Overview Features Instructions Performance Forum Downloads Products OrderV4 Reseller Contact

Welcome to the Apollo Forum

This forum is for people interested in the APOLLO CPU.
Please read the forum usage manual.



All TopicsNewsPerformanceGamesDemosApolloVampireAROSWorkbenchATARIReleases
Performance and Benchmark Results!

Benchmark FxPaint 2.02R - IOBench

Simo Koivukoski
(Apollo Team Member)
Posts 588
03 Nov 2017 20:02


Gold2.7/x11, femu 0.11-WIP : FxPaint 2.0 Update 202R / 040 build.
 
Menu -> Plugins -> IOBench:
 



Mr Niding

Posts 452
04 Nov 2017 08:03


Nifty!

Anyone have comparative stats from non-Vampire machines to evaulate the benchmarks ShK put up here?


Simo Koivukoski
(Apollo Team Member)
Posts 588
04 Nov 2017 11:09


Mr Niding wrote:

Nifty!
 
  Anyone have comparative stats from non-Vampire machines to evaulate the benchmarks ShK put up here?

There is some results: EXTERNAL LINK


Saladriel Amrael

Posts 148
04 Nov 2017 11:12


I think it is mandatory to subscribe in order to see the pictures


Vojin Vidanovic

Posts 770
04 Nov 2017 12:26


Saladriel Amrael wrote:

I think it is mandatory to subscribe in order to see the pictures

Yup, one needs to be registered German-only forum user.

Can you please convert both Vamp and non Vamp results to a plain text?


Simo Koivukoski
(Apollo Team Member)
Posts 588
04 Nov 2017 12:42


Gold2.7/x11, femu 0.11-WIP : FxPaint 2.0 Update 202R / 040 build
I/O: Loading picture:      0,11 sec
I/O: Saving picture:      0,38 sec
F/X: Zoom blur:          15,66 sec
F/X: Waves:              202,29 sec
F/X: Lightsource:          9,59 sec
F/X: Pixelise:            2,82 sec
MEM: Byte fill 100:        1,17 sec
DRW: Ellipse 'Silent':    5,84 sec
DRW: Lines 'Silent':      2,12 sec



Vojin Vidanovic

Posts 770
04 Nov 2017 13:01


Thanks Simo, but it was more about benchmark results
from other Amigas, from German forum (hope some member
can do it)


Michael AMike

Posts 142
04 Nov 2017 14:01


Simo Koivukoski wrote:

  Gold2.7/x11, femu 0.11-WIP : FxPaint 2.0 Update 202R / 040 build
I/O: Loading picture:      0,11 sec
  I/O: Saving picture:      0,38 sec
  F/X: Zoom blur:          15,66 sec
  F/X: Waves:              202,29 sec
  F/X: Lightsource:        9,59 sec
  F/X: Pixelise:                2,82 sec
  MEM: Byte fill 100        1,17 sec
  DRW: Ellipse 'Silent':    5,84 sec
  DRW: Lines 'Silent':      2,12 sec

 

 
  Hi Simo, thanks for the benchmark. Is better than I expected despite the fact that only a X11 Core is used. Here are the linked benchmarks of a1k.
 
  A4000_060_PPC_366Mhz
 

  I/O: Loading picture:      0,07 sec
  I/O: Saving picture:      0,14 sec
  F/X: Zoom blur:            4,53sec
  F/X: Waves:                  9,93 sec
  F/X: Lightsource:          9,93 sec
  F/X: Pixelise:                2,55 sec
  MEM: Byte fill 100:        0,79 sec
  DRW: Ellipse 'Silent':    5,85 sec
  DRW: Lines 'Silent':      2,11 sec

 
  A4000_060_66Mhz
 

  I/O: Loading picture:      0,15 sec
  I/O: Saving picture:      1,56 sec
  F/X: Zoom blur:          20,33 sec
  F/X: Waves:                111,66 sec
  F/X: Lightsource:        37,80 sec
  F/X: Pixelise:                8,96 sec
  MEM: Byte fill 100:        0,79 sec
  DRW: Ellipse 'Silent':    5,85 sec
  DRW: Lines 'Silent':      2,12 sec

 


Nicolas Sipieter
(Needs Verification)
Posts 115/ 1
04 Nov 2017 14:19



amike was faster than me, but since i took the time to make the document, i link to it anyway.
for those that don't want to register on a1k forums:

EXTERNAL LINK


Simo Koivukoski
(Apollo Team Member)
Posts 588
04 Nov 2017 14:29


Thanks guys for these comparison results! :)


Samuel Devulder

Posts 246
04 Nov 2017 14:39


Waves F/X is noticeably slower on the vampire than on the 060. Is there any reason for that like a missing fpu opcode that take ages by FEmu to emulate ?


Niclas A
(Apollo Team Member)
Posts 218
04 Nov 2017 15:17


EXTERNAL LINK


Gunnar von Boehn
(Apollo Team Member)
Posts 4972
04 Nov 2017 15:20


Samuel Devulder wrote:

  Waves F/X is noticeably slower on the vampire than on the 060.

 
I think in many tests the Vamp is faster.
 
 


Vojin Vidanovic

Posts 770
04 Nov 2017 15:25


Gunnar von Boehn wrote:
  I think in many tests the Vamp is faster. 

If "lower is better" yes, Vamp outclasses 060, somethimes coming close to PPC results. But in that one tests,lags behind.

      202,29 sec (Vamp)    111,66 sec (060 66Mhz)

It will be interesting to see these "gap" tests redone once v4
with full FPU is out. Outclassing BPPC would be nice :-)




Gunnar von Boehn
(Apollo Team Member)
Posts 4972
04 Nov 2017 15:30


Niclas A wrote:

EXTERNAL LINK 

Your image is not correct, for example Zoom bars look wrong.


Saladriel Amrael

Posts 148
04 Nov 2017 15:31


Thanx for sharing the results.
  The one thing that surprise me is the "Byte Fill 100": on the 060 is as fast as on PPC, while Vampire results much slower.
 
  Sorry if dumb question but shouldn't Vampire be much faster than 060 in memory fill operations? (Not wanting to complain, just trying to understand, maybe this is not even a memory fill operation as the name suggests).

For the rest it looks pretty much impressive


Niclas A
(Apollo Team Member)
Posts 218
04 Nov 2017 15:43


Gunnar von Boehn wrote:

 
Niclas A wrote:

  EXTERNAL LINK   

  Your image is not correct, for example Zoom bars look wrong.
 

 
  Yes. Too quick for my own best :P
  Updated now. Sorry.
 
  Hate google docs. Dont have Excel installed right now.

Always hard to make nice charts with huge difference in numbers.
Easy in Excel to put some data on second axis. Cant find that in google docs.


Michael AMike

Posts 142
04 Nov 2017 16:25


nicolas sipieter wrote:

  amike was faster than me, but since i took the time to make the document, i link to it anyway.
  for those that don't want to register on a1k forums:
 
  EXTERNAL LINK 

Perfect - thanks :)


Michael AMike

Posts 142
05 Nov 2017 21:06


I've made some charts
 



Michael AMike

Posts 142
05 Nov 2017 21:07


I've made some charts
   
EXTERNAL LINK  EXTERNAL LINK  EXTERNAL LINK 

posts 20