hdrw@ibmpcug.co.uk (Howard Winter) (04/01/91)
I have been given a number of recommendations to go for the Conner CP3204 200M IDE disk drive as a good way to upgrade my machine. Before I order it, a couple of questions: 1. I've seen it advertised as 15mSec in US adverts, but 19mSec from the only English firm I've seen advertising it. What is the correct access time, and is this the TRUE speed, or only apparent due to the controller hiding the actual head/track/sector configuration ? 2. What is the achievable data-transfer rate ? I have tripped up on this before, when I bought an ST296N, only to find it cannot be formatted at better than 3:1 in my machine, so limiting it to 300K/Sec. I am hoping for about 1M/Sec from the Conner. Is this possible ? Are there any implications (such as processor or AT Bus bottleneck, or anything else like that) ? 3. Am I really getting the best value for money ? How does this drive compare with the ST2139A, which has similar basic spec. (~200M, 15mS) ? 4. Has anyone got Conner's phone/fax number, so I can request data sheets from them (I have been going on adverts and dealers' words so far) ? Thanks in anticipation, Howard. -- Automatic Disclaimer: The views expressed above are those of the author alone and may not represent the views of the IBM PC User Group. -- hdrw@ibmpcug.Co.UK Howard Winter 0W21' 51N43'
john@jwt.UUCP (John Temples) (04/03/91)
In article <1991Mar31.162532.9509@ibmpcug.co.uk> hdrw@ibmpcug.co.uk (Howard Winter) writes: >2. What is the achievable data-transfer rate ? I've used the CP3204 under both DOS and UNIX, and the transfer rate was pathetic. I/O benchmarks (ones that do more than just read the same 64K over and over) showed between 80 and 150 KB/sec read and write rates. Even worse than 1:1 ST-506/MFM. I suspect the controller on the drive isn't fast enough to run at 1:1 interleave, but I don't have the appropriate software to reformat it and find out. I would love to hear about someone's experiences after reformatting one of these to a looser interleave. >I am hoping for about 1M/Sec from the Conner. Is this possible ? Sure, if your only application is Coretest. I'm starting to suspect people have been wooed by the small, quiet packaging of IDE, its fast seek times, cheap interface cards, easy cabling, and misleading Coretest results. I'm starting to suspect it's a well packaged, well marketed performance dog. I bought a 15 MHz ESDI drive, and it absolutely blows the Conner away. I won't recommend another IDE drive until I see them show me respectable transfer rates. -- John W. Temples -- john@jwt.UUCP (uunet!jwt!john)
phr@lightning.Berkeley.EDU (Paul Rubin) (04/03/91)
In article <1991Apr3.035310.11481@jwt.UUCP> john@jwt.UUCP (John Temples) writes: Path: agate!bionet!uwm.edu!zaphod.mps.ohio-state.edu!samsung!uunet!jwt!john From: john@jwt.UUCP (John Temples) Newsgroups: comp.sys.ibm.pc.hardware,connect.audit Date: 3 Apr 91 03:53:10 GMT References: <1991Mar31.162532.9509@ibmpcug.co.uk> Organization: Private System -- Orlando, FL Lines: 25 In article <1991Mar31.162532.9509@ibmpcug.co.uk> hdrw@ibmpcug.co.uk (Howard Winter) writes: >2. What is the achievable data-transfer rate ? I've used the CP3204 under both DOS and UNIX, and the transfer rate was pathetic. I/O benchmarks (ones that do more than just read the same 64K over and over) showed between 80 and 150 KB/sec read and write rates. Even worse than 1:1 ST-506/MFM. I suspect the controller on the drive isn't fast enough to run at 1:1 interleave, but I don't have the appropriate software to reformat it and find out. ... Maybe your cpu isn't fast enough. I get 645k/sec from my CP3204 (1:1 interleave) on a 25MHz cacheless 386dx (Norton sysinfo benchmark). It seems pretty fast in practice, too, subjectively, but I don't have any other measurements. A friend of mine has measured 1.1MB/sec from a similar (non-Conner) 200MB drive on a 486-33 system. The PC-AT bus is 16 bits wide and runs at 8 MHz (even for faster cpus), so 2 MB/sec is the upper bound on the transfer speed for any 16-bit controller. I am very happy with my IDE CP3204.
tinyguy@cs.mcgill.ca (Yeo-Hoon BAE) (04/03/91)
In article <1991Apr3.035310.11481@jwt.UUCP> john@jwt.UUCP (John Temples) writes: >In article <1991Mar31.162532.9509@ibmpcug.co.uk> hdrw@ibmpcug.co.uk (Howard Winter) writes: >>2. What is the achievable data-transfer rate ? > >I've used the CP3204 under both DOS and UNIX, and the transfer rate was >looser interleave. > >>I am hoping for about 1M/Sec from the Conner. Is this possible ? > >-- >John W. Temples -- john@jwt.UUCP (uunet!jwt!john) I have used the 105Meg version of Conners SCSI drive, and although I'm running it on Amiga with DMA controller, it gives roughly 600-750k/s depending on the environment... I'm sure 200Megger is faster, given the right controller... -TG
john@jwt.UUCP (John Temples) (04/04/91)
In article <PHR.91Apr3023701@lightning.Berkeley.EDU> phr@lightning.Berkeley.EDU (Paul Rubin) writes: >Maybe your cpu isn't fast enough. I tried it on a non-cached 386/20 and a cached 386/33 with the same results. >I get 645k/sec from my CP3204 (Norton sysinfo benchmark). What exactly does the Norton benchmark do? Did you have any software cacheing enabled when you ran it? I do find it interesting that the result obtained by Norton is nearly 50% slower than what Coretest quotes, though. >A friend of mine has measured 1.1MB/sec from a similar (non-Conner) >200MB drive on a 486-33 system. What benchmark was used to measure this transfer rate? -- John W. Temples -- john@jwt.UUCP (uunet!jwt!john)
john@jwt.UUCP (John Temples) (04/04/91)
In article <1991Apr3.113533.1777@cs.mcgill.ca> tinyguy@cs.mcgill.ca (Yeo-Hoon BAE) writes: >I have used the 105Meg version of Conners SCSI drive, and although I'm >running it on Amiga with DMA controller, it gives roughly 600-750k/s How was this transfer rate measured? In any event, it can't be assumed that the transfer rate of a SCSI version of the drive would have any bearing whatsoever on the IDE version. -- John W. Temples -- john@jwt.UUCP (uunet!jwt!john)
tinyguy@cs.mcgill.ca (Yeo-Hoon BAE) (04/04/91)
In article <1991Apr4.040715.13946@jwt.UUCP> john@jwt.UUCP (John Temples) writes: >In article <1991Apr3.113533.1777@cs.mcgill.ca> tinyguy@cs.mcgill.ca (Yeo-Hoon BAE) writes: >>I have used the 105Meg version of Conners SCSI drive, and although I'm >>running it on Amiga with DMA controller, it gives roughly 600-750k/s > >How was this transfer rate measured? In any event, it can't be assumed >that the transfer rate of a SCSI version of the drive would have any >bearing whatsoever on the IDE version. >-- >John W. Temples -- john@jwt.UUCP (uunet!jwt!john) The transfer rate was measured with a program called DiskSpeed which is a common program used on Amiga System to measure HD speed. It gives read/write speed as well as no. of seeks/create/delete and scans per second. In general, it seems pretty accurate, and I sware, I didn't use any software cache program... Infact, cache shouldn't affect the overall tansfer speed anyway, it does affect the no. of seeks/delete etc though... -TG
phr@tsunami.Berkeley.EDU (Paul Rubin) (04/05/91)
In article <1991Apr4.040424.13848@jwt.UUCP> john@jwt.UUCP (John Temples) writes: >In article <PHR.91Apr3023701@lightning.Berkeley.EDU> phr@lightning.Berkeley.EDU (Paul Rubin) writes: > >>I get 645k/sec from my CP3204 (Norton sysinfo benchmark). > >What exactly does the Norton benchmark do? Did you have any software >cacheing enabled when you ran it? I do find it interesting that the >result obtained by Norton is nearly 50% slower than what Coretest >quotes, though. I don't know what the Norton benchmark does, but the disk makes a lot of noise while the benchmark is running, so it can't be simply reading from the on-drive cache repeatedly. I did not have any software caching enabled. Throughput on a Unix system using this same drive is comparable: around 150k/sec to copy one large file to another using "dd". Because both the old file is being read and the new one being written, and because the Unix file system imposes some overhead and because the files are not necessarily contiguous on the disk, I think this is consistent with the 645k/sec transfer speed from the disk. >>A friend of mine has measured 1.1MB/sec from a similar (non-Conner) >>200MB drive on a 486-33 system. > >What benchmark was used to measure this transfer rate? Norton sysinfo again.