dougl@westfort.UUCP (Dougras Roos) (05/26/89)
I recently pieced together a 10 MHz XT clone system that I purchased from JDR Microdevices. I purchased a Seagate 30M hard drive and a 1.44 meg 3 1/2" drive (Teac?) to put in the system. I put it all together, and things were working beautifully until I tried to work with the floppy drive. To make the thing work on an XT, JDR includes a device driver ("jdrdisk.sys"). I set that up just right, and the drive *does* work, but it does so at an appalingly slow pace, considerably slower than a 360K drive! I called JDR to ask about this, and they said that there was nothing that they could do about it -- I was told that the driver slows the drive down so it can "cope with the high amount of data on the disk." Now, I've dealt with JDR before, and they've been very good at customer service and technical help, but this sounds a bit fishy. Why would it slow the disk down so much that it runs *slower* than a normal (5 1/4") floppy drive? My computer is as fast or nearly as fast as some low-end AT's, and the AT I use at work with the small disk drive just sings along. Does anyone know about this? Perhaps someone could either second JDR's verdict or point me to a source for a better driver? Thanks, Douglas Luce Modesty International dougl@westfort.UUCP || tut!westfort!dougl || westfort!dougl@tut.cis.ohio-state.edu
harnyo@grad2.cis.upenn.edu (Suwandi Harnyo) (05/26/89)
Well, I think using the high density drives whether it's 1.44 or 1.2 on XT will definitely be slow. First, the disk controller operates in 8 bits slots as compared to 16 bits in the AT. So, even though you may have Turbo Speed, you are still restricted to slower bus. Secondly, the huge amount of data on the high density disk can slow down the access time significantly. The only thing you can do to alleviate the problem without getting an AT is by using cache program, or use program like fastopen in Dos 3.3. Also, if you want to copy the disk to your harddisk, use xcopy program instead of normal dos copy. Xcopy reads as much as data into memory and write it on the target drive. So, you don't have to read and write so many times. All these methods may improve the disk performance. I hope it answers your question. Andy harnyo@grad2.cis.upenn.edu
ray@philmtl.philips.ca (Raymond Dunn) (05/27/89)
In article <11389@netnews.upenn.edu> harnyo@grad2.cis.upenn.edu.UUCP (Suwandi Harnyo) writes: >Well, I think using the high density drives whether it's 1.44 or 1.2 >on XT will definitely be slow. First, the disk controller operates >in 8 bits slots as compared to 16 bits in the AT. So, even though >you may have Turbo Speed, you are still restricted to slower bus. >Secondly, the huge amount of data on the high density disk can slow >down the access time significantly. Sometimes a spade should be called a spade. This is such a time: The above is PURE NONSENSE! Please! Please, don't disseminate this drivel! 16 bit/8 bit issues have *NOTHING* to do with floppy operation which is always done at 8bits, whether on an XT or an AT. The "huge amount of data" has *NO* effect on the "access time": Both 720K drives and 1.44Mb disks have 80 tracks, so that the average seek time is the *SAME*. A 1.44Mb disk has 18 sectors/track as opposed to 9 for a 720K disk. The data transfer rate of a 1.44Mbyte drive is *TWICE* that of a 720K drive, and is *ALWAYS* limited by the sectors/track and rotation speed, and *NOT* the CPU speed (except of course when e.g. disk caching is brought into the picture). grump... grizzle... -- Ray Dunn. | UUCP: ..!uunet!philmtl!ray Philips Electronics Ltd. | TEL : (514) 744-8200 Ext: 2347 600 Dr Frederik Philips Blvd | FAX : (514) 744-6455 St Laurent. Quebec. H4M 2S9 | TLX : 05-824090
gharring@enprt.Wichita.NCR.COM (Gary Harrington) (05/27/89)
I have an AT-class machine with both 5 1/4 and 3 1/2 drives. The 5 1/4 is suitably fast but the 3 1/2 is pitifully slow. Does anyone have a 3 1/2 drive that is actually just as fast as the 5 1/4? Gary Harrington
harnyo@grad2.cis.upenn.edu (Suwandi Harnyo) (05/27/89)
In article <543@philmtl.philips.ca> ray@philmtl.UUCP (Raymond Dunn) writes: >In article <11389@netnews.upenn.edu> harnyo@grad2.cis.upenn.edu.UUCP (Suwandi Harnyo) writes: >>Well, I think using the high density drives whether it's 1.44 or 1.2 >>on XT will definitely be slow. First, the disk controller operates >>in 8 bits slots as compared to 16 bits in the AT. So, even though >>you may have Turbo Speed, you are still restricted to slower bus. >16 bit/8 bit issues have *NOTHING* to do with floppy operation which is always >done at 8bits, whether on an XT or an AT. I am not sure about this. I always thought that having a 16 bits floppy controller on my AT is definitely faster than a 8 but controller. That's why I pay more for the 16 bit controller than the cheaper 8 bit one. Comment ? Andy
karl@ddsw1.MCS.COM (Karl Denninger) (05/29/89)
>Response 15 of 15 (3260) by harnyo at grad2.cis.upenn.edu on Mon 29 May 89 >1:18 >[Suwandi Harnyo] >>16 bit/8 bit issues have *NOTHING* to do with floppy operation which is >>always >>done at 8bits, whether on an XT or an AT. > >I am not sure about this. I always thought that having a 16 bits floppy >controller on my AT is definitely faster than a 8 but controller. That's >why I pay more for the 16 bit controller than the cheaper 8 bit one. Floppy I/O should not be dependant on 8/16-bit controller configuration; the ultimate limiting transfer rate might be greater with the 16-bit controller, but EITHER of them can transfer faster than the 1.44M or 720K drive can read or write. I doubt this really enters the picture at all --- Karl Denninger (karl@ddsw1.MCS.COM, <well-connected>!ddsw1!karl) Public Access Data Line: [+1 312 566-8911], Voice: [+1 312 566-8910] Macro Computer Solutions, Inc. "Quality Solutions at a Fair Price"
ray@philmtl.philips.ca (Raymond Dunn) (05/30/89)
In article <11389@netnews.upenn.edu> harnyo@grad2.cis.upenn.edu.UUCP (Suwandi Harnyo) writes: > [absolute drivel] In article <543@philmtl.philips.ca> I write: > [grumbles about the above drivel] In article <11412@netnews.upenn.edu> harnyo@grad2.cis.upenn.edu.UUCP (Suwandi Harnyo) continues to try to justify the drivel: >I am not sure about this. I always thought that having a 16 bits floppy >controller on my AT is definitely faster than a 8 but controller. That's >why I pay more for the 16 bit controller than the cheaper 8 bit one. >Comment ? Yes, I can see why you are not sure about it even though you always thought it. (:-( (:-( That is what was wrong with your original posting, and I'm not sorry to be given the opportunity to re-iterate this: - DO NOT POST AS FACTS, THAT WHICH YOU ARE NOT SURE ABOUT, EVEN THOUGH "YOU ALWAYS THOUGHT IT" Expressing opinion during a discussion is, of course, another matter - in that case you'd better check your facts even MORE carefully or you'll be REALLY flamed (:-) (:-). Now to respond to your request for comment. Excuse me for being rude, but it only takes 30 seconds prior to posting to verify the validity of FACTS - in the AT technical reference manual, Norton, or your own favourite bible. On an AT, the floppy disk controller is connected to channel 2 of DMA Controller 1. DMA Controller 1 is physically only capable of supporting 8-bit transfers. The system BIOS and most software which directly accesses the floppy, does so via the DMA. And finally, the speed of transfer to a floppy disk is limited by the physical rotation speed of the media, NOT the width of the data bus. A 64-bit bus wouldn't give you faster data transfer to a floppy!! This is the end of the real story. Period. Lets look at where the myth of a 16-bit floppy controller might come from: AT *hard* disk controllers are 16-bit, and many do have embedded floppy controllers. Data is transferred to the hard disks 16-bits at a time, interestingly enough *NOT* using the DMA, but by programmed I/O - it's faster than the DMA on the 286. The floppy controllers on these cards are still 8-bit devices. It *would* be possible to have a floppy controller that could be controlled by 16-bit wide programmed I/O, but it would also have to be capable of being driven in a compatible 8-bit mode. Such a controller card would have to have its own BIOS chip to replace the system BIOS which contains the compatible floppy driver code. I don't know of any standalone floppy controller cards for AT's - that doesn't of course mean that there aren't any. The functionality is nearly always provided on the hard disk controller card. Please name one 16-bit floppy controller card. -- Ray Dunn. | UUCP: ..!uunet!philmtl!ray Philips Electronics Ltd. | TEL : (514) 744-8200 Ext: 2347 600 Dr Frederik Philips Blvd | FAX : (514) 744-6455 St Laurent. Quebec. H4M 2S9 | TLX : 05-824090
tmurphy@wpi.wpi.edu (Tom [Chris] Murphy) (05/30/89)
In article <484@enprt.Wichita.NCR.COM> gharring@enprt.UUCP (Gary Harrington) writes: >I have an AT-class machine with both 5 1/4 and 3 1/2 drives. The 5 1/4 >is suitably fast but the 3 1/2 is pitifully slow. Does anyone have a >3 1/2 drive that is actually just as fast as the 5 1/4? I have both in my AT, and if anything the 3.5 is faster. Tom -- Thomas C. Murphy Worcester Polytechnic Institute CAD Lab Internet: tmurphy@zaphod.wpi.edu tmurphy@wpi.wpi.edu BITNET: TMURPHY@WPI BIX: tmurphy CompuServe: 73766,130 -- Guns don't kill people, people kill people - with guns. --
johne@hpvcfs1.HP.COM (John Eaton) (06/01/89)
<<<< < < I am not sure about this. I always thought that having a 16 bits floppy < controller on my AT is definitely faster than a 8 but controller. That's < why I pay more for the 16 bit controller than the cheaper 8 bit one. < Comment ? < Andy ---------- Sounds like a real market opportunity in the making. We could offer a 16 bit serial card and advertise that it receives 9600 baud data in half the time as an ordinary 8 bit card. Charge a 50% premium and laugh all the way to the bank. Anyone want in on the ground floor? John Eaton !hpvcfs1!johne