Jump to content
tmp

AVGCART

Recommended Posts

11 hours ago, tmp said:

is the same happening on 800xl or is it 800 specific?

 

Just finished testing the exact same procedure I listed on my 800 / Incognito and 800XL / U1MB. In those two platforms, the test completes successfully with wither DMA = ON or OFF. NO timing problems reported there.

 

Having said that, you may want to take a look at post #1587 , on this very same matter, across other platforms:

 

 

This is a serious, deeper issue that (at this moment) appears to me related to how AVG times itself with respect to the system bus (but that is my 80,000 ft. assessment on the matter, though).

Share this post


Link to post
Share on other sites

that incognito 800 has phi2 or ras on its cart slot?

Share this post


Link to post
Share on other sites
1 hour ago, tmp said:

that incognito 800 has phi2 or ras on its cart slot?

RAS.

 

My 800 / Incognito has no permanent / non-reversible modifications (of any kind) on its MoBo. That means, it has factory-fitted RAS signal on LEFT-cart port, and PHi2-signal on RIGHT-cart port. If you are wondering, SIDE-2 also works without problems or any modifications on that platform, as well.

 

My prior thoughts on this are suggest that, once Incognito is properly installed on the 800 (and removal of almost 40+ active components spread in different boards), it eventually alters (favorably) the entire timing chain, or at least once it hits the Personality slot (where Incognito resides). This in turn, may help explaining why stuff that it is supposed NOT to work with RAS, ends up working there, without any problems.

 

However, per my last post (and reference), it seems this problem is also appearing in other XL-based systems. What we need at this point would be a black-and-white guidance as to what AVG timing tolerance or requirements are... because it DOES manage to work, for a good part, on my reference 800 (non-Incognito).

 

 

 

 

 

Share this post


Link to post
Share on other sites

on that incognito 800, did you have anything pbi/hdd related disabled in incognito setup so side2.sys from cart was used as on stock 800?

Share this post


Link to post
Share on other sites
13 minutes ago, tmp said:

did you have anything pbi/hdd related disabled in incognito setup so side2.sys from cart was used as on stock 800?

An important question, for sure.


APT drive IDs on Incognito are setup DIFFERENT than APT ids in AVG. Therefore, and for full clarity:

 

  1. Plug AVG cart on left-port (APT ids are DJ and DK).
  2. Power-up i-800 to BIOS, and DISABLE SDX (do not disable PBI & HD access, since APT ids are D8 and D9).
  3. Press "B" and boot into AVG selector.
  4. Press "0" which is my shortcut for latest SDX load for AVG.
  5. SDX boots from AVG, and RECOGNIZES its own HD storage, plus also sees (independently) Incognito HD storage (this only possible on the 800).
  6. Load SI226 and perform AVG-based HD tests. I can also perform Incognito-based HD tests, since the drive IDs are also reported to SI226.

 

All test pass on i-800, without any problem, with or without DMA active. No matter which drives / HD you choose. AVG is plugged on left-cart port, with RAS active on it.

Share this post


Link to post
Share on other sites

so can you please try to disable pbi in order to make stock and incognito 800s more comparable regarding code that's doing the hdd access? (there could be some version differences and afaik if you have pbi hdd enabled, side2.sys won't load and code on incognito will be used instead)

i want to rule out sw since there's not much hw difference if you're saying that both machines have ras instead of phi2

 

it would be also helpful if anyone with stock 800 (no incognito, no cart port "fix") could run the same test

Edited by tmp

Share this post


Link to post
Share on other sites
7 minutes ago, tmp said:

so can you please try to disable pbi in order to make stock and incognito 800s more comparable regarding code that's doing the hdd access?

Sure, I will test and report back...

Share this post


Link to post
Share on other sites
2 hours ago, tmp said:

so can you please try to disable pbi in order to make stock and incognito 800s more comparable

Ok, not only tested on XL-mode (with PBI disabled on BIOS), but also on COLLEEN mode (where there is no PBI to speak of, and being as close you can get to stock 800).

 

The results are the same, SI-v226 HD (DOS) test works with or without DMA. All of this always booting SDX from AVG directly (nowhere else).

 

As a side note, I re-tested with Incognito's PBI-enabled (on XL-mode), so I could access all APT HD partitions, from both AVG and Incognito's own HD. It is worth mentioning that when running tests with System Information 2.26 set to DOS (not OS), Incognito's HD speed reports around 68 KB/sec, whereas AVG reports 60 KB/sec, at most (this is with DMA=on). The latter is also the same readout when running this test on a 800XL platform. This is over 10% of speed-increase when reading Incognito's internal HD, though.

 

Let me know what else would you like to test.

Share this post


Link to post
Share on other sites
On 8/23/2021 at 2:33 PM, Faicuai said:

Ok, not only tested on XL-mode (with PBI disabled on BIOS), but also on COLLEEN mode (where there is no PBI to speak of, and being as close you can get to stock 800).

 

The results are the same, SI-v226 HD (DOS) test works with or without DMA. All of this always booting SDX from AVG directly (nowhere else).

 

As a side note, I re-tested with Incognito's PBI-enabled (on XL-mode), so I could access all APT HD partitions, from both AVG and Incognito's own HD. It is worth mentioning that when running tests with System Information 2.26 set to DOS (not OS), Incognito's HD speed reports around 68 KB/sec, whereas AVG reports 60 KB/sec, at most (this is with DMA=on). The latter is also the same readout when running this test on a 800XL platform. This is over 10% of speed-increase when reading Incognito's internal HD, though.

 

Let me know what else would you like to test.

My 800 with retrocanada's designed 512K axlon board, assembled by Jeffrey Worley, SI2 reports for AVGCart (DD21 FloB edition) with old stock Sandisk 32GB SD card (class 4) a rate of 62 KB/sec (best rate, *sequential read*, I think) with DMA disabled.  With DMA enabled it reports 98 KB/sec. There are no modifications aside from the Axlon.   I don't own an Incognito equipped 800.

 

The very same AVGCart, now in my 800XL, yields similar numbers, also 62 KB/sec with DMA but now in the *random sector* test.  60 KB/sec for *read sequence*.  The number were reversed on the 800.  97 KB/sec best rate with screen .DMA disabled.

 

In the same 800XL I have an internal MyIDE with an IDE to CF adapter.  Numbers are also similar here as well, 62/60 KB/sec (*read sequence*/*random sectors*) and 96 KB/sec (across all tests).

 

-SteveS

 

P.S.

These are all the "HDD" options I own.

Edited by a8isa1

Share this post


Link to post
Share on other sites

so no "ERROR 144" during those tests on 800?

 

also dma/no dma speeds being swapped on 800 is weird and doesn't make much sense (unless it's a bug in SI2 but i don't know what would be the reason to handle 800 and 800XL differently there)

Share this post


Link to post
Share on other sites
3 hours ago, tmp said:

so no "ERROR 144" during those tests on 800?

 

also dma/no dma speeds being swapped on 800 is weird and doesn't make much sense (unless it's a bug in SI2 but i don't know what would be the reason to handle 800 and 800XL differently there)

oops my *dyslexia* kicked in.   reversed "enabled" and "disabled".

  • Like 1

Share this post


Link to post
Share on other sites
22 hours ago, tmp said:

so no "ERROR 144" during those tests on 800?

 

also dma/no dma speeds being swapped on 800 is weird and doesn't make much sense (unless it's a bug in SI2 but i don't know what would be the reason to handle 800 and 800XL differently there)

no "Error 144" .

 

  • Thanks 1

Share this post


Link to post
Share on other sites

 

On 8/23/2021 at 1:43 PM, tmp said:

if anyone with stock 800 (no incognito, no cart port "fix") could run the same test

 

Alrighty!

 

Been quote busy testing Incognito's PBI facilities with latest JED, but eventually managed to return to these tests... with more HW and firepower (and also completing tests for a similar problem with Side/2 cart).

 

I fired my 800 "lab"/test-bed, for which I have multiple OS, RAM and CPU cards. NO mods whatsoever, just a naked Rev-9 MoBo, identical to my 040-0077 reference 800. Once setup, I fired-up AVG and tested with multiple cards combinations, including (x2) plain 6502 and (x1) Sally spare CPU boards.

 

The conclusion is that I can now reproduce and also mitigate AVG's Error-144 code on SDX, at will. During my tests,  I found the source of the problem to be the CPU board, being SALLY-version the one that mitigates the problem, and older-6502B the offenders.

 

I DID NOT test an extra Sally board to confirm that the problem would go away with any of such board-type, but it is clear that AVG cannot run properly (for some reason) with both 6502B boards I tested. Coincidentally, the CPU board running on my 040-0077 A800 is also 6t02B (old-type).

 

So there we have it. And when it comes to Side-2 tests, I actually managed to boot it successfully, including SDX, from specifically one (1) of the two spare 6502B boards, with 8014 (date) CPU, the oldest (!!!) I thought the newer Sally board would be the cure-all remedy (with RAS signal on left-port) but it turns out to be a bit more complex than it looks.

 

Any comments here are more than welcome.

Share this post


Link to post
Share on other sites
On 8/30/2021 at 3:27 PM, Faicuai said:

The conclusion is that I can now reproduce and also mitigate AVG's Error-144 code on SDX, at will. During my tests,  I found the source of the problem to be the CPU board, being SALLY-version the one that mitigates the problem, and older-6502B the offenders.

 

 

Alright, just wanted to circle-back here and close the loop.

 

After a much longer testing, I finally got my 040-0077 800 to reliably work with BOTH AVG *and* SIDE-2 carts, with absolutely NO modifications of any kind the MoBo, and still with stock RAS-signal (instead of Phi2) on the left-port. The intermittent nature of this problem has been a long-standing, unresolved mystery on my end, which now seems finally addressed.

 

What (absolutely) made the difference was replacing the older 6502 A/B type CPU board (as well as any of my same-kind spare boards) for a fresher (1983) 6502-Sally type board. ALL timing errors disappeared in a single-stroke, being everything reduced to an occasional (but ineffectual) ATA-command misfire during IDEDIAG tests in SDX, which does not show up anywhere else in any of external-app tests and latest versions of AVG and Side2 Loaders + SDX. I also confirmed this significant improvement to be effective irrespective of my OS boards (stock-form or RAMROD with 52K enabled).

 

FYI (and just in case), I noticed that, when testing several older-style 6502 A/B CPU boards, one could fare partially better than other, in that specific group. However, I eventually tracked the difference down to the CPU itself (which I swapped between such like-boards, improving, but not solving the problem entirely on such boards).

 

Furthermore, I cannot discard the chance of finding an 6502 A/B outlier that will drive perfectly both AVG and Side2 carts' HDs, but as far as I could test, the possibility seems remote. So you may save plenty of time by testing Sally-boards, right from the get-go. 

 

In line with above findings, there is a possibility that when swapping such older boards for a fresher Sally-type, you may not instantly mitigate the offending timing issues, BUT testing 2 or 3 different 6502/Sally chips on that newer board, may very well get rid of such problems, if you are experiencing them. (also important for any XL users out there randomly experiencing these timing errors).

 

Just trying to close the case for any 800 owners out there, that would like to bring the best out of the system's architecture and potential with essentially no other internal HW mods, but a few plug-and-play enhancements (e.g. AXLON RAM, RAMROD 52K & multi-OS card, Hard-Drive add-on via AVG or Side-2, latest SDX DOS, etc.)

 

Enjoy!

 

 

 

 

Share this post


Link to post
Share on other sites

interesting, thanks for the testing and figuring out what's going on

 

so it's either that 6502 and sally (6502c) have some different characteristics that come into play or there's some supporting circuitry on these cpu boards that is different

 

was using hdd emulation in sdx the only way to trigger the issue or was it showing also when using any other avgcart functionality?

Share this post


Link to post
Share on other sites
11 hours ago, tmp said:

so it's either that 6502 and sally (6502c) have some different characteristics that come into play or there's some supporting circuitry on these cpu boards that is different

Or BOTH, actually. In any case, that conclusion seems correct.

 

AVG problems showed when booting SDX and attempting disk access through it (even System Information 2.26 would fail its HD benchmarks, except if DMA was set to OFF, which was already a warning sign of the nature of the problem).

 

In Side-2, however, the problems manifested differently: it would not even see the CF card from Loader menu, and much less boot SDX properly ("No Media" in both cases).

Share this post


Link to post
Share on other sites

In view of one of the contestants in the ABBUC software contest 2021 which needs BASIC XE 4.1, I have a request for @Mr Robot's De-Re-AVGCART:

 

AVG can simultaneously load a CAR and disk images, but only if all are in the same directory of the SD card.

First assign the disk image to disk drive D1 by pressing SHIFT-1, then select the CAR file and press RETURN to boot the computer.

 

This does not work if you need to change the directory of the SD card in between.

Share this post


Link to post
Share on other sites

Yes, it's a bit of a pain for people like me who don't like big mixed dirs, I keep stuff grouped by what they are..

 

I wonder if there's a way that can be changed.. @tmp

Share this post


Link to post
Share on other sites

this has been discussed a few times, i thought changing this would be easy but it turned out that it would require a more involved code change

 

until i find time to look into that, the workaround with shortcut (set shortcut to CAR, go to dir with ATR, assign it to D1: and run CAR via shortcut) should work

  • Thanks 1

Share this post


Link to post
Share on other sites
On 9/11/2021 at 3:20 PM, Faicuai said:

 

Alright, just wanted to circle-back here and close the loop.

 

After a much longer testing, I finally got my 040-0077 800 to reliably work with BOTH AVG *and* SIDE-2 carts, with absolutely NO modifications of any kind the MoBo, and still with stock RAS-signal (instead of Phi2) on the left-port. The intermittent nature of this problem has been a long-standing, unresolved mystery on my end, which now seems finally addressed.

 

What (absolutely) made the difference was replacing the older 6502 A/B type CPU board (as well as any of my same-kind spare boards) for a fresher (1983) 6502-Sally type board. ALL timing errors disappeared in a single-stroke, being everything reduced to an occasional (but ineffectual) ATA-command misfire during IDEDIAG tests in SDX, which does not show up anywhere else in any of external-app tests and latest versions of AVG and Side2 Loaders + SDX. I also confirmed this significant improvement to be effective irrespective of my OS boards (stock-form or RAMROD with 52K enabled).

 

FYI (and just in case), I noticed that, when testing several older-style 6502 A/B CPU boards, one could fare partially better than other, in that specific group. However, I eventually tracked the difference down to the CPU itself (which I swapped between such like-boards, improving, but not solving the problem entirely on such boards).

 

Furthermore, I cannot discard the chance of finding an 6502 A/B outlier that will drive perfectly both AVG and Side2 carts' HDs, but as far as I could test, the possibility seems remote. So you may save plenty of time by testing Sally-boards, right from the get-go. 

 

In line with above findings, there is a possibility that when swapping such older boards for a fresher Sally-type, you may not instantly mitigate the offending timing issues, BUT testing 2 or 3 different 6502/Sally chips on that newer board, may very well get rid of such problems, if you are experiencing them. (also important for any XL users out there randomly experiencing these timing errors).

 

Just trying to close the case for any 800 owners out there, that would like to bring the best out of the system's architecture and potential with essentially no other internal HW mods, but a few plug-and-play enhancements (e.g. AXLON RAM, RAMROD 52K & multi-OS card, Hard-Drive add-on via AVG or Side-2, latest SDX DOS, etc.)

 

Enjoy!

 

 

 

 

Time to add to the confusion. Sorry. My 1981 800 has a 6502B and is perfectly happy with AVGCart and AXLON compatible 512K SRAM board.  No modifications or other additions.

 

a8isa1_6502B.thumb.png.ddf01aed2d81d3cd2d37c47b7c0f4b13.png

 

-SteveS

Edited by a8isa1

Share this post


Link to post
Share on other sites
30 minutes ago, a8isa1 said:

Time to add to the confusion. Sorry. My 1981 800 has a 6502B and is perfectly happy with AVGCart and AXLON compatible 512K SRAM board

Mine worked too (minus occasional 144 error...) But SIDE2 did not (you can try that one, as well).

 

In any case, I will not doubt that some older 6502B or similar CPU (plus supporting logic chips) could nail the timings right, under the right conditions. But for anyone that did not, there is a remedy now.

 

I have been running both carts (SIDE2 and AVG), essentially error free, since switching to Sally CPU board, and no HW modifications of any kind on host 800's MoBo (RAS still present on left-cart port).

Share this post


Link to post
Share on other sites

Awesome!

 

You are truly making use of the field-programmability of the FPGA! So many options to explore with the same hardware. And kudos for designing the whole cart to support all these possibilities from the start.

 

Nice tease at the end too.

  • Like 5

Share this post


Link to post
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

Loading...

  • Recently Browsing   0 members

    No registered users viewing this page.

×
×
  • Create New...