Beat It - If You Can | page 1 2
|
---|
|
---|
| | Gunnar von Boehn (Apollo Team Member) Posts 6254 16 Oct 2017 21:10
| genuine AMIGA 600 with VAMPIRE / 68080 running AIBB Beachball
| |
| | Vojin Vidanovic
Posts 770 16 Oct 2017 21:35
| Its nice to say Vamp accels A600 by factor x1324 and beats A4000-040 just by 155 times ... Sounds like MJs commercial of the days, before he started to be a white girl, member? EXTERNAL LINK
| |
| | Markus B
Posts 209 17 Oct 2017 08:38
| Vojin Vidanovic wrote:
| Its nice to say Vamp accels A600 by factor x1324 and beats A4000-040 just by 155 times ...
|
I think you misinterpreted the numbers. They're all related to the base, so Vampire is 1324x as fast as the A600 (which has no FPU) and a 68040 is 155x as fast. So the Vampire is 8,5x faster than the 040.
| |
| | Peter Heginbotham
Posts 214 17 Oct 2017 11:52
| very nice, for context it's 4.16 faster than a 060 EXTERNAL LINK CyberStorm II (060/50, Piccolo SD64, 3.1 in RAM) 318.06
| |
| | Thierry Atheist
Posts 644 17 Oct 2017 11:56
| Markus B wrote:
| I think you misinterpreted the numbers. They're all related to the base, so Vampire is 1324x as fast as the A600 (which has no FPU) and a 68040 is 155x as fast. So the Vampire is 8,5x faster than the 040. |
And since a 68060 is about 3 times faster than a 68040;Vampire 2 = 1324.45/(155.97*3) Vampire 2 = 2.83 times faster than a 50MHz 68060 Vampire 2 =~141.5 MHz 68060 However Vampire 2 has many other enhancements that are far beyond what a 68060 is capable of. That makes the Vampire 2 as much as DOUBLE to TRIPLE the performance of a 68080 running at 141.5 MHz!!!!!!
| |
| | Markus B
Posts 209 17 Oct 2017 12:26
| Do all the other FPU tests have similar results compared to the 040 results?
| |
|
| | Peter Heginbotham
Posts 214 17 Oct 2017 13:41
| so now you have blown away the 060 fpu what are you doing this afternoon :) EXTERNAL LINK
| |
| | Claudio Guglielmotti (Apollo Team Member) Posts 185 17 Oct 2017 14:28
| catching bugs EXTERNAL LINK
| |
| | Markus B
Posts 209 17 Oct 2017 15:31
| Gunnar von Boehn wrote:
| Yes, we Dpainted then all between 8x to 12x 68040 speed.
|
Good DPaint skills! The manipulation is barely noticable.
| |
| | Szyk Cech
Posts 191 18 Oct 2017 09:51
| Some times when I open this forum page I feel like a dream: How is this possible to reach such outstanding test results?!? It is just "modern technology" or maybe some genius behind this?!? In true I think: both...
| |
| | Gunnar von Boehn (Apollo Team Member) Posts 6254 18 Oct 2017 10:10
| Szyk Cech wrote:
| Some times when I open this forum page I feel like a dream: How is this possible to reach such outstanding test results?!? It is just "modern technology" or maybe some genius behind this?!? In true I think: both... |
Shall I tell you something sad..FPGA technology that we use today in the VAMPIRE is available since the year 2002. In the theory 68080 CPU cards for AMIGA could have been build already then. Of course the Cyclone FPGA at that time was little bit more expensive, so the VAMP would have cost 2 or 3 times the money in year 2002/3.
| |
| | Andrew Copland
Posts 113 18 Oct 2017 10:13
| Gunnar von Boehn wrote:
| Shall I tell you something sad.. FPGA technology that we use today in the VAMPIRE is available since the year 2002. In the theory 68080 CPU cards for AMIGA could have been build already then. Of course the Cyclone FPGA at that time was little bit more expensive, so the VAMP would have cost 2 or 3 times the money in year 2002/3.
|
"The future is already here it is just not evenly distributed" -William Gibson
| |
| | Markus B
Posts 209 18 Oct 2017 10:54
| Is there a chance to run some other FPU application for comparison? There is a collection of Cinema4D results here. Maybe we can extend them with FPU results.
| |
| | Gunnar von Boehn (Apollo Team Member) Posts 6254 18 Oct 2017 11:04
| Markus B wrote:
| Is there a chance to run some other FPU application for comparison? There is a collection of Cinema4D results here. Maybe we can extend them with FPU results.
|
Sure will all come.. But please mind the FPU runs still in "slow/debug" mode. With proper code and full FPU speed features turned on, results can be 3-4 times faster.
| |
| | Markus B
Posts 209 18 Oct 2017 12:03
| Gunnar von Boehn wrote:
| Sure will all come.. But please mind the FPU runs still in "slow/debug" mode. With proper code and full FPU speed features turned on, results can be 3-4 times faster.
|
Thumbs up!
| |
| | Nixus Minimax
Posts 416 18 Oct 2017 12:29
| Szyk Cech wrote:
| How is this possible to reach such outstanding test results?!? It is just "modern technology" or maybe some genius behind this?!? In true I think: both... |
It is not only modern technology and genius (of which there is plenty involved) but also a lot of hard work and dedication. Not only hard work by the geniuses at the core of the project but also by the support people around them. It's just too easy to hide behind one's own lack of genius and use it as a reason to not contribute.
| |
| | Samuel Devulder
Posts 248 18 Oct 2017 14:43
| Is this Soft-FPU or Hard-FPU ?
| |
| | Martin Soerensen
Posts 232 18 Oct 2017 15:11
| Samuel Devulder wrote:
| Is this Soft-FPU or Hard-FPU ?
|
Hard. FEMU is still slower than a regular 040 FPU, although it is moving closer. No soft FPU on the Vampire will ever be this fast. :-)
| |
| | Gunnar von Boehn (Apollo Team Member) Posts 6254 18 Oct 2017 15:50
| Martin Soerensen wrote:
|
Samuel Devulder wrote:
| Is this Soft-FPU or Hard-FPU ? |
Hard. FEMU is still slower than a regular 040 FPU, although it is moving closer. |
Where did you see this?
| |
|