delay is much greater than the requested delay. I
see occasionally it being 2/100 sec greater, usually when I'm doing
something else on the system at the same time. Quite reasonable.

 >How big are the packets you send between client and server?

They vary. There is a fixed header of 12 bytes, optionally followed
by other data, perhaps upto 1 or 2K in the extreme. Most will be
quite short during the 'build the world' phase, which is where my
concern is.

 >Are you running between two machines?  If so, try to benchmark it
 >against bulk transfer (ftp).

No - they are all local on one machine.

--
Chris Gray     cg at ami-cg.GraySage.Edmonton.AB.CA




More information about the mud-dev-archive mailing list