Path: utzoo!utgpu!news-server.csri.toronto.edu!mailrus!uwm.edu!rutgers!faatcrl!jprad From: jprad@faatcrl.UUCP (Jack Radigan) Newsgroups: comp.sys.amiga Subject: Re: JR-Comm 1.10 Features Message-ID: <77@faatcrl.UUCP> Date: 21 Aug 90 01:46:51 GMT References: <72@faatcrl.UUCP> <1054.26cdefd9@desire.wright.edu> Organization: FAA Technical Center, Atlantic City NJ Lines: 43 arc@desire.wright.edu writes: > Jack, I have used your JRComm for a while now, and find it very nice. About >the internal ZModem, I don't find it that efficient... I have used NComm 1.9 >with it's external ZModem, and it takes MUCH less CPU time (I used XOper, plus >by doing some mtasking, etc.) for all xfering. I run your term at a 0 >priority, just as I do NComm 1.9 (NComm 1.9 is buggier than JRComm 1.0), but >the Zmodem library does an AWSOME job. I will be uploading with JRComm 1.0 at >a 0 priority (on its own screen) and then when I go to resize a window on my >workbench screen (interlaced) your term pause the transmit for a split >second... This really bothered me that you'd say that an external ZModem >wouldn't achieve the thoughput that your internal does... Well, since you asked... My tests were conducted between a 25MHz A3000 w/4M SCRAM and an A1000 w/'010 and 4M reg. RAM. All transfers were done to and from RAM: using ZMODEM, no handshaking with a null modem. With XOPER running, 4800bps was the highest speed that would complete a file transfer without errors, cpu utilization was at 100%. JR-Comm averaged about 43% with 38 and 50 as the extremes. NCOMM averaged about 51% with extremes of 47 and 63. Both of these transfers were at 4800bps using a 162k file with ZMODEM receive. Additionally, JR-Comm was using an 8 color screen to match screen DMA with NCOMM. So, on the whole, cpu usage was 10 points lower with JR-Comm than with NCOMM. But, cpu utilization was *not* what I was refering to, "throughput" was. And, throughput is best shown with uploads since you can only receive as fast as it is sent to you. NCOMM can transmit at only 1573cps @ 19.2kbps, or about 81.9% throughput. On the other hand, JR-Comm gets 1850cps, about 96.35% efficiency. NCOMM doesn't include 38.4kbps, so I couldn't test it at that rate. It's probably better that way too, considering that it is already almost 15 points behind JR-Comm. Who knows how much slower it would be then the 3600cps that JR-Comm obtains at that rate... Now, all this is really only of value to those with high speed modems. But, the point remains. JR-Comm has the more efficient ZMODEM implementation. -jack-