stan@phx.mcd.mot.com (Stan Fisher) (07/13/90)
I just upgraded the ParNET (1.0?) I was using to the newly distributed version that came over binaries. (2.0?) Has anyone noticed it being about half (or less) the speed of its predecessor??? I do backups across it from an A500 w/48 meg SCSI drive to an A2000 w/tape and and what used to take a little over an hour to handle the whole backup (to tape) was only about 19megs (out of 44 meg) into the backup after 2.5 hours!! I finally stopped the backup figuring something is wrong here. Does the new parpnet.handler do something much slower than the earlier parnet.handler? For now, I'm going to go back to 1.0 ParNet. I havn't tried any diskperf tests across ParNET to compare the two.. if I get a chance I'll do that and post the comparisons here (if they show the differences I suspect) Stan Fisher - stan@teroach.phx.mcd.mot.com - asuvax!mcdphx!teroach!stan Motorola Microcomputer Division, Tempe, Arizona - Voice (602) 438-3228 Call our User Group BBS "M.E.C.C.A." running Atredes 1.1 @ (602) 893-0804
kent@swrinde.nde.swri.edu (Kent D. Polk) (07/14/90)
In article <13197@mcdphx.phx.mcd.mot.com> stan@phx.mcd.mot.com (Stan Fisher) writes: >I just upgraded the ParNET (1.0?) I was using to the newly distributed >version that came over binaries. (2.0?) > >Has anyone noticed it being about half (or less) the speed of its >predecessor??? Hadn't noticed it, but ran diskperf & it was very close to 1/2 the speed of ParNet Version 1. >Does the new parpnet.handler do something much slower than the earlier >parnet.handler? For now, I'm going to go back to 1.0 ParNet. Fixes some problems, introduced a few others. The Amiga shell can't cd to a net: directory. Cshell4.01 can cd, but can't find anything when it gets there. I have to use absolute pathnames for everything but my old friend... Browser!!! (Peter: Thanks. I bet you thought everyone was going to throw away Browser when 2.0 came out. Fact is that Browser is about an order of magnitude faster in moving files than Workbench 2.0 is. Think I'll keep it around :^) Anyway, Browser works fine with parnet v.2. One other interesting item: I started up cputrap on the 2500 parnet server when diskperf was running across parnet & got a rash of invalid memory accesses. FAULT ADDRESS: 32 TASK NAME: PARNET.DEVICE Kent Polk: Southwest Research Institute (512) 522-2882 Internet : kent@swrinde.nde.swri.edu UUCP : $ {cs.utexas.edu, gatech!petro, sun!texsun}!swrinde!kent
walker@unx.sas.com (Doug Walker) (07/18/90)
In article <13197@mcdphx.phx.mcd.mot.com> stan@phx.mcd.mot.com (Stan Fisher) writes: >I just upgraded the ParNET (1.0?) I was using to the newly distributed >version that came over binaries. (2.0?) > >Has anyone noticed it being about half (or less) the speed of its >predecessor??? > I wasn't aware that the speed difference was THAT great, but I did accidentally compile it with a 1k buffer size instead of 8k because I had been working with TSSNet, which was limited to 1k buffers. Matt had earlier upped the buffer size to improve performance. As soon as work cools down a little I will be getting parnet3 out, which should fix it. ***** =*|_o_o|\\=====Doug Walker, Software Distiller====== BBS: (919)382-8265 = *|. o.| || | o |// For all you do, this bug's for you! ====== usenet: ...mcnc!rti!sas!walker plink: dwalker bix: djwalker
dillon@overload.Berkeley.CA.US (Matthew Dillon) (07/19/90)
>Hadn't noticed it, but ran diskperf & it was very close to 1/2 the >speed of ParNet Version 1. > >>Does the new parpnet.handler do something much slower than the earlier >>parnet.handler? For now, I'm going to go back to 1.0 ParNet. Ow, there is something wrong here... At least on my A1000 the speed stayed around 28KBytes/sec between versions. Perhaps the speed of the 2500 is causing some inefficiency that I did not notice with my A1000. >Fixes some problems, introduced a few others. The Amiga shell can't cd >to a net: directory. Cshell4.01 can cd, but can't find anything when it >gets there. I have to use absolute pathnames for everything but my old >friend... ???? Doesn't happen to me! I can CD just fine! >One other interesting item: >I started up cputrap on the 2500 parnet server when diskperf was running across parnet >& got a rash of invalid memory accesses. > >FAULT ADDRESS: 32 >TASK NAME: PARNET.DEVICE Grr... I need some of these debugging programs! Actually, I need to get myself an A3000 to make use of the MMU for debugging. The 1.06 UUCP UUCico also has memory problems. >Kent Polk: Southwest Research Institute (512) 522-2882 >Internet : kent@swrinde.nde.swri.edu >UUCP : $ {cs.utexas.edu, gatech!petro, sun!texsun}!swrinde!kent -- -Matt Matthew Dillon dillon@Overload.Berkeley.CA.US 891 Regal Rd. uunet.uu.net!overload!dillon Berkeley, Ca. 94708 USA