ZORRAM

_________________
I HAVE ABS OF STEEL
--
CAN YOU SEE ME? CAN YOU HEAR ME? OK FOR WORK

@MEGA_RJ_MICAL

The ZoRAM is a memory module board not a CPU. Stay with us man!

_________________
"Art challenges technology. Technology inspires the art."
John Lasseter, Co-Founder of Pixar Animation Studios

@cdimauro

Quote:

Does it matter? My statement on Vampire was on a very different context, right?

Yes, it matters when you use the word "Vampire" when you mean Apollo V4, which isn’t Vampire. This isn’t just you, people have been using "vampire" for anything with apollo core, while it has always been important to distinguish the two projects, now so even more important. From what I have seen, the page you link to is an exception, the term "vampire" is gone from all V4 product pages.

And on your context - do you know the chipset+RTG capabilities of other FPGA systems? How do you define "powerful" in that context, and "chipset" for that matter?

_________________
B5D6A1D019D5D45BCC56F4782AC220D8B3E2A6CC

kolla Quote:

Yes, it matters when you use the word "Vampire" when you mean Apollo V4, which isn’t Vampire. This isn’t just you, people have been using "vampire" for anything with apollo core, while it has always been important to distinguish the two projects, now so even more important. From what I have seen, the page you link to is an exception, the term "vampire" is gone from all V4 product pages.

The 'V' standing for "Version" in V4 instead of "Vampire" is easy enough. You are correct that the word "Vampire" is missing on the Apollo-core.com website and Apollo hardware. In my opinion, the "Vampire" name sucks for marketing anyway but not as much as the overused "Apollo" name. "Vampire" made sense for the original Vampire 600 model which had teeth that gripped onto a chip.



The Vampire imagery appears to still be used by both groups and resembles a winged Vampire with a 'V' that surely stands for "Vampire". Maybe this is not a problem as long as it is not trademarked although it is being used like a trademark by two entities. Now that I mention it, I wouldn't be surprised if Hyperion soon registers the trademark and holds it hostage from the Apollo-core team and Majsta. Only Amiga makes it possible.

kolla Quote:

And on your context - do you know the chipset+RTG capabilities of other FPGA systems? How do you define "powerful" in that context, and "chipset" for that matter?

All the Amiga chipset cores are in FPGA and only MiSTer has a FPGA that competes with newer Apollo-core hardware. Maybe some FPGA hardware is able to use a RTG ASIC somehow but I expect most RTG implementations are integrated with the Amiga chipset in FPGA as adding chunky support doesn't need much space, at least without a chunky blitter. I expect the memory bandwidth to the chunky buffer/bitmap is the most important performance factor where FPGA hardware isn't too bad (I believe DDR3 memory is used for new Apollo core hardware for example). I'm not so sure a chunky blitter is used much on modern hardware as the 3D hardware is often used for 2D rendering instead using OpenVG or Gallium3D (the latter practically antiquated but worth a mention since AROS uses it?).

https://en.wikipedia.org/wiki/OpenVG
https://en.wikipedia.org/wiki/Mesa_(computer_graphics)#Gallium3D

The 3D hardware itself may just use a pointer to a chunky bitmap, at least with a SoC using a unified memory space for both the CPU and GPU (we have some 3D gurus in the Amiga community who could correct me if I'm wrong). For smaller resolution 2D rendering, I doubt OpenVG would gain much over direct rendering using the CPU, especially with thin Amiga graphics abstraction. More SoC memory bandwidth would make higher resolutions more practical where OpenVG may be worthwhile and of course we are just starting to see primitive 3D support even in FPGA SoCs. I don't foresee a performance bottleneck with an enhanced Amiga chipset even with 3D added but then I'm not a hardware guru. Other perspectives, opinions and knowledge would be interesting.

Last edited by matthey on 30-Aug-2022 at 09:08 PM.
Last edited by matthey on 30-Aug-2022 at 08:50 PM.
Last edited by matthey on 30-Aug-2022 at 08:46 PM.
Last edited by matthey on 30-Aug-2022 at 08:44 PM.

@matthey

Quote:

BusSpeedTest is an old and respected benchmark. It's not some new benchmark created to make Vampire hardware look good.

I never claimed that. There are many variables when it comes to measuring performance.

Quote:

I don't own a V4 to test even if there was a 68k compiled version of that "stream" benchmark

There appears to be a 68k version in the MOS archive on Aminet.

Quote:

. That benchmark must be popular for NG users though as those numbers were given for the X1000 on the Apollo-core forum along with the bandwidth numbers by Vojin Vidanovic.

I've no idea on the popularity, I picked t up from probably the same thread.

Vox's scores are higher than mine because he most likely used the version that Dan had compiled with a more recent version GCC. The link is dead now so I to use V5.4 from back in 2005.

Stream, Ragemem and Gunnars own Sortbench all show X1000 memmory operation in GB/s so why would I trust the outlier that shows only a few hundred MB/s as being the reflection of it's true performance.

Quote:

Vampire/Apollo core hardware has good memory bandwidth which is where much of the performance comes from as it is weak in other areas like CPU clock speed.

Sam440 has faster memory than A1 XE, DDR1 vs SDR but a G4 at 800Mhz still flatens a 440 at 800Mhz. 68080 is a good design and takes advantage of silicon that wasn't available back in the 90's but Gunnars assertion that it means better than an Amigaone is a little off.

@Rob

Quote:

Stream, Ragemem and Gunnars own Sortbench all show X1000 memmory operation in GB/s so why would I trust the outlier that shows only a few hundred MB/s as being the reflection of it's true performance.

As someone who has spent a lot of time profiling, what I will say to this is that unless you know exactly what the tools are measuring you can't trust any of them. Moreover, in cases like this it's possible that the "outlier" is telling the truth, or at least telling something of interest the others are not due to some commonality they share that it doesn't. For memory bandwidth tests, one of the most important factors will be the size of the memory regions involved and whether or not the data cache is impacting it (constructively or destructively, both are possible depending on the exact operation being performed).

_________________
Doing stupid things for fun...

@kolla

Quote:

kolla wrote: @cdimauro Quote: Does it matter? My statement on Vampire was on a very different context, right? Yes, it matters when you use the word "Vampire" when you mean Apollo V4, which isn’t Vampire. This isn’t just you, people have been using "vampire" for anything with apollo core,

Which was the context...
Quote:

while it has always been important to distinguish the two projects, now so even more important.

Now yes, but it wasn't important looking at the context.
Quote:

From what I have seen, the page you link to is an exception, the term "vampire" is gone from all V4 product pages.

Sure. But there were many Vampire boards sold with the Apollo core, which had a subset of features available and yet very powerful compared to other FPGA projects.
Quote:

And on your context - do you know the chipset+RTG capabilities of other FPGA systems?

I've some info about Minimig, MIST, Chameleon .
Quote:

How do you define "powerful" in that context,

The number of available features and how they perform compared each other.
Quote:

and "chipset" for that matter?

The set of hardware features available in the system besides the CPU (albeit here it's tricky, since the Apollo core uses the second CPU hardware thread to implement an "RTG Blitter").


@Karlos

Quote:

Karlos wrote: @Rob Quote: Stream, Ragemem and Gunnars own Sortbench all show X1000 memmory operation in GB/s so why would I trust the outlier that shows only a few hundred MB/s as being the reflection of it's true performance. As someone who has spent a lot of time profiling, what I will say to this is that unless you know exactly what the tools are measuring you can't trust any of them. Moreover, in cases like this it's possible that the "outlier" is telling the truth, or at least telling something of interest the others are not due to some commonality they share that it doesn't. For memory bandwidth tests, one of the most important factors will be the size of the memory regions involved and whether or not the data cache is impacting it (constructively or destructively, both are possible depending on the exact operation being performed).

That's why I repeat since years to do NOT use synthetic benchmarks but only real applications for checking CPUs (and system) performances...

@cdimauro

And now Gunnar has decided that all old "legal" V2 cards no longer are Vampire cards, and instead use the term "Apollo V2". Majsta doesn’t work on Apollo cards, never never has. Apollo V2 cards are doubleplus good.

http://apollo-core.com/knowledge.php?b=1&note=38875

_________________
B5D6A1D019D5D45BCC56F4782AC220D8B3E2A6CC

@kolla

Quote:

kolla wrote: @cdimauro And now Gunnar has decided that all old "legal" V2 cards no longer are Vampire cards, and instead use the term "Apollo V2".

Questionable. From your link:
Quote:

Gunnar von Boehn wrote: There are over 10 thousand Vampire owners worldwide. The 2.16 release gives thousands of people wonderful new features.

So, he continues to call them Vampire. As many users do (see the comments).

Sure, the picture reports "Apollo V2 Accelerators", which looks ambiguous. It's something like Vampire V2 + ApolloCore = Apollo V2 Accelerator.
Quote:

Majsta doesn’t work on Apollo cards, never never has. Apollo V2 cards are doubleplus good. http://apollo-core.com/knowledge.php?b=1¬e=38875

In fact there are no Apollo V2 cards. The picture always talks about "V2 accelerators" which could use the new ApolloCore.

As I've said, it's ambiguous, for sure, but the cards and the softcore are always split all over the picture, besides the title which reports Apollo V2 Accelerators.

@all

This nitpicking Amiga legal shit sucks me up.. I spend my Money elsewhere. Basta!

In Science they say:

Publish or perish!

Sorry I have enough Computers:

I am saturated:

https://www.gebrauchtcomputer24.de/computer.html

Shut down and Offline!

Last edited by QBit on 31-Aug-2022 at 06:01 PM.
Last edited by QBit on 31-Aug-2022 at 05:59 PM.
Last edited by QBit on 31-Aug-2022 at 05:56 PM.
Last edited by QBit on 31-Aug-2022 at 05:53 PM.

@kolla

Yes, we get it!
The official "Vampire" trademark is owned by Majsta, and the going of separate ways of Apollo Team and Majsta means that officially the Apollo boards do not carry the mark of "Vampire".

I too can often be pedantic, but there is this thing called "common usage". It is not a violation of the trademark, nor is it some conversational faux pas to refer to all Apollo FPGA boards as 'Vampire' boards outside of a commercial context.

The same way it is acceptable common usage to refer to sparkling white wine as 'champagne' even if it is not from the Champagne region.

Last edited by agami on 01-Sep-2022 at 04:38 AM.

_________________
All the way, with 68k

@cdimauro

“This brings your Apollo V2 closer to the new V4 series…”

My what? I have two Vampire V2 accellerator cards. Why is the word “Vampire” left out when these are the actual Vampire products? And at the same time, what you could call the “collective vampire” term is used also for V4, because old habits? It is as if they are afraid of using “vampire” in any sort of advertisement, yet want to “own” the term, in daily speak. Passive-aggressive marketing.

I also have the original Minimig, MiST, FleaFPGA and MiSTer, and know first hand what all these are capable of in terms of chipset+RTG vs the actual Vampire products (not much capable of anything chipset wise, though ok-ish RTG).

How well does the Apollo V4 SAGA work with HighGFx monitor definitions? Can it do Amiga native 1280x720 modes for legacy software titles like DPaint?

Last edited by kolla on 01-Sep-2022 at 03:50 PM.

_________________
B5D6A1D019D5D45BCC56F4782AC220D8B3E2A6CC

@kolla

Quote:

kolla wrote: @cdimauro “This brings your Apollo V2 closer to the new V4 series…” My what? I have two Vampire V2 accellerator cards. Why is the word “Vampire” left out when these are the actual Vampire products? And at the same time, what you could call the “collective vampire” term is used also for V4, because old habits? It is as if they are afraid of using “vampire” in any sort of advertisement, yet want to “own” the term, in daily speak. Passive-aggressive marketing.

Sorry, kolla, but I don't understand where do you want to go with those continues changes of topics (after that you got proper replies).

BTW, why don't you ask those questions to Gunnar?
Quote:

I also have the original Minimig, MiST, FleaFPGA and MiSTer, and know first hand what all these are capable of in terms of chipset+RTG vs the actual Vampire products (not much capable of anything chipset wise, though ok-ish RTG).

OK, and? Care to give a precise comparisons of all of those products? A nice table with the list of features in the rows and the list of products in the columns, and cells representing the status (implemented, non implemented, missing X, Y or Z, etc).
Quote:

How well does the Apollo V4 SAGA work with HighGFx monitor definitions? Can it do Amiga native 1280x720 modes for legacy software titles like DPaint?

I know a guy which says that he owns two vampire cards and that could give a try.

Besides that, I've read in the Apollo forum that people has tested it in FullHD but with 256 colours. But it's the Apollo forum and I have no vampire card to check it.

@cdimauro

I’m not the one making claims about systems I don’t have access to, and it’s not my job to be your data source - the solution is simple enough, you just buy the hardware you want to compare. Buy some V2 accellerators, buy a V4, buy a MiST clone, buy a Minimig, buy a MiSTer. Then compare the lot and and now you can post some well backed up opinions about these systems. Bonus, you now have tons of hardware to test out your code and games on.

FullHD 1280x720 256 colour RTG chunky mode, or native AGA 8bit planar mode?
I speak of the latter.

Last edited by kolla on 02-Sep-2022 at 08:37 PM.

_________________
B5D6A1D019D5D45BCC56F4782AC220D8B3E2A6CC

@kolla

Quote:

kolla wrote: @cdimauro I’m not the one making claims about systems I don’t have access to, and it’s not my job to be your data source - the solution is simple enough, you just buy the hardware you want to compare. Buy some V2 accellerators, buy a V4, buy a MiST clone, buy a Minimig, buy a MiSTer. Then compare the lot and and now you can post some well backed up opinions about these systems. Bonus, you now have tons of hardware to test out your code and games on.

So, you have all those systems, you said that the ApolloCore-equipped boards (to avoid using the Vampire word, since it's "distracting" you) aren't the more advanced FPGA systems for Amigas, but you don't want to give any features-list for comparing them...

Well, I will NOT buy any of those because to me it's enough reading their specs.
Quote:

FullHD 1280x720 256 colour RTG chunky mode, or native AGA 8bit planar mode? I speak of the latter.

I spoke of the former.

@cdimauro

Quote:

cdimauro wrote: @kolla Quote: kolla wrote: @cdimauro I’m not the one making claims about systems I don’t have access to, and it’s not my job to be your data source - the solution is simple enough, you just buy the hardware you want to compare. Buy some V2 accellerators, buy a V4, buy a MiST clone, buy a Minimig, buy a MiSTer. Then compare the lot and and now you can post some well backed up opinions about these systems. Bonus, you now have tons of hardware to test out your code and games on. So, you have all those systems, you said that the ApolloCore-equipped boards (to avoid using the Vampire word, since it's "distracting" you) aren't the more advanced FPGA systems for Amigas, but you don't want to give any features-list for comparing them...

“most advanced FPGA systems for Amigas”? Or just this:

Quote:

The Vampire, which has the most powerful Amiga chipset + RTG implementation

to which I replied - V4 maybe, but certainly not the actual Vampire cards.

Quote:

Well, I will NOT buy any of those because to me it's enough reading their specs. Quote: FullHD 1280x720 256 colour RTG chunky mode, or native AGA 8bit planar mode? I speak of the latter. I spoke of the former.

And from reading the apollo forums your impression is that this is what Vampire 2 cards can do?

The MiST (and its clones) can do 1280x720 8bit RTG. Vampire V2 boards can even do 1280x720 32bit RTG - but the only Apollo core with actual chipset (severely limited and broken AGA) for Vampire V2 cannot do any RTG (known as the "gold 3 beta" core). MiSTer also does 1280x720 32bit RTG and native 1280x720 with AGA, also HAM modes, using the HighGfx monitor definitions (originally made for Indivision) Can V4 do that? I don’t know.

* http://aminet.net/package/driver/moni/HighGFXnmore

_________________
B5D6A1D019D5D45BCC56F4782AC220D8B3E2A6CC

@kolla

Quote:

kolla wrote: @cdimauro Quote: cdimauro wrote: @kolla So, you have all those systems, you said that the ApolloCore-equipped boards (to avoid using the Vampire word, since it's "distracting" you) aren't the more advanced FPGA systems for Amigas, but you don't want to give any features-list for comparing them... “most advanced FPGA systems for Amigas”? Or just this: Quote: The Vampire, which has the most powerful Amiga chipset + RTG implementation to which I replied - V4 maybe, but certainly not the actual Vampire cards.

So, are you telling that Vampire's V2 competitors are doing better? Source(s) for this?
Quote:

Quote: I spoke of the former. And from reading the apollo forums your impression is that this is what Vampire 2 cards can do? The MiST (and its clones) can do 1280x720 8bit RTG. Vampire V2 boards can even do 1280x720 32bit RTG - but the only Apollo core with actual chipset (severely limited and broken AGA) for Vampire V2 cannot do any RTG (known as the "gold 3 beta" core).

So you've bought TWO Vampire V2 cards which aren't working and I don't understand why you decided to fail two times.

Anyway: http://apollo-core.com/knowledge.php?b=1¬e=38875&x=1&z=-RhaqU

Antonio la Marca

Posts 10
03 Sep 2022 22:49

Hello ,i have the new 2.16 Core Flashed to my Vampire V500 and the result was thats all ist function but all RTG Compatible Demos have red color and disorted graphic !
I flashed back to Core 2.15 with USB-Blaster and now all RTG Demos graphics is OK !

So, if it's not working for you it's maybe you've used the wrong firmware. Because for that guy RTG works with the 2.15 one.

Who's right then?
Quote:

MiSTer also does 1280x720 32bit RTG and native 1280x720 with AGA, also HAM modes, using the HighGfx monitor definitions (originally made for Indivision) Can V4 do that? I don’t know.

Me neither, but see above.
Quote:

* http://aminet.net/package/driver/moni/HighGFXnmore

OK, and? What's the problem? Is it the ECS chipset implementation which isn't working? Which looks strange again, since there were some threads where it was shown how to create new video modes with specific resolutions.

@cdimauro

What? You really struggle so hard to read? Didn’t I write that the V2 cards can do 1280x720 32bit? Yes I did. Did I write that the Vampire V2 cards don’t work for me? No, they have stability issues, but they are working. Am I saying that there are other FPGA solutions that are better than V2 for RTG+chipset? Yes, certainly - as they actually HAVE chipset implemented on FPGA - which the 2.x series pf Apollo Core DO NOT HAVE - maybe that is what flew above your head? With V2 there is ONLY RTG, all native modes use the actual chipset on the Amiga motherboard, NOT FPGA!

This is why you should buy the products yourself, so this confusion can end.

_________________
B5D6A1D019D5D45BCC56F4782AC220D8B3E2A6CC

@kolla

Quote:

kolla wrote: @cdimauro What? You really struggle so hard to read?

I don't.
Quote:

Didn’t I write that the V2 cards can do 1280x720 32bit? Yes I did. Did I write that the Vampire V2 cards don’t work for me? No, they have stability issues, but they are working.

Let me quote yourself again:

Vampire V2 cannot do any RTG

You really struggle so hard to read... yourself!
Quote:

Am I saying that there are other FPGA solutions that are better than V2 for RTG+chipset? Yes, certainly - as they actually HAVE chipset implemented on FPGA - which the 2.x series pf Apollo Core DO NOT HAVE - maybe that is what flew above your head? With V2 there is ONLY RTG, all native modes use the actual chipset on the Amiga motherboard, NOT FPGA! This is why you should buy the products yourself, so this confusion can end.

That's wrong (again): it depends on the firmware.

2.x doesn't support AGA. 3.x supports it.

PADDING V4

_________________
I HAVE ABS OF STEEL
--
CAN YOU SEE ME? CAN YOU HEAR ME? OK FOR WORK