exxos blog - random goings on

[exxos]

DECODER & BOOSTER

IMG_6087.JPG

IMG_6088.JPG

IMG_6089.JPG

The scores are a tiny bit slower than the V2.2 booster. Part of it was the clock switching logic was glitching because of a faster PLD :roll: So I had to add a small delay in the logic to prevent it.. the side effect is a few % slower speeds.. But still, its still not far off anyway..

16CPUROM_result.png

EDIT:

Saying that I changed the compiler to "slow" outputs and while there is a small glitch, with the original code its still a few % slower :shrug:

[exxos]

Bit more progress...

IMG_6095.JPG

[exxos]

Built up 4 boosters, just testing out the second one now.. I had a issue of crashing after its been on a bit, thought it was the CPU, but seems theres more noise on the clock than I would have liked. So a small mod and the first one has been running for the past few hours..

IMG_6100.JPG

[exxos]

Been trying to get the TF536 working with EMUTOS 1.0.. There is odd stuff going on all over the place. Without the blitter I cant even access the floppy.. But when I do put the blitter in, things are better, but when running GB6 from alt-ram, I get this error ? @czietz @thorsten.otto ?

IMG_6111.JPG

Running in STram works fine..

IMG_6102.JPG

EMUTOS finds the 64MB atl-ram as well..

IMG_6108.JPG

[czietz]

What the error message means: An illegal ("line F") CPU instruction was executed within GEM Bench. Since I assume that there are no illegal instructions in GEM Bench: This can happen when (for whatever reason: faulty RAM, wrong timing of RAM reads) 0xFFFF is read by the CPU instead of the correct instruction.

[stephen_usher]

Looking at the timings, are you sure you don't have a load of ground bounce going on?

That's the sort of nonsense my system was doing without the ground plane on my board. Those 99%s look suspiciously like lost clock cycles.

[exxos]

Looking at the timings, are you sure you don't have a load of ground bounce going on?

That's the sort of nonsense my system was doing without the ground plane on my board. Those 99%s look suspiciously like lost clock cycles.

All the boosters I have ever seen with 020, 030 always seem to have slightly slower speeds than expected.

[JezC]

@exxos I have seen a similar issue with the floppy on my H4 with TF536 fitted and EmuTOS v1.0...but the same setup seems to work fine with a slightly older EmuTOS.
I'm on the phone at present but the earlier version that agranlund posted still works fine.

I boot EmuTOS from floppy with his special EmuTOS boot ROM image & with the v0.9x it all works fine...but change to loading v1.0 from floppy & it boots ok but then the floppy stops working. :shrug:

May not be the same...but maybe it is...?

[exxos]

@JezC Its all to confusing with all the various softwares , firmwares and patches about. Its hard to work out whats what with it all. Things mostly seem to start going nuts when alt-ram is enabled.. but it still isn't stable with it off.. Its gone back "on the shelf" for a bit, as its not going to be simple to work it all out.

[exxos]

I have gone back to my SEC booster on my H4.. Took a bit of working out as there seemed to be a bad connection on the CPU preventing it from booting :roll:

I'm going to try and hook up the fast-ram boards I got made over a year ago.. So I can just run GB6 from my own alt-ram and see where it all ends up.

IMG_6113.JPG

IMG_6114.JPG