Date: Mon, 5 Feb 90 16:34:39 EST From: Dwight D. McKay (The Moderator) Reply-To: Suns-at-Home@orchestra.ecn.purdue.edu Subject: Suns-at-Home Digest V3 #5 To: Suns-at-Home-List Suns-at-Home Digest Mon, 5 Feb 90 Volume 3 : Issue 5 Today's Topics: SCSI disk performance Sun 2/120 for sale Sun 2/50 Diagnostic ROMs Sun2 mem? errors +--------------------------------------------------------------------+ | Submissions: suns-at-home \ @orchestra.ecn.purdue.edu | | Requests: suns-at-home-request > -- or -- | | Archives: suns-at-home-archives / ...rutgers!pur-ee!... | +--------------------------------------------------------------------+ ---------------------------------------------------------------------- Date: Tue, 30 Jan 90 16:50:30 GMT From: dutrun!dutetvg!pieter@relay.EU.net (Pieter Venemans) Subject: SCSI disk performance To: Suns-at-Home@orchestra.ecn.purdue.edu Hello folks, Recently, I bought a new ST506 disk for my SUN-2/50. I connected it to the Adaptec ACB4000 controller inside my shoebox and it works just fine. But, having an empty disk in my system now, it is a good time to play around with interleaving, tunefs(8) etc. You won't do that so lightly when disks are in use. Reading a 1Mb file from an empty file system takes around 20 seconds, so the transfer rate is 50 Kb/s. Quite disappointing: an old PC/XT is often twice as fast. 50 Kb/s means that the system can read 2 sectors per revolution. The Adaptec controller has a buffer of 1 Kb (2 sectors), so that sounds reasonable for a non-interleaved disk and a not-so-fast machine. As far as I understand, increasing the interleave factor should improve the performance in such a situation. But, whatever interleave factor I specify in /stand/diag, nothing changes. When I simulate interleaving by seeking in /dev/rsd1a to the (simulated) interleave positions while reading sectors, I get a transfer rate around 90 Kb/s, so a proper interleave factor could really speed up things... I know that it is impossible for the device driver to allocate blocks and sectors optimally due to the invisible bad sector slipping on SCSI disks, but doesn't it even try? Why does diag ask for an interleave factor when it is used neither by the formatter nor by the device driver? Is this a bug? Or did *I* miss something ? Do you know how to increase the performance of an SCSI disk, or should I just be happy with this 50 Kb/s? [ My configuration: SUN-2/50, 2 Mb, SCSI (501-1045), SunOS-3.5, ] [ Adaptec ACB4000, Miniscribe 3085 (80Mb, ST-506, 22 ms av.seek, ] [ 17 sect/track, no defects in the partition I used). ] --- Pieter H.A. Venemans --- pieter@et.tudelft.nl ------------------------------ Date: Sun, 29 Jan 89 21:51:03 -0500 From: lemming!cspencer@ursa-major.SPDCC.COM Subject: Sun 2/120 for sale To: spdcc!orchestra.ecn.purdue.edu!suns-at-home Sun 2/120 Multibus workstation for sale. 5 Mb RAM Micropolis 1325 71Mb hard disk 20 Mb cartridge tape drive Thick wire ethernet SCSI bus Adaptec ACB400 SCSI<==>ST506 Adapter Sky floating point board 6 serial ports 1152x900 19" Mono Display Sun OS release 3.5 uucp X11 R3 Window System $2,300 takes it home. I don't want to get involved with shipping so I'm looking for a buyer in the greater Boston area. Cliff Spencer lemming!cspencer days 617-225-0500 x565 eves till 9:30 617-641-1822 ------------------------------ Date: Sat, 3 Feb 1990 16:36:54 PST From: Ismail Dalgic Subject: Sun 2/50 Diagnostic ROMs To: Suns-at-Home@orchestra.ecn.purdue.edu I'm trying to run the Sun 2/50 diagnostic ROMs, but so far I couldn't successfully be able to make the system communicate with the RS-232 terminal (I used an h-19 for this purpose, if that matters). I first tried the suggested connection in the field service manual, which is cross-connecting lines 2 (TXD) and 3 (RXD) (This implies that the Sun RS-232 connector is configured as DTE), looping back lines 5 (CTS) and 6 (DSR), and connecting together line 7 (Ground). The screen of the terminal remained completely blank. Of course, I set the speed of the terminal to 9600 baud and it was in full-duplex mode, so this is not the problem. Then I tried the standard null-modem connection, and also connecting lines 4 (RTS) and 5, and 6(DSR),8(DCD) and 20(DTR) together (i.e., no handshake), with the same unsuccessful result. In fact it doesn't make sense to me to connect lines 5 and 6 together, which are both inputs (CTS and DSR). Can you help me? --Ismail Dalgic dalgic@cs.stanford.edu ------------------------------ Date: Tue, 30 Jan 90 10:25:25 PST From: Joe Pruett Subject: Sun2 mem? errors To: Suns-at-Home%orchestra.ecn.purdue.edu@RELAY.CS.NET I've just added a 4meg board to bring my 2/120 up to 7 meg. Now I keep getting parity errors that cannot be found (the kernel says "transient?"). And then when I try to reboot, I get memory protection violation messages. I'm going to check voltages later this week, but I was wondering if people had any ideas. Previously, I had 5 1meg boards running with no problems at all. I've tried running with and without the P2 terminator board. My CPU has rev N PROMS and thinks that it's a 1/100U or 1/150U with VT100 keyboard (which it doesn't currently have). In the past I sometimes would get kernel crashes about vm_hat during initial boot, but the reboot thereafter would always work. PS: Mister moderator, any chance we could change suns-at-home to more of a mail reflector instead of a digest? The turnaround time can be quite frustrating when you have a dead machine. sun-nets and sun-managers seem to work quite well as mail reflectors, and the turnaround time is usually hours instead of days or weeks. Please take this as a constructive suggestion because I find suns-at-home quite useful. [ I understand your frustration. Noone likes to have a dead Sun on his hands ] [ and have to sit waiting for assistance. We did discussed switching over to ] [ a mail reflector for SaH last year and the response at that time was almost ] [ 10 to 1 in favour of staying a digest. Traffic on the list has been ] [ sufficient to allow weekly publishing which seems to be just fine for most ] [ things. ] [ If any of you out there in netland have a strong feeling that SaH ought to ] [ switch over to a mail-reflector, please let me know by private mail to: ] [ mckay@harbor.ecn.purdue.edu (or rutgers!pur-ee!mckay). If I get more then ] [ two or three responses I'll ask for a vote in the next issue of the digest. ] [ --ddm ] ------------------------------ End of Suns-at-Home Digest ******************************