Overview Features Coding ApolloOS Performance Forum Downloads Products Order Contact

Welcome to the Apollo Forum

This forum is for people interested in the APOLLO CPU.
Please read the forum usage manual.
Please visit our Apollo-Discord Server for support.



All TopicsNewsPerformanceGamesDemosApolloVampireAROSWorkbenchATARIReleases
Performance and Benchmark Results!

300Mhz Targetpage  1 2 3 4 

Mario Angelucci

Posts 11
28 Aug 2018 18:50


Hi all, I'm following this project since it was called Natami, Apollo team has done an amazing and remarkable job. Truly.
  There is one "but" though.
  The "but" is that what is it for having such a performance in cpu if all we need to run is 25 years old software ? Old 060 040 or even 030 accelerators are far more than enough for that such an old software. The most uptodated software to run on Apollo Vampire can be found on Macintosh emulation side, that is 20 years old only !!!!.....What I mean is that once we have an 080 we need also software to use nowadays and web browsing only is quite heavy now for a 90Mhzish clocked cpu. Please do not missunderstand me, I'll buy one vampire standalone as it is available in view of I already own A1200/030 and A4000/060. The point is that we need raw power more than SAGA, any discrete graphic card, on a fast expansion slot, can be hundredfold faster than SAGA....and also can be cheap if we could use common marketplace graphic cards. For a start ? A 300Mhz+ running 080 maybe with a different FPGA but it can be even simpler in view of it has to work as AGA only as graphic counterpart. Just an idea. Have fun all :)


Hugo Pereira

Posts 72
28 Aug 2018 19:22


Hello all Amigans! When the goal is to have an actual Amiga (native hardware) without emulation, too much stone has to be broken. I trust the safe steps (AMMX, OoOE, etc.) that the Apollo Team has been giving. This way we already have a very powerful FPGA 080 and in the future we will have one of the best systems of the present time. It must be taken into account that this process is evolving and have the patience to take the train that the Amiga had lost in the last decades.
Congratulations Apollo Team.


Vojin Vidanovic
(Needs Verification)
Posts 1916/ 1
28 Aug 2018 20:03


Mario Angelucci wrote:

    A 300Mhz+ running 080 maybe with a different FPGA but it can be even simpler in view of it has to work as AGA only as graphic counterpart. Just an idea. Have fun all :)
   

   
    I am sure some V5 Vamp might run 300Mhz, without ASIC.
   
    Current V4 are ~110-120Mhz and coupled with AMMX, faster RAM, improved cores, larger cache it gives a decent performance TO BE USED. There is nothing old that can choke it, and it should handle current trouble children like NeoGeo emulator, PC Task 486 performance and NetSurf. Hope laisy AEON could deliver 68k Enhancer and AmiStore promised, since there is enough horsepower now, even for Odyssey backport.
 
  To I, some simple OpenGL compatibility or Warp3D in hardware, would enable more older software and more development and backports then solely x3 CPU horsepower with AGA and P96 only.
 
  Why would you need that 300Mhz NOW (in V5 Vamp future days we might have an app or backport of Quake 3 or similar)?


Matthew Burroughs

Posts 59
28 Aug 2018 20:10


Until i read about the Vampire i knew nothing of FPGA

Whatever way you look at it, an FPGA based system is delivering results cheaper and far, far quicker then a PPC based upgrade.




Vojin Vidanovic
(Needs Verification)
Posts 1916/ 1
28 Aug 2018 20:23


Matthew Burroughs wrote:

  Whatever way you look at it, an FPGA based system is delivering results cheaper and far, far quicker then a PPC based upgrade.

We have all seen Acubes A500 FPGA implementation and other FPGA systems, but only Vampire demonstrated new development and not the retro cores.



Mario Angelucci

Posts 11
28 Aug 2018 20:28


I agree 100% but nowadays what's it for to have a pctask as fast as 385 25Mh ?!? I mean, once You have done something on it You fall back on using your x86 octacore cpu pc system to.......to read your eamils, to web surfing, to play......I'm sure that an amiga capable of a 300+Mhzish 080 power we could start to ditch our x86 pc system. Otherwise is just a (very nice and amusing) nostalgia affair.


Matthew Burroughs

Posts 59
28 Aug 2018 20:45


nostalgia is a huge part of peoples interest in the Amiga.

While i don't ever think that the Amiga is going to rule the high street again, its hobbiest appeal is apparent for a machine that has been "dead" for over 20 years.




Vojin Vidanovic
(Needs Verification)
Posts 1916/ 1
28 Aug 2018 21:01


Matthew Burroughs wrote:

  While i don't ever think that the Amiga is going to rule the high street again, its hobbiest appeal is apparent for a machine that has been "dead" for over 20 years.
 

 
  It finally raises, not us mummy, not as zombie, as a Vampire.

Why rushing it? We need time to "drink blood" and regain strenght (68k software and eco system).


Mario Angelucci

Posts 11
28 Aug 2018 22:18


RULEZ ! :)



Mario Angelucci

Posts 11
28 Aug 2018 22:19


ok 500+Mhz 080 cpu power :)


Vojin Vidanovic
(Needs Verification)
Posts 1916/ 1
28 Aug 2018 22:23


I agree "life starts after native 200Mhz" but understand its a evolution, not a revolution. Something others have failed often promised too much, delivering gimmick on s/w or h/w side.

And in my eyes, jumps from x10 to x15 (v4) + core evolution is way more anyone has done, so show some ... humanity, patience and gratitude.


Mario Angelucci

Posts 11
28 Aug 2018 22:25


agree


Nicolas Sipieter
(Needs Verification)
Posts 115/ 1
28 Aug 2018 23:12


nooo friend,
you see this project is nice but i know how it can get better.
you know what would be real cool ? apollo-core make new transition to become the first bicycle cpu. it implement a x86 and a 68k and it execute each one after the other in cycle.
so you can have two live OS running, each running on his own screen.
one eye on each screen you can play two games in the same time (with free roto-zoom), so you get x2 free time.
it's the future. and now the team will follow a new direction i'm sure.

roadmap for "M-apollo":
make a x86 fpga
put it in a room, alone, with apollo-core .. put nice music .. and let nature do it's job.
tomorrow you have mixed instruction set (misc).
in 9months i know it can happen.



Roy Gillotti

Posts 517
28 Aug 2018 23:40


^ I'm sure Gunnar would confirm all that.


Mr-Z EdgeOfPanic

Posts 189
29 Aug 2018 05:24


For the next few years FPGA's that would give us 300 Mhz cores will stay very expensive.
As much as i would like to have that extra raw horsepower i do not see it happening soon.
Cards like that will be far beyond the 1000 euro mark i'm afraid.



Gunnar von Boehn
(Apollo Team Member)
Posts 6207
29 Aug 2018 05:47


Mr-Z EdgeOfPanic wrote:

  For the next few years FPGA's that would give us 300 Mhz cores will stay very expensive.
  As much as i would like to have that extra raw horsepower i do not see it happening soon.
  Cards like that will be far beyond the 1000 euro mark i'm afraid.
 

 
This is 100% correct.
 
FPGA technology offers the possibility to be reconfigured, but FPGA by design do never offer high clockrate.
 
Nevertheless APOLLO is very fast, lets compare it with the original 68000. When you compare APOLLO with another 68000 then you can see the following.
 
Example instruction:
ADDI.l #$121234,40(a0)
Needs ~ 32 clock on Motorola 68000.
APOLLO 68080 needs 0.5 cycle for the same instruction.
 
Lets say your Vamp runs at 85 MHz, now do he math yourself and tell me what 68000 clockrate that equals to for this instruction?
 
 


Mario Angelucci

Posts 11
29 Aug 2018 09:09


How fast can we expect the future V4 be clocked ? Just a safe range


Renee Cousins
(Apollo Team Member)
Posts 142
30 Aug 2018 01:44


Mario Angelucci wrote:

  How fast can we expect the future V4 be clocked ? Just a safe range
 

I'd like to hope for a 50% bump. I think that's achievable.


M Rickan

Posts 177
30 Aug 2018 01:52


Mr-Z EdgeOfPanic wrote:

As much as i would like to have that extra raw horsepower i do not see it happening soon.

Cards like that will be far beyond the 1000 euro mark i'm afraid.

That may be a given but if someone were to bankroll a high end FPGA board that used the Apollo core you can be sure that there would be a market for it.


Gregthe Canuck

Posts 274
30 Aug 2018 04:49


Renee Cousins wrote:

I'd like to hope for a 50% bump. I think that's achievable.

 
I don't think it is realistic to expect that much of a clock speed jump. I believe a 14x core was demoed which is a nice bump up from the 12x on V2. But that isn't guaranteed. *Maybe* if at some point the team does a "black edition" with a higher speed grade of the chip used in the V4 they could maybe go one speed level higher? But as Majsta has noted earlier the speed grading is a bit of a crapshoot.
 
The best thing the team can do is keep working on the efficiency per clock cycle. That is where the "win" is. The new Cyclone V has a lot more logic elements to make that possible. And the DDR3 memory will help as well. And of course other core enhancements, many of which will make it back into the V2 (such as OoO).
 
In theory, given the moderately higher potential clock rate (14x vs 12x) and other core improvements the V4 could have much faster performance than the V2. But I still think 50% is stretching it. But then again who I am to doubt the Vampire wizards?
 
  Cheers
 

posts 68page  1 2 3 4