English Amiga Board

English Amiga Board (https://eab.abime.net/index.php)
-   News (https://eab.abime.net/forumdisplay.php?f=29)
-   -   Gold 2.7 release for the Vampire accelerators (https://eab.abime.net/showthread.php?t=91049)

Sinphaltimus 04 March 2018 07:36

Perhaps when my power is restored (March 7th maybe) all of this will be worked out and I can try again. My report May not be needed by then.

emufan 04 March 2018 10:01

Quote:

Originally Posted by Sinphaltimus (Post 1224019)
My report May not be needed by then.

I'm sure you will make it work. report your findings, this will help and motivate others :)

systmcrsh 04 March 2018 12:20

Quote:

Originally Posted by emufan (Post 1223960)
can anyone try and report what you get using calculator with fpu disabled?

i tried it with the fpu disabled and got the correct result 6.5

Seiya 04 March 2018 14:00

however, fpu implementation in 2.7 core is needed to run Quake and some Demoscene productions that require fpu :) also if this one is not very precise or accurate.

emufan 04 March 2018 19:26

Quote:

Originally Posted by systmcrsh (Post 1224070)
i tried it with the fpu disabled and got the correct result 6.5

thanks. above Alb42 explained already, how these libraries are working.
this all makes sense (no judging here) - so the vampire fpu needs some tweaking.
Quote:

Originally Posted by Seiya (Post 1224094)
however, fpu implementation in 2.7 core is needed to run Quake and some Demoscene productions that require fpu :) also if this one is not very precise or accurate.

they did an terrific job implementing the fpu, this after all the discussions we had.
just remember the most recent cpu flaws (Meltdown/Spectre) - that should be of more concern.

emufan 04 March 2018 23:42

ALB42 and Chain-Q did help to track down errors in the fpu:
Code:

flype  sagt:
 3. März 2018 um 23:31 Uhr
Ok, thanks ALB42 and Chain-Q for the inputs.
The Round() issue is now fixed in the next Vampire core.
MUIMapparium runs well now, both CPU and FPU versions are OK.

:great

emufan 05 March 2018 12:38

Quote:

Originally Posted by Sinphaltimus (Post 1223851)
Second problem, sdnet doesn't work. That was the dealbreaker. [/url]

Henryk Richter released a new version of the sdnet.device
Code:

Henryk Richter - 05 Mar 2018 07:30

For anyone interested, an update to the SDNet package is available
here: EXTERNAL LINK 
I've made some timing adjustments and changed the default SPI clock divider to 2,
so that it works more reliably with x12 and x13 cores.

For x11 cores, I'd recommend to set SPISPEED to 1, as outlined in the README.


alexh 05 March 2018 13:47

I am surprised the SPI core doesn't use a clock enable and a fixed rate tick to make it insensitive to clock speed changes.

emufan 05 March 2018 21:36

I have no idea :)

#1) I found something useful for V600 V2 owners, those boards need a fix for Gold 2.7
Code:

What capasitors do I need to upgrade? What are their numbers on board?
Code:

C3 and C6.. NOTE that it's only needed for the V600 V2.

Best way to test if the mod is needed is to flash the 2.7 core.
If it doesn't boot then you need the mod. Alternatively measure
the cap values and if it's less than 22uf then needs to be modded


TjLaZer 05 March 2018 23:16

Gold 2.7 release for the Vampire accelerators
 
After updating mine it bricked it with a red screen, but then I use the USB blaster and I got it to boot up! No caps mod needed.

Did they do a up to date registered compile for the 2.7 version? I registered my board last week right before they released the 600 version...

emufan 06 March 2018 01:32

no update for 2.7 yet. nothing i'm aware of. you have that black/white screen now?
IIRC, the problem with V600 V2 may occure under certain circumstances,
disk access or something. screen flicker?! maybe depends on the resolution used.
I cannot remember exactly. so you may do some more tests.

emufan 06 March 2018 15:43

Problem With SFS Partitions With Gold2.7:
Code:

The bug causes Workbench to display a second ghosted icon for the SFS partition
with the name "DHx:SFS\0"

This patch fixes a problem with SmartFilesystem 1.279 on a 68k Amiga with an
exec.library newer than 45.20, such as those found in Cloanto's 45.6x ROMs.

http://www.doobreynet.co.uk/
http://www.doobreynet.co.uk/files/amiga/SFS_Fix.lha

Sinphaltimus 07 March 2018 12:53

OK, so power has returned and there is so much new information out there with so many issues worked around and lot's of people sharing their experience. As promised I published my story of issues in the Apollo forums so rather than repost here, I'll just share the link. I feel lost in a sea of information right now so hopefully I can get the info I need in one place, here:

http://forum.apollo-accelerators.com...php?f=7&t=1872

emufan 07 March 2018 13:01

thanks for the link. i'll follow what will be written over there :)

#1) yesterday Gunnar gave a link to a Gold 2.7 "SAFE" x10 core V600, which is worth a look too.

#2) above I wrote about a sdnet update.

#3) plipbox problem fix for Gold 2.7:
Code:

  It needs to be placed at ENVARC:SANA2/plipbox.config.
 Just add "NOBURST" as the top line of the file.

#4) SAGADriver.v1.1.lha - 2018-03-06
Code:

Changelog:
[1.1] (06.03.2018)
* Updated VampireGFX.card to 1.12
    - Adds Scanlines settings
    - Assumes SOFTSPRITE=YES by default if not set in tooltypes
* Updated SDNet.device to 1.1
    - Fix for x12 cores
* Added VampireGFX tooltypes documentation


grond 07 March 2018 14:37

It may be worth to point out that the project has reached a level of complexity where it is very hard to test and debug everything. The past few days have shown a few bugs that were not found before in a year of development and testing but could be fixed in a relatively short time. This has been very positive for the project.

Djole501 07 March 2018 14:43

So it would be better to release publicly more often :) pointing out its a beta release.... Testing "in-house" is good but i can imagine, very slow. The users could decide for them self if they want to run the beta release or keep the production one....

Sinphaltimus 07 March 2018 15:51

Quote:

Originally Posted by grond (Post 1224916)
It may be worth to point out that the project has reached a level of complexity where it is very hard to test and debug everything. The past few days have shown a few bugs that were not found before in a year of development and testing but could be fixed in a relatively short time. This has been very positive for the project.


Absolutely agree - It's why I'm trying to provide as much detail as possible. I figure the more data the team can collect, the better...

emufan 08 March 2018 16:34

Gunnar just anncounces 2.7.1:
Quote:

GOLD 2.7.1 JIC is out

This new file gives color to all people which did register since the original 2.7 release.
Quote:

what other changes are in the 2.7.1?

No changes in the core.

Sinphaltimus 08 March 2018 20:39

Quote:

Originally Posted by Sinphaltimus (Post 1224908)
OK, so power has returned and there is so much new information out there with so many issues worked around and lot's of people sharing their experience. As promised I published my story of issues in the Apollo forums so rather than repost here, I'll just share the link. I feel lost in a sea of information right now so hopefully I can get the info I need in one place, here:

http://forum.apollo-accelerators.com...php?f=7&t=1872

So far so good on my round 2 attempt:

http://forum.apollo-accelerators.com...2&p=2819#p2819

emufan 08 March 2018 21:17

Quote:

Originally Posted by Sinphaltimus (Post 1225297)
So far so good on my round 2 attempt:

good news. what else is missing/not working?

#1) as a sidenote: if you get a new version of Apollo/CoffinOS,
at least archive the standard folder on your boot partition, such as C: L: S: DEVS: LIBS: ...
so you saved the config files, which you may use on the new Coffin version :)


All times are GMT +2. The time now is 23:20.

Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2024, vBulletin Solutions Inc.

Page generated in 0.05068 seconds with 11 queries