thats not a standard term, i still think it means the cpu/dma controller on the mobo will be doing the task;

PIO generally means "Programmed I/O", which doesn't use DMA at all. The last pages of the manual simply refer to the jumper as turning DMA on or off. I don't know if the AHA-152x does bus mastering (1st party DMA), it might rely on the motherboard's DMA controller (3rd party DMA), which isn't all that much faster then PIO.

The AHA-154x on the other hand is a true bus mastering card and would likely be the faster choice. Its also modern enough to have a SCSISelect ROM on it.

NJRoadfan wrote:

I don't know if the AHA-152x does bus mastering (1st party DMA), it might rely on the motherboard's DMA controller (3rd party DMA), which isn't all that much faster then PIO.

but still when the dma controller does his thing (can this even be overclocked on the mobo?) the cpu is thoeretically free to perform tasks that do not require memory access; although i'm not sure if this is in reality of any help.

The newer 1522B and 1520B controller has the SCSISelect BIOS and an option to set the DMA channel, yet the transfer rate is fixed on PIO mode. On this older 1522 card, if you set the DT jumper, the text on the screen at boot claims the transfer mode is DMA. Perhaps this is a feature Adaptec removed from the newer 1522B units?

At any rate, the older 1522 card hangs on trying to query the SCSI HDD. I checked all the jumpers. I'm not sure why this is as the same setup works fine on the newer 1520B card I have. Perhaps the issue is related to using the terminator jumpers on the HDD rather than a terminated cable?

I also have a 1540CP card with true DMA bus mastering. This card doesn't work on some 386 boards in combination with the DLC/SXL series of CPUs.

Ideally, I would like to benchmark the 1522 in "2nd-party DMA" and PIO, and compare the results to the newer 1520B and the 1540CP. Anyone run through the various DOS HDD benchmarks to see which are the most sensitive in benchmarks? It is likely that these 3 controllers will be on par, so a very sensitive benchmark would be ideal.

The manual for the AHA-1522A has an interesting note about DMA.

http://ftp.uma.es/Drivers/SCSI/ADAPTEC/AHA1UIDE.PDF

"DMA transfer mode is not supported under DOS or Windows. It is supported under OS/2 on EISA-bus computers."

Its possible the 1522 has the same limitation. The card likely takes advantage of the superior DMA capabilities of EISA bus machines (the better DMA controller is likely accessible to ISA devices) and only OS/2 had a driver to run the card in that mode.

Why would anyone with an EISA machine want to use a crappy ISA SCSI controller? Makes no sense.

No clue, perhaps the card was released before Adaptec readied the AHA-1740 series.

Curious about ISA cards in EISA systems, I found the following:

Source: http://www.ece.ufrgs.br/~fetter/eng04476/data … ts/eisabook.pdf

EISA DMA channels may be programmed to use one of four DMA bus cycle types when transferring data between the I/O device and memory. The default DMA bus cycle type, ISA-compatible, delivers a higher data transfer rate than ISA-compatible computers. The improvement is the result of EISA's faster bus arbitration and requires no hardware or software modifications to ISA-compatible DMA devices. Type A and B cycles are EISA modes that permit some ISA-compatible DMA devices to achieve higher performance. The burst DMA (Type C) bus cycle type is the highest-performance DMA bus cycle and is only available to DMA devices designed specifically for EISA burst.

Page 103 shows the difference between DMA Type A and B cycles.

this has a separate floppy connector. is that a standard thing or do scsi floppies still connect to a different channel than all the other drives? would this be any faster/ yield less cpu load than standard floppy controller?

The floppy controller is nothing special, same as what you get on an I/O card or motherboard.