HIgh Speed SIO on IDE+ v2

[MrFSL]

These changes make SDX on IDE+ much nicer than running SDX on a cart. So much fun.

 

 

[drac030]

If still feeling brave, here is a next version to test:

bios17d.arc

[MrFSL]

 

 

 

I will try it tonight I am sure!!!!

 

[MrFSL]

So with IDE+ SIO off --- we are about the same. Index 6 works fine. 5 seems ok too. 4 isn't reliable.

 

BUT!!!!!! with IDE+ SIO on --- ummmm..... wow. All the way to Index 0 seems stable. So stable in fact that I threw extra testing at it. To include loading yoomp from SDX (log included) and other binaries.

 

Nice work. I simple thought that IDE+ SIO should perform better than SDX SIO... but you sir... you have managed to max things out. What did you change this time?

IDE_HSIO_OFF.zip IDE_HSIO_ON.zip

[drac030]

5 hours ago, MrFSL said:

What did you change this time?

The same stuff exactly as before: shortened the setup time between the consecutive stages of the SIO exchange. Many thanks for the tests.

 

The "fun" point is that here it nothing changed, on my "main" Atari the max. SIO speed is still 8 as it always has been...

 

5 hours ago, MrFSL said:

So with IDE+ SIO off --- we are about the same

Yes, because SDX is the same. IDE+ SIO is a separate entity residing on a separate ROM together with the HDD driver and such.

Edited November 6, 2020 by drac030

[MrFSL]

Well, this is unfortunate for you. You make it work,  but cannot enjoy the fruits of your labor.  

 

At least your changes haven't made things worse. Perhaps,  new capacitors or no capacitors as suggested earlier. For me, if I remove the IDE÷ and use just a HIAS patched OS ROM I could achieve index 0. This was my indication that it might be possible with the IDE+.

 

Also, i had no idea how the device switches between SIO methods. I expected no change for SDX SIO, but couldn't be sure there wasn't a connection in software so i was testing both ON and OFF. Now I know.

 

Lastly, my pleasure to test. Why are any of us fooling around with these old ataris if not too tinker and marvel at them as they do those things we could have not imagined 30+ years ago. 

 

Thank you, and all the other guys who do what you do here at atariage.

[+Larry]

On 11/4/2020 at 2:03 PM, drac030 said:

SDX and its SIO drivers is a separate question. SIO.SYS is slow and it has been known for ages. Worse, I am not sure if this is fully fixable, but the current betas (which some have) do contain a SIO.SYS driver which should be faster.

 

As for IDE+, the thing I could do is to shorten the intervals the procedure is spending in setting up the consecutive phases of the transfer. I can already see under the emulator that if the data block is sent by the external device immediately after the 'C' acknowledge (at divisor 0), the receiver may have a problem with catching up. I will see what can be done, unfortunately the only means of testing this for me is Altirra.

 

Is there an general explanation about why SIO.SYS is slow and may not be fixable?  Just curious.

I rarely use anything faster than UltraSpeed on any device, but I realize that some folks want to go faster.

-Larry

 

 

[MrFSL]

Here ya go Larry. This was posted recently in a different thread.