exxos's Atari, Amiga & retro forum

For those using the Starstruck-Final..don`t it`s buggy use the Starstruck-partyV, I have put a link below

https://mega.nz/file/ItZl3Sxa#3Ky6wEjds ... gMGHJypti8

Re: starstruck-final. I've moved the card to my 1D4 board and changed clock speed to 75Mhz and the glitching is gone for me.

Plus it works fine on my 2B at 75Mhz (I must have forgotten to up the speed previously). Anyway I'm sick of that demo now! 😂 Back to work... or a cheeky game of Kick Off 2 🤔

A big thanks you to Superduper for the quality build and to Stephen for designing such a wonderful accelerator.

Rocking a 060 LC @ 62Mhz on the new Workbench 3.2 release, with fresh Kichstart roms burned yesterday.

Just need to sort out some cooling and I'm done, many thanks to all involved.

Excellent...

Let me clear up something... This card was really only meant for me and a few friends... so Speed Limit 106Mhz... Air do shocair.. is a teuchter joke and a reference to



Really do not expect anything but sarcasm on my silkscreens. Like everyone was told when they bought it.. you'll get maybe 75 to 80mhz out of the card max.

terriblefire wrote: 02 Jun 2021 08:34 Well it says you have specified PHASE and CRYSTAL when you have not.

cpuspeed.c

If anyone wants to fix it... knock yourself out.

Hi Stpehen. Yesterday a friend of mine worked on these cpuspeed codes and assembled a newer one. He'll post details and new verison of cpuspeed. I wanted to be sure it's stable and as far as i tested it doesn't give 4294 mhz error. On startup , after TBL demos , after reset or after manually make system crash by overclocking too much i tested every possible cause of this 4294 mhz error and didnt get one. Thanks for sharing codes, as i said he will post deatils and newer version.

terriblefire wrote: 02 Jun 2021 08:34 Well it says you have specified PHASE and CRYSTAL when you have not.

cpuspeed.c

If anyone wants to fix it... knock yourself out.

Challenge accepted.

I do not own a TF1260 but I've looked at the code by request. It looks quite straightforward. I refactored it to compile with gcc6.5.0 and added some control routines for the problems my friend ekalkici reported. He is happy with this version so far. For those who face similar issues, here it is.

Tested, seems to work fine for me, no crashes so far, speed is reported correctly, can still clock to 80mhz ok.

@Alpyre What are the differences between this version of cpuspeed and the original from Stephen?

trixster wrote: 06 Jun 2021 14:52 Tested, seems to work fine for me, no crashes so far, speed is reported correctly, can still clock to 80mhz ok.

@Alpyre What are the differences between this version of cpuspeed and the original from Stephen?

This one is compiled with gcc 6.5.0, so it should have better soft-float routines hopefully. Also it is linked with the native version of printf() and these two only are probably solving the reported issues.

But to be more reliable, I've added this routine that checks the status of the card after the new values are set. And if they do not match it will try 2 more times before returning an error message. Older code just used to report the values from the card status.

And finally this one also carries a version tag.

That new compiled version works OK here too! Stephen - Are the clock speed increments related to the board design btw or the FPGA code?