rreiner@yunexus.YorkU.CA (Richard Reiner) (03/12/91)
lam@hyper.hyper.com (Edmund C. Lam) writes: >The drawback with IDE comes from the fact that the interface limits >your data transfer rates to an observed maximum of 400K/s. >The drive might be quick on average access time, but IDE drives >suffer from low transfer rates. As a generalization, this is false. I have observed 950 Kb/sec from IDE drives on ISA 386-33 machines. However, it is no doubt true that higher data transfer rates can be obtained on ISA with bus-mastering DMA techniques, such as those used by the better SCSI controllers, than with the PIO approach used by IDE. But the limits of PIO are not nearly as low as Lam suggests. //richard
davidg%aegis.or.jp@kyoto-u.ac.jp (Dave McLane) (03/13/91)
rreiner@yunexus.YorkU.CA (Richard Reiner) writes: > lam@hyper.hyper.com (Edmund C. Lam) writes: > > >The drawback with IDE comes from the fact that the interface limits > >your data transfer rates to an observed maximum of 400K/s. > >The drive might be quick on average access time, but IDE drives > >suffer from low transfer rates. > > As a generalization, this is false. I have observed 950 Kb/sec from > IDE drives on ISA 386-33 machines. For what it's worth, I spent some time investigating whether to use "IDE" or "ESDI" before I got my UNIX box (Dell 333D with 320 MB ESDI). The conclusion I came to is that one cannot make any kind of intelligent decision based upon the names; instead you need to know both the average seek time *and* the transfer rate. My simple minded comparison test to concatenate a 96 K file 8 times to make one 768 K file took the following times: 80386 SX 16 Mhz with 40 MB IDE 7.5 Mbps controller: 10 sec 80386 SX 16 Mhz with 190 MB IDE 14.8 Mbps controller: 4 sec 80386 33 33 Mhz with 320 MB ESDI 20 Mbps controller: 2 sec Kinda makes sense if you look at the Mbps, doesn't it? --Dave