64NIC+ in fast Mode
Guest
|
Post by 64NIC+ in fast Mode on Jan 20, 2015 6:07:22 GMT
I have learned since there is sadly very little network stuff that has been written for native 128 mode, but what I could find CONTIKI 128 does work with the 64NIC+ for the 128 in native mode, but only if the computer is in SLOW mode, FAST mode and nothing seems to work.... Not being a hardware guy I don't have any clue as to why this is the case, does anyone know what might be the culprit and if a fix/hack could be done to make it work in Fast mode properly?
|
|
|
Post by VDC 8x2 on Jan 20, 2015 20:58:54 GMT
Could it be timing issues?
|
|
|
Post by hydrophilic on Jan 22, 2015 5:31:16 GMT
I've heard of 64NIC, but don't know the hardware details... okay a quick google tells me it is cartridge based. So without special programming it won't work at 2MHz... for example, it is known that REU will not work at 2MHz. In short, the software must be specially written to adjust CPU speed upon access to the cartridge I/O (anything in $DE00 ~ DFFF range). Off-topic, access to cartridge ROM is safe at any speed. I would buy one if it had 16MB REU... yeah, call me selfish, but I want *both* RAM expansion and Ethernet access in one cart
|
|
|
Post by Guest on Jan 22, 2015 13:58:56 GMT
I don't recall every having issues using the REU or even the SwiftLink cartridge I had in fast mode back when I had one, but it has been about 20 years since I did any coding/playing with the 128. I haven't repurchased either of those yet since getting back into some retro work. Maybe I had the changes to access them both already in my code, been way too long since I did that work. I'll do some digging.
|
|
|
Post by hydrophilic on Jan 23, 2015 8:09:04 GMT
Yeah, I don't own an REU myself, but based on my experiments in VICE, and feedback from users of MP128 and CP/M, any I/O cartridge access must occur at 1MHz or it will fail! Slightly off-topic, I have been thinking about a method to allow 2MHz access to a C128 cartridge.... simply said, a *special* program would need to run on the C128 so that the cartridge can monitor the 'dot clock' (8MHz) and determine when/how to communicate at full 2MHz speed. Sadly it is just a theory, and I don't have the hardware to actually build/test it
|
|
|
Post by Guest on Jan 23, 2015 15:20:44 GMT
Well If you get it working I think I'd be more than willing to try to utilize it. AFter doing a little reading, for the REU I'd say you are right, I probably was using the building stash and fetch commands from the kernel which do indeed change the setting to slow when interacting with the REU. I honestly don't recall playing around with the CPU speed when I was using the Swfitlink at all, but maybe I was and just don't remember... I have found information that does imply you must be in 1mhz when interating with the IO for the cart port so that's probably what's going on.
|
|
|
Post by VDC 8x2 on Jan 23, 2015 19:40:38 GMT
You think it has something to do with the vicIIe chip handling the ram?
|
|
|
Post by Guest on Jan 23, 2015 21:49:01 GMT
It certainly could be, though it was my understanding, and I am going back 20+ years from when I would have known better how this worked, was the whole point of FAST mode was to pretty much take the VIC accessing the ram out of the equation, or minimizing it.. hence the blank screen. In normal 1 mhz mode, the vic must share access to the bus with the CPU, so they effectively have to take turns... IE when the VIC needs the ram it tells the CPU, STAY OFF THE BUS UNTIL I'M DONE... and until the vic clears the signal the CPU will effectively just sit in a wait state waiting on the bus... In fast mode, the VIC isn't attempting to access the ram hence the reason the 40 column screen is just blanked... so the CPU can have the bus uninterrupted, and since they weren't trying to share the bus, the CPU could run at its full 2 MHZ mode because keeping the 2 chips in sync wasn't going to be a problem timing wise....
Of course I'm not a hardware guy, and my memory is not nearly what it used to be.... I assume that if access to the cartridge had to be done at 1mhz there had to be a reason, just note fully sure why that was. I don't think, and I could be completely wrong here, that the VIC had anything to do with the IO, and the memory bus certainly could handle 2mhz.. and I would assume the CIA's are mapped to the I/O on the cartridge port, not the VIC, which I would think too could handle 2mhz...
I know I am obviously wrong about something, since its clearly stated you need to be in 1MHZ to use the REU at the very least... though perhaps that's a limitation on the REU, not the 128, since the 1764, and the 1700/50 versions are exactly the same with the exception of I believe a resister or something on their boards.. perhaps the REU's bus and controller operated only at 1mhz, so if the 128 was in 2mhz mode it would not work properly... this to me makes some reasonable sense since the 64 only ran at the 1mhz mode.
I will completely and humbly accept that I am completely wrong on any of the assumptions made above if someone claims otherwise.
|
|
|
Post by guest on Jan 23, 2015 22:03:45 GMT
Wait, I think I'm wrong about the CIA's I don't think they come into play on the cartridge at all... if I am right about that, I am going to guess that the REC in the REU's themselves run at 1mhz, so if the bus isn't at 1mhz they just aren't going to work... which would make it a limitation caused by the REU not by the computer itself having problems using IO on the port at 2mhz... but again, as I said before, I could be completely and utterly wrong.
|
|
|
Post by donno128 on Jan 24, 2015 0:30:19 GMT
The inability to work in fast mode seems to be REU specific, or maybe specific cartridge specific. Both Warpspeed128 and Partner 128 work fine in fast mode.
|
|