Welcome, Guest. Please login or register.

Author Topic: FastATA 4000 MK-V released  (Read 6768 times)

Description:

0 Members and 1 Guest are viewing this topic.

Offline mboehmer_e3b

  • Sr. Member
  • ****
  • Join Date: Aug 2002
  • Posts: 312
    • Show all replies
    • http://www.e3b.de/usb/
Re: FastATA 4000 MK-V released
« on: August 21, 2008, 11:29:12 PM »
Quote

I forgot this think works on ZorroIII :). I can't tell from "experience", maybe someone?


Here. You can make about 15.3MB/s on Zorro III if you are nasty and occupy the bus with 100% load, leaving no bandwidth for any other device sitting there (like a gfx card), and eating up 100% CPU time (which is not what multitasking likes).
I assumed a single cycle time of 250ns (including the small delay between cycles).

You can measure it yourself with "bustest" (which is amazingly accurate when comparing to a scope or logic analyzer).

Some values from my setups here:

A4000D, CS MKII 060@50MHz:

DENEB: 7.5MB/s read, 11.8MB/s write
CV64:  5.4MB/s read,  6.7MB/s write

Quote

BTW. Personally I'm waiting for something truly, really etc. new.


It is. At least the EPROM and the MACH chips on it are "new", the rest of the board looks like the 1999 version of Fast ATA 4000, compare yourself in amiga.resource.cx) :-)

Michael
 

Offline mboehmer_e3b

  • Sr. Member
  • ****
  • Join Date: Aug 2002
  • Posts: 312
    • Show all replies
    • http://www.e3b.de/usb/
Re: FastATA 4000 MK-V released
« Reply #1 on: August 22, 2008, 07:37:21 AM »
Quote

Its like USB2.0 which can make up to 480 Mbit/s but in real life you will get only about 320 Mbit/s or less even on modern PCs which makes it slower than FireWire 400.


No, sorry. Both are serial busses, and both "suffer" from the same bandwidth eaters: 8b/10b coding, bit stuffing, sync frames, protocol overheads.

So if you compare both serial busses you need to give both a reduction of 33%, to be fair :-)

Michael