Welcome, Guest. Please login or register.

Author Topic: New drivers for FastATA 1200 and online speed tests  (Read 2759 times)

Description:

0 Members and 1 Guest are viewing this topic.

Offline DJSTopic starter

  • Newbie
  • *
  • Join Date: Apr 2003
  • Posts: 36
    • Show only replies by DJS
    • http://www.elbox.com
New drivers for FastATA 1200 and online speed tests
« on: June 17, 2004, 03:10:45 AM »
ELBOX COMPUTER has released today the NEW drivers for FastATA 1200 (Power Flyer 1200, Winner 1200 High-Speed) controllers.

FastATA 1200

New software is available free of charge for all registered users of FastATA 1200.

The update includes new versions of:
• ATA3.driver
• ATA3Prefs
• AllegroCDFS

Changes:
ATA3.driver ver.7.0
- compatiblity with the newest big hard drive geometries (support up to 2TB)
- write speed improvement

ATA3Prefs ver.7.0
- changes in the SPLIT/NOSPLIT detection algoritm

AllegroCDFS ver.3.4
- improvement in reading CDs with surface defects

Speed tests of the EIDE hard drives connected to the FastATA 1200 MK-II controller have been published in the News | Product Reviews section.

Tests were done with the most popular A1200 turbo cards: Apollo 1240, Blizzard 1230, Blizzard 1260 and BlizzardPPC.

If you own an old version of the FastATA controller, you can upgrade it now with the newest FastATA 1200 MK-II firmware using a DIY Upgrade Pack.
 

Offline SHADES

  • Sr. Member
  • ****
  • Join Date: Apr 2002
  • Posts: 355
  • Country: au
    • Show only replies by SHADES
Re: New drivers for FastATA 1200 and online speed tests
« Reply #1 on: June 18, 2004, 08:36:34 AM »
Is there some reason as to why Hardware manufactures insist on using PIO modes? The Zorro III slots on AMIGA computers are quite capeable of using DMA transfers. Is there some limitation I have overlooked or am I correct in my assumption that it's just too costly to re-design and make another controller.

PIO modes for everyone who is not sure on what it stands for are Programmed I/O  or Input/Output.
This is a technique whereby the system CPU and support hardware directly control the transfer of data between the system and the hard disk. There are several different speeds of programmed I/O None of which are all that fast mind you.

 First of all, the PIO modes are defined in terms of their cycle time, representing how many nanoseconds it takes for each transfer to occur. The maximum transfer rate is the reciprocal of the cycle time, doubled because the IDE/ATA interface is two bytes (16 bits) wide. There was also talk of a "PIO mode 5", which does not really exist and was never implemented in any IDE/ATA hard disks. Apparently, at one point some discussion occurred about creating a faster PIO mode, which was tentatively called "PIO mode 5". This mode was to support a transfer rate of 22.2 MB/s, but it was never implemented (probably because the much faster 33 MB/s Ultra DMA mode 2 was on the horizon). Some motherboard manufacturers made a point of providing early support for this proposed mode in their BIOS setup programs, so you may occasionally see it mentioned.

But there you have it. Even the amazingly ancient ATA Ultra DMA mode 2 is faster than the ambiquious PIO 5 and guess what? As the name suggests, the DMA modes do not involve the CPU of your computer, they Directly Access the computers Memory DMA (Direct Memory Access)


Oh, for a comparison??


PIO Mode 4
Mode 4 -< Industry name
120 -< nano seconds (Cycle time)
16.7 -< 16.7 MB/s(Transfer rate) Dep on CPU load!
ATA-2 -< Name of mode.
 
This next bit could pose a potental problem on Zorro 3 because of bus width, but it's a comparison none the less and you could use other (slower) modes that still don't use the CPU.
LIKE Multiword DMA !!
Mode 2 -< Industry name
120 -< nano seconds (Cycle time)
16.7 -< 16.7 MB/s(Transfer rate)
ATA-2 -< Name of mode.
 
CPU independant!

Then ther's the beast

Ultra DMA 5
Mode 5 -< Industry name
40 -< nano seconds (Cycle time)
100.0 -< 100 MB/sec (Transfer rate)
ATA/ATAPI-6 -< Name of mode.

Just wondering.........
It's not the question, that is the problem, it is the problem, that is the question.
 

Offline mikeymike

  • Hero Member
  • *****
  • Join Date: Nov 2002
  • Posts: 3420
  • Country: 00
    • Show only replies by mikeymike
Re: New drivers for FastATA 1200 and online speed tests
« Reply #2 on: June 18, 2004, 11:45:45 AM »
DMA transfers aren't "CPU independent", but they do make something like 99% better use of CPU time than PIO modes do.
 

Offline odin

  • Colonization had Galleons
  • Hero Member
  • *****
  • Join Date: Feb 2002
  • Posts: 6796
    • Show only replies by odin
Re: New drivers for FastATA 1200 and online speed tests
« Reply #3 on: June 18, 2004, 11:47:58 AM »
Correct me if I'm wrong, but the A1200 stuff is all incapable of DMA. Only (harddisk)controllers on processorboards in the trapdoorslot will be able to provide that.

(Then again you only mentioned Z3).

Offline SHADES

  • Sr. Member
  • ****
  • Join Date: Apr 2002
  • Posts: 355
  • Country: au
    • Show only replies by SHADES
Re: New drivers for FastATA 1200 and online speed tests
« Reply #4 on: June 21, 2004, 11:54:24 PM »
@mikeymike
True, you do still need to setup the word transfer. CPU independant was probably a silly choice. I was just trying to show the massive release of CPU time by using DMA Vs the PIO transfer but yes it does use some CPU, just very minimal.
It's not the question, that is the problem, it is the problem, that is the question.
 

Offline SHADES

  • Sr. Member
  • ****
  • Join Date: Apr 2002
  • Posts: 355
  • Country: au
    • Show only replies by SHADES
Re: New drivers for FastATA 1200 and online speed tests
« Reply #5 on: June 21, 2004, 11:59:41 PM »
@odin
Yes, I specificaly chose ZorroIII because I am aware of it's specs. I was not aware that the 1200 couldn't do DMA. I was sure it could although it was through the trapdoor expansion which is the way most 1200's are expanded. Inbfact I thought the 1200's trapdoor was sort of like a zorro bus type expansion. I remember reading of the similarites somewhere although never really looked up the specs for it, I'm sure it had the ability to do DMA. Could be wrong though.

Even so, I think HW manufactures should impliment more use of the AMIGA's hardware. These now antiquated systems already lack CPU power, trying to save this already limited cycle CPUs for other uses other than hard disk transfer just seems sensible.
It's not the question, that is the problem, it is the problem, that is the question.