Return to BSD News archive
Newsgroups: comp.os.386bsd.questions Path: sserve!newshost.anu.edu.au!munnari.oz.au!ihnp4.ucsd.edu!ames!cnn.nas.nasa.gov!wilbur.nas.nasa.gov!tweten From: tweten@wilbur.nas.nasa.gov (Dave Tweten) Subject: Re: 1740 adaptec questions Message-ID: <CMEzFn.MG8@nas.nasa.gov> Sender: news@nas.nasa.gov (News Administrator) Nntp-Posting-Host: wilbur.nas.nasa.gov Organization: NAS Systems Division, NASA Ames References: <gokingsCMDvqr.Cq@netcom.com> Date: Wed, 9 Mar 1994 20:26:11 GMT Lines: 49 In article <gokingsCMDvqr.Cq@netcom.com> gokings@netcom.com (Russell Marrash) writes: >I am installing an old 1740 EISA controller in my system and I have a few >questions. > > 1. I am running FreeBSD 1.0, will I have to re-compile the kernel or > re-install FreeBSD. The reason I ask is because my pc hangs now > while trying to boot FreeBSD. My 1740 has always worked fine with the distributed kernels oriented toward the 1742, the GENERICAH systems. One concern with both the 1742 and the 1740 is the version of the boot records you are using. If you still have old 386BSD boot records on your hard disk you will only be able to use your 17xx in 15xx compatibility mode. With the new boot records, you can configure your disklabel using "BIOS geometry" (the 1 megabyte pseudo-cylinder fake geometry favored by DOS) and run your 17xx in either compatibility or enhanced modes. > 2. After installing the 1740, I ran some benchmark programs, ie checkit > and sysinfo. The transfer rate is the same as it was with my 1542, > regardless of whether the 1740 is in enhanced or standard mode. That's not surprising. Unless you have a very high bandwidth device (approaching 5 MB/s asynchronous transfer) the device, and not the choice of ISA or EISA host bus adapter, is the rate-limiting factor for any single device. Also, the 1740's mode has no effect on its performance. The "enhancements" of enhanced mode are the ability to tailor things like asynchronous transfer negotiation, BIOS support, start commands, etc. on a target-by-target basis. In compatibility mode, your choices apply to all targets uniformly. Slow (worst case, synchronous transfer) devices use up a much greater portion of the SCSI bus bandwidth than their meager performance would suggest. As a result, multi-device aggregate performance measurements can be throttled by a single device. As an example, consider two kinds of SCSI asynchronous transfer disks. One transfers 1.5 megabytes per second at the head and does asynchronous SCSI transfers at 3.3 megabytes per second. The other transfers 2.2 megabytes per second at the head and does asynchronous SCSI transfers at 10 megabytes per second. I own one of each. If I were to populate a SCSI bus with enough disks of one type to fully load the bus, the effective bus bandwidth would be 3 megabytes per second at two disks in the first case, and 8.8 megabytes per second at four disks in the second case. -- --------------------------------------------------------------------------- Dave Tweten tweten@nas.nasa.gov NASA Ames Research Center, M/S 258-5 (415) 604-4416 Moffett Field, CA 94035-1000 FAX: (415) 604-4377