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!

Apollo Hardfpu Resultspage  1 2 3 4 5 6 7 8 

Gunnar von Boehn
(Apollo Team Member)
Posts 6207
03 Oct 2017 18:34


Hi,
     
New performance Updates:
 
 
 
 

    As you know the APOLLO HARD-FPU is coming for Vampire4.
    We will post here some early benchmarks.
     
   
     
   
     
   
 
     
    Please mind, that these results are preliminary and will likely improve soon. Mind the HARD-FPU is not fully tuned yet.


Vojin Vidanovic

Posts 770
03 Oct 2017 18:49


Nice, especially 20 000 Mhz (relative to 68 000?) >:-)

I ve seen first Sysinfo results on FB
EXTERNAL LINK


Lord Aga
(Apollo Team Member)
Posts 119
03 Oct 2017 18:51


This will be HARD for some people all right :)


Niclas A
(Apollo Team Member)
Posts 219
03 Oct 2017 19:31


Insane.. But in a good way!


Gunnar von Boehn
(Apollo Team Member)
Posts 6207
03 Oct 2017 21:26


Update faster FPU now 60 MFLOPS in SYSINFO




Hbarra 2Pi

Posts 8
03 Oct 2017 21:43


Wow! Give me a dozen, please!
Bravo!
Great Job.


Mo Retro

Posts 241
03 Oct 2017 22:43


Great work, let the fight begin ;)


Rod March

Posts 119
03 Oct 2017 23:31


Just epic.



Renee Cousins
(Apollo Team Member)
Posts 142
03 Oct 2017 23:48


Augh! This is like watching the food network while being on a diet. "Now fold the butter gently into the melted chocolate."

Mmmm, MFLOPS....


Eric Gus

Posts 477
04 Oct 2017 00:19


Nice!..

the FB amiga forums were eerily quiet and lacked the usual amount of any vampire rhetoric they usually have today when someone posted this.. though there were the comments about the test being invalid due to timers and other such "well there must be a bug" type comments.


Thierry Atheist

Posts 644
04 Oct 2017 01:15


BRAIN HAS BEEN BLASTED TO NEXT NEAREST SOLAR SYSTEM!!!

! ! ! W O W ! ! !

WE'RE BACK, BABY!!!!!!


Wawa T

Posts 695
04 Oct 2017 03:01


i actually expected v4 would have some edge over v2 in cpu speed itself.


Ian Parsons

Posts 230
04 Oct 2017 03:44


I guess it depends on what it's clocked at. It's probably being used in accelerator configuration currently with a multiplier of the base 7.09MHz CPU clock signal. It's been said many times that the Cyclone 5 would only give a modest boost in integer performance.


Thierry Atheist

Posts 644
04 Oct 2017 03:54


wawa t wrote:
i actually expected v4 would have some edge over v2 in cpu speed itself.

Hi wawa,

The highest MIPS rating was 160. The Vampire 4 is supposed to be about 25% faster. Even though we saw 113 MIPS in the picture, it should be MUCH higher. And the corresponding MFLOPs should reflect that too.



Sean Sk

Posts 488
04 Oct 2017 05:58


Renee Cousins wrote:

Augh! This is like watching the food network while being on a diet. "Now fold the butter gently into the melted chocolate."
 
  Mmmm, MFLOPS....

Hahaha this is the funniest thing ever....yet so true....  :(


Nixus Minimax

Posts 416
04 Oct 2017 13:36


wawa t wrote:

i actually expected v4 would have some edge over v2 in cpu speed itself.

This is still running on the v2.


Cunn Pole

Posts 29
04 Oct 2017 14:06


Doesn't that not contradict the first statement from Gunnar: "the APOLLO HARD-FPU is coming for Vampire4."


Sean Sk

Posts 488
04 Oct 2017 14:09


Nixus Minimax wrote:
 
  This is still running on the v2.
 

 
  It is? I don't think so. I thought this was running on a V4 since there was not going to be a hard FPU for the V2, only a soft FPU.


Nixus Minimax

Posts 416
04 Oct 2017 14:29


Cunn Pole wrote:

Doesn't that not contradict the first statement from Gunnar: "the APOLLO HARD-FPU is coming for Vampire4."

Not necessarily. The hard-FPU and everything else fit the v4 FPGA easily but even though these tests are being done on the v2 (a well-known reliable hardware suitable for testing), some features planned for v2 Gold3/Platinum are probably missing or removed in order to make space for the FPU in the v2.

Also note that Gunnar said that the v4 FPU would do one or even two FLOPS per clock cycle. The results shown above are still below that meaning that there is still work to be done.


Martin Soerensen

Posts 232
04 Oct 2017 15:20


Gunnar von Boehn wrote:
  Please mind, that these results are preliminary and will likely improve soon. Mind the HARD-FPU is not fully tuned yet.

Yes you're right, those numbers are terrible. :-)

Too bad the FPGA on V2 does not have space for this as it looks really amazing. But at least we have the soft-emu which is also nice, although it will never see this kind of performance.

posts 158page  1 2 3 4 5 6 7 8