FPGA Replay Board

[Darrin]

I have a friend that has a Chameleon and another that has an MCC-216, but have never seen anyone put together a review comparing all FPGA based systems with each other.  It would be interesting to see in which ways they are the same, or similar and which features are different on each of these systems.

I have not kept up with the features of the Chameleon, MCC-216, Natami, or expanded v1.1 MiniMig.  I think MikeJ has done a better job of keeping most of us Amiga users up to date with the progress of his work and the work of Yaqube, than some of the other system designers have done with their offerings.  Although the Natami Team seems to be very active on their forum, not as much of that info spills over to sites like this one.  The same is probably true for the Chameleon and MCC-216 on their own sites.

I did actually start a thread here with the intention of comparing the same software running on the Minimig v1.1, Chameleon and FPGA Arcade (obviously the FPGA would test AGA versions where available).  There didn't seem to be much interest and I got very busy with work so I shelved it (plus I was expecting a core update for the FPGA Arcade so it seemed to make sense to wait).

Here's all 3 machines running ClassicWB:

[Darrin]

The Minimig v1.1 is built into the black keyboard on the far left.  You should just to able to make out the yellow Chameleon cartridge sticking out of the cartridge port expander on the back of the C64C and the FPGA Arcade is that tiny board to the left of the keyboard on the right.

Another problem with testing is that the Chameleon Minimig core doesn't recognise the joystick ports on the C64 (yet).  The IR controller is supposed to work if I use a CDTV remote as a joystick, but I don't have one.  My only option was to use the keypad on a PS2 keyboard as a joystick.

[matthey]

I hope the FPGA Arcade can do better with the update. If it's the 68020 enhancements slowing it down, then it would be hard to imagine the Natami N68k achieving 68060 performance.

[freqmax]

Natami N68k is a whole another ballgame. It's done in AHDL which is more specialized for the FPGA they use at the price of flexibility and portability. There's likely more people working on the core as the current FPGA Arcade source is closed in practice atm. It's also that the original source is an 68000 in a hardware description language (HDL), which has then been modified into a 68020 core.

[amigadave]

The Minimig v1.1 is built into the black keyboard on the far left.  You should just to able to make out the yellow Chameleon cartridge sticking out of the cartridge port expander on the back of the C64C and the FPGA Arcade is that tiny board to the left of the keyboard on the right.

Another problem with testing is that the Chameleon Minimig core doesn't recognise the joystick ports on the C64 (yet).  The IR controller is supposed to work if I use a CDTV remote as a joystick, but I don't have one.  My only option was to use the keypad on a PS2 keyboard as a joystick.

I think AmigaKit still has a ton of brand new CDTV remote controllers for sale.

I for one am interested in testing comparisons of your three FPGA based Amiga computers, so I hope you will find time to renew your interest in testing and reviewing them for the rest of us.  You can always update the results for the FPGA Arcade Replay board after you get the next firmware update.  That will also show how much improvement has been made between the firmware you have now, and the next version.

I should have my own Replay board within the next few days.  I can hardly wait to get it and test it out.

[mikej]

I hope the FPGA Arcade can do better with the update. If it's the 68020 enhancements slowing it down, then it would be hard to imagine the Natami N68k achieving 68060 performance.

The cache and prefetch must be disabled in the current core. This is what is hurting performance, not the cpu extensions. The new memory controller and cache system does a lot better.
/MikeJ

[matthey]

The cache and prefetch must be disabled in the current core. This is what is hurting performance, not the cpu extensions. The new memory controller and cache system does a lot better.
/MikeJ

That should make a HUGE difference in performance. Maybe up to a 68030@50MHz? The fpga CPU isn't really pipelined is it? How big of cache? Is the branch logic like the 68020/68030 or the better 68040/68060/CF where branches backward are predicted taken and branches forward are assumed not taken. The latter provides much improved performance, especially on loops. The latter is what the N68k CPU will use also. I suggested the least significant bit of branches (no odd branch targets) be used to reverse this logic. It would make for prettier, more compact code, easier design for compilers, and easier branch optimization that could for example, change 68020/68030 branches to the newer style. A setting could allow this bit in the instruction to be inverted on the fly as the code is executed. Please consider cooperating with the Natami team to create standard ISA enhancements to the 68k after you have the fpga Arcade soft CPU working better .

[freqmax]

How big of cache?

The amount of SRAM in a FPGA is quite limited. So maybe a "real" cache should be built with off-chip SRAM. But..

Earlier results show 256 byte instruction- and data cache improves performance to twice that of A4000 68040 @ 25 MHz. Dhrystone 9868, MFlops 10.30 measured with sysinfo.

[billt]

Please consider cooperating with the Natami team to create standard ISA enhancements to the 68k after you have the fpga Arcade soft CPU working better .

Are Natami team open to this as well, or are they a closed-project group wanting to do their own thing?

[mikej]

1 block ram is about 2K byte. This is quite sufficient as a cache in this case.

/MikeJ

[Darrin]

I think AmigaKit still has a ton of brand new CDTV remote controllers for sale.

I for one am interested in testing comparisons of your three FPGA based Amiga computers, so I hope you will find time to renew your interest in testing and reviewing them for the rest of us.  You can always update the results for the FPGA Arcade Replay board after you get the next firmware update.  That will also show how much improvement has been made between the firmware you have now, and the next version.

I should have my own Replay board within the next few days.  I can hardly wait to get it and test it out.

Thanks for that info.  I'm flying to the UK next month and I'll be in Cardiff.  I keep saying I'm going to call in to AmigaKit's store and this will give me a perfect excuse.  I hope they'll have some Spider PCI cards in too.

[Darrin]

The cache and prefetch must be disabled in the current core. This is what is hurting performance, not the cpu extensions. The new memory controller and cache system does a lot better.
/MikeJ

Also, the Chameleon's soft 68000 isn't really running at 110MHz, despite what SysInfo says.  :-)

[freqmax]

1 block ram is about 2K byte. This is quite sufficient as a cache in this case.

Quite enough then

Motorola caches:
68000 - none
68020 - 256 B I-cache
68030 - 256 B
68040 - 4 kB
68060 - 8 kB

Also, the Chameleon's soft 68000 isn't really running at 110 MHz, despite what SysInfo says.  :-)

Run a C compilation 10 times in sequence. That ought to give a sufficient comparision ballpark figure when Sysinfo fails.

[matthey]

\
Earlier results show 256 byte instruction- and data cache improves performance to twice that of A4000 68040 @ 25 MHz. Dhrystone 9868, MFlops 10.30 measured with sysinfo.

You mean Mips=10.30 not MFlops right? There's probably not room to add FPU support in the fpga although it would only take 4 FPU instructions to do well in the SysInfo MFlops test. I posted the SysInfo FPU MFlops test on the Natami board...

http://www.natami.net/knowledge.php?b=2¬e=35355

The SysInfo tests are a joke. SysSpeed is a little better.

Are Natami team open to this as well, or are they a closed-project group wanting to do their own thing?

I think they are open to standards for 68k ISA enhancements and (S)AGA/RTG enhancements. It would mean more software would support it. Software needs to support enhancements to get a benefit from them. More than 1 CPU standard could be supported if it's too burdensome to add full support. For example, a 68CF standard could add ColdFire instructions only and 68CF2 could add more enhancements for example. A defined specification, name with which to specify it and willingness of developers to support it is what is needed. I would be willing to help if I could find enough interest. I think developers that would support it include the Natami Team, Frank Wille (vasm, vbcc) and me (new version of ADis disassembler and also worked with Frank on some CF optimizations for vasm including more to come). Rune Stensland may do a Asm-Pro update as well. The 68k is a great development platform and we can make it better with a little bit of effort and cooperation.

1 block ram is about 2K byte. This is quite sufficient as a cache in this case.

Even a 2k cache will cause trouble for poorly written games using self-modifying code. Can the cache size be changed or is there bus sniffing (snooping)? I believe the N68k will get bus sniffing eventually in order to have as large of caches as they want.

[freqmax]

I hope there won't be any huge list of extra instruction ISA that all Amiga now should implement. Rather a few instructions that would give a great boost.

The FPGA capacity is limited.

Bloatwarning ..

[freqmax]

You mean Mips=10.30 not MFlops right?
.
.
The SysInfo tests are a joke. SysSpeed is a little better.

Mips, I read the screenshot wrong.

What method difference between Sysinfo and SysSpeed is there that makes SysSpeed benchmark so much more accurate?