Franken-A3640-060 project-Anybody Game?

[jbenam]

Originally Posted by SpeedGeek

Many thanks to my good friend SuperCosmos (and thanks to matthey also) for the 060 adapter and Rev. 1 CPU!!!

Ohhh I want some pics of the 060 adapter!! Pretty please? Give us some hardware pr0nz

[SpeedGeek]

@huepper

Thanks!

@Retrofan

The A3000 already has an Amber chip (which provide a basic scan doubler function). While it may not offer all the features of Indivison ECS (which should work perfectly fine with the A3000). I'm sure most A3000 owners will find it difficult to justify the price of the Indivision ECS.

@jbenam

There is already plenty of hardware prOn on the A1k.org thread! Also, I've been very busy with software patching and debugging... I now have a patched exec.library (45.20b) working with 030, 040 and 060! I'm sending it to Cosmos for additional beta testing, but that's far more important to me than more hardware prOn...

[SpeedGeek]

OK, as promised (to Cosmos) the first episode of A3660 overclocking has begun!

I attached a small cooling fan to the Rev. 1 CPU + Ramsey's CLK90 is modified for the A3000*.

Hopefully, as I acquire the additional components needed and time permitting clock speeds will increase with future episodes.

*Note: There is no practical overclocking limit with the adapter itself.

[alexh]

Looking very nice.

Cosmos said in his blog this adapter works with the Apollo A1240 with mach 130 chips.

I've got an Apollo A4040 which has Mach130 chips on it... is there a chance this could be upgraded to 060?

[protek]

I have just one question: where can one buy these adapters?

[SpeedGeek]

@alexh

I don't know why these Mach130 chips would make any difference unless they fixed timing problems with the Apollo which could cause problems for the 060 adapter.

The adapter was designed for 040 accelerators which use a clock doubler PLL to compensate for skew between BCLK/PCLK. But if your 040 accelerator uses a clock divider there's a good chance the designer ignored the skew simply because the 040's internal PLL would provide the skew compensation.

@protek

Cosmos has a few extra PCBs, but I don't know if he has all the extra parts needed to make a complete adapter. Also, he probably has more people asking to buy them now than he can possibly satisfy. It's really a DIY project.

[protek]

Does it make any difference what the nominal speed of the 060 is?

[SpeedGeek]

OK, episode #2 of A3660 overclocking is completed!

This is the upper limit for this Rev. 1 CPU... Also, SDMAC is working at it's maximum CLK speed. I made a custom oscillator and added an extra 10 uF cap to the top PCB for this project.

Higher clock speeds are still possible but will require a late Rev. CPU and a custom timing hack to keep SDMAC happy....

[Zetr0]

@SpeedGeek

my friend its delicious!

truly TRULY delicious!

[SpeedGeek]

@ZetrO

If your getting hungry my friend, there are a few threads on Amibay which may be able to satisfy your appetite!

@Thread

Cosmos has released the 060 FPU disable patch in a few recent updates here:

http://cosmosunivers.free.fr/download/exec.readme
http://cosmosunivers.free.fr/downloa...3000%29.readme
http://cosmosunivers.free.fr/downloa...4000%29.readme

[alexh]

I don't know if anyone has noticed but this product has attracted the attention of some hardware hoarders.

I've had an eBay alert for Sonnet Quaddoubler for many years, and not a sniff of any and then this product is released and all of a sudden two in two weeks

http://www.ebay.co.uk/itm/321530623695

[alenppc]

Those are only readme files, is there a link to download the patch?

[SpeedGeek]

Originally Posted by alenppc

Those are only readme files, is there a link to download the patch?

Has Google search been banned in Canada recently?

http://leblogdecosmos.blogspot.com/p/coding.html

[alenppc]

Thanks. I did try googling it but the only things I was getting were a) the readme and b) "this blog is now closed" page.

[alenppc]

OK, so I built the board. It is pretty amazing, works "out of the box" on the A3640!

I got what was supposed to be an EC060 rev 4 CPU but turned out to be an LC060, not bad.

The system actually boots with regular 68060 libraries, it doesn't seem to require the special LC version, but I have not done extensive software testing yet.

I tried patching the kickstart to be able to use the full 68060 by taking the regular 40.68 A4000 kickstart and replacing only the patched exec.library. The result is that the Amiga does not boot with the patched kickstart (no matter which CPU card is used). Replacing other kickstart files is not practical since the A4000 can only use 512K ROMs. If someone has successfully patched the kickstart and is willing to share the details, I would appreciate it.

Here some pics of the assembled adapter

[tnt23]

Cool, now I want one too Can you post SysInfo speed test results?

[jbenam]

I'm definitely going to build one too

Why do you need to patch exec.library? I remember reading something on it on a1k.org but it was mostly german

[alexh]

Originally Posted by SpeedGeek

You only need to disable the FPU for early startup then 68060.library enables the FPU at boot time.

http://www.amiga.org/forums/showpost...5&postcount=15

Originally Posted by Thomas Richter

It is more complicated than this. Yes, exec init uses an FRESTORE to test the presence of a FPU, but the problem does not go away by patching of this. The second problem is the exec task dispatcher, or rather the stack frame exec uses for saving the FPU context when it switches between tasks. The exec scheduler thus also requires patching to be able to work with 060's.

What the 060 library does is that it *in addition* includes a patch to exec that discovers some popular mis-use cases of FRESTORE, most notably the MANX/C Compiler startup code which uses FRESTORE to initialize the FPU. The patch just detects this startup code and works around it.

[SpeedGeek]

Originally Posted by alenppc

OK, so I built the board. It is pretty amazing, works "out of the box" on the A3640!

Have you tried it on a state machine modded A3640 yet?

Originally Posted by alenppc

I tried patching the kickstart to be able to use the full 68060 by taking the regular 40.68 A4000 kickstart and replacing only the patched exec.library. The result is that the Amiga does not boot with the patched kickstart (no matter which CPU card is used). Replacing other kickstart files is not practical since the A4000 can only use 512K ROMs. If someone has successfully patched the kickstart and is willing to share the details, I would appreciate it.

I used the A3640 MapROM function to test 512K Kickstart builds before burning them into ROM. If you have a Romy you can use 1MB ROMs but that's really another topic... .

Based on a recent email from Cosmos, he appears to have introduced a bug in his exec patch by moving the patch code to a different location. (He also reported a problem with WinUAE but I assumed it was due to a WinUAE bug).

I have warned him about this and hopefully he will get it solved soon. I have not actually tested any of the Cosmos exec patches myself, but if you are having problems please email him and report the version you are having problems with.

Originally Posted by tnt23

Cool, now I want one too Can you post SysInfo speed test results?

SysInfo (AKA MisInfo) is not a reliable benchmark program. Would you trust a program which gives bogus MHz ratings, uses a small cache based MIPS benchmark, and 9 out of 10 times reports N/A for Ramsey and Gary on the A3000?

[alenppc]

Originally Posted by SpeedGeek

Have you tried it on a state machine modded A3640 yet?

No, I gave up on that since I can't find GALs/PALs that work reliably with my programmers, and I have two of them by now. The Chinese ones they sell on ebay are crap, and I can't get the new ATF ones from Digikey to work with my programmers.

Originally Posted by SpeedGeek

I used the A3640 MapROM function to test 512K Kickstart builds before burning them into ROM.

Yes, that's what I am doing too, I am going to burn ROMs once I can make a working version. Right now, attempting to softkick a modded 40.68 file (with exec 45.20a) makes the machine not boot with either 040/060 or third party cpu cards, so I think the patched exec is broken.

Originally Posted by SpeedGeek

I have not actually tested any of the Cosmos exec patches myself, but if you are having problems please email him and report the version you are having problems with.

Okay, I am going to email him. How are you disabling the FPU at boot?

Originally Posted by SpeedGeek

SysInfo (AKA MisInfo) is not a reliable benchmark program. Would you trust a program which gives bogus MHz ratings, uses a small cache based MIPS benchmark, and 9 out of 10 times reports N/A for Ramsey and Gary on the A3000?

Indeed, SysInfo is mostly useless on 060 since the entire test will fit inside the CPU cache and it will report values comparable to that of a real 060 accelarator with fast ram, probably best to test with SysSpeed.