On Aug 23, 2004, at 12:52 PM, Mat Hayashibara wrote:
At 04:09 AM 8/22/2004 +0200, you wrote:There is a FireWire standard called 1394b which allows transfer speed up to
800Mbps. Is there a CF card reader that can handle this speed, or will any
FireWire CF reader automatically support this transfer rate?
(chuckling...) The bottleneck isn't the 1394 interface itself, it's how the data gets to the processor (eg, PCI bus or hard drive interfaces like ATA133). A lot of these new interfaces like Firewire 800, USB2 (480MB/s) and the upcoming UWB (a wireless version of USB2!) have all kinds of speed, but the ability of current PC technology to keep up isn't there. As it is, a buffer will fill up and firmware will halt the transaction till the PC can handle the data. PCI-X is going to help, but it ain't in widespread use yet...
Er, in this case, the bottleneck is in the CF card itself--none of them are faster then 10 MB/sec, which doesn't touch the ~35 MB/sec limit of traditional firewire.
Traditional PCI is good for around 100-120 MB/sec (132 on paper, but there's a lot of overhead). PCI-X is good for around 8x the speed, but you won't find it on cheap desktops, just servers, workstations, and the higher-end models of Apple's PowerMac G5. PCI Express (also known as PCI-E) is starting to show up on cheaper PC, but it'll take a couple years before it's really common.
One thing that a lot of people miss--most networking and communications standards (Ethernet, Firewire, USB) have speeds quoted in mega*bits* per second (abbreviated Mb/sec or Mbps), which almost everything else in computers talks about mega*bytes* per second (abbreviated MB/sec or MBps). There's a factor of 8 difference between 800 MBps and 800 Mbps. So, Firewire 800 is a little bit slower then normal PCI. USB 2 is around half of the speed of PCI. ATA133 is roughly the same speed as PCI, or faster then Firewire 800, even though 133<800. It's not really any more complex then the math needed to get your exposures right, but that doesn't make it simple.
Scott
* **** ******* *********************************************************** * For list instructions, including unsubscribe, see: * http://www.a1.nl/phomepag/markerink/eos_list.htm ***********************************************************
