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!

Talk About Sysinfopage  1 2 3 

Adam A

Posts 106
11 Jun 2019 17:03


its a a3640 card with 68060 (rev6) using cpu adapter
 
  EXTERNAL LINK


A1200 Coder

Posts 67
11 Jun 2019 18:59


But this can't be right for 68060 results, according to your Sysinfo results it has about the same performance as a 68040 at same clock speed. Or 68040@50 MHz is the same as 68060@50 MHz with integer performance.

SysSpeed seems to give 65 MIPS for 68060@50 MHz, but it is also flawed in many ways, seems to overestimate most CPU's performance and scores not comparable to Sysinfo.


Knight Stone

Posts 119
11 Jun 2019 19:20


Niclas A wrote:

Nice :)
  Can someone with a 060 also run this version so we can see how it behaves for that too?

  i would be happy to run it on an A1200 with Apollo 1260 @66mhz.

is there a download link?




Vojin Vidanovic

Posts 1719
11 Jun 2019 19:52


Knight Stone wrote:

 
  i would be happy to run it on an A1200 with Apollo 1260 @66mhz.
 
  is there a download link?
 
 

Yes, Gunnar posted it in this thread.
CLICK HERE


Knight Stone

Posts 119
11 Jun 2019 19:53


A1200 1D4 Board. Apollo 1260 68060 (50mhz part,@ 66mhz) full 68060 FPU and MMU, 64MB RAM


49,804  Dhrystones.
51,98  MIPS
37.07  MFlops

Hope this helps.



Estrayk / PaRaDoX

Posts 19
11 Jun 2019 23:33


A1200 Blizzard 1260 060/66Mhz.

49287 Dhrystones
51.44 MIPS
36.66 MFlops



Gunnar von Boehn
(Apollo Team Member)
Posts 4899
12 Jun 2019 06:56


A1200 coder wrote:

But this can't be right for 68060 results, according to your Sysinfo results it has about the same performance as a 68040 at same clock speed. Or 68040@50 MHz is the same as 68060@50 MHz with integer performance.
 
SysSpeed seems to give 65 MIPS for 68060@50 MHz, but it is also flawed in many ways, seems to overestimate most CPU's performance and scores not comparable to Sysinfo.

We should first clarify that each benchmark is different.
And as each benchmark is different and uses different instructions - it will measure something different.

The 68000 supports MANY different instructions, some are simple, some are very powerful. Some instructions are calculated very quickly, some instructions take longer to finish.

SYSINFO executes 1 test program which executes a mix of instructions.
The work code in SYSINFO is roughly 1000 instructions long.
SYSINFO measures how many times this program can be executed per second and based on this it then calculates the MIPS.

Its obvious the result does depend very much on the "mix" of instructions.

The 68040 CPU is a good CPU and has its own strength.
In some areas the 68040 is even stronger than the 68060.

Lets look at some example


Instruction        LENGTH in BYTE    Clocks 68040  Clocks 68060
ADDQ.L  #1,D0      2                  1              0.5
ADD.W  #11,D0      4                  1              1
ADD.L  #111,D0    6                  1              2

Lets calc the MIPS


Instruction        LENGTH in BYTE    Mips 68040@50  Mips 68060@50
ADDQ.L  #1,D0      2                  50            100
ADD.W  #11,D0      4                  50            50
ADD.L  #111,D0    6                  50            25

As you see the result of the 68060 can be higher than the 68040 but also can be lower in some cases. It depends very much on the tested code.




Thellier Alain

Posts 118
12 Jun 2019 10:24


Hello

>But Sysinfo does not the official "Dhrystone" routine for this.

Perhaps but the only result I want to know is:
How many times the V4 standalone that I will buy will be faster than the last classic I owned an A4040/25 ?

I mean using the same benchmark for A4040/V4 standalone



Ronnie Beck
(Apollo Team Member)
Posts 191
12 Jun 2019 11:48


thellier alain wrote:

  Perhaps but the only result I want to know is:
  How many times the V4 standalone that I will buy will be faster than the last classic I owned an A4040/25 ?

Then take the first post of this thread and add some salt.


Vojin Vidanovic

Posts 1719
12 Jun 2019 12:09


thellier alain wrote:

      Perhaps but the only result I want to know is:
      How many times the V4 standalone that I will buy will be faster than the last classic I owned an A4040/25 ?
   

   
    10x in CPU in 040-FPU code, more once 060/080-MMX code is used (RIVA player, few datatypes,version of Netsurf, few games and emulators). Several times in IDE, many many times in fast RAM / chip RAM and RTG, and some even in combined AGA performance. Nice :) Add having USB, Ethernet, SD Card, RTG-DIGITAL-VIDEO,16 bit sound and plenty of RAM out of box, plus likely warp3D support in future, and there is a real "Amiga 5000" like.


Simo Koivukoski
(Apollo Team Member)
Posts 587
15 Jun 2019 16:13


Life begins at 200 MIPS! (Standalone x16)




Niclas A
(Apollo Team Member)
Posts 218
15 Jun 2019 18:55


Simo Koivukoski wrote:

Life begins at 200 MIPS! (Standalone x16)
 
 

LOL awsome :D
NovaCoder needs a new slogan :P


Chris Edwards

Posts 43
17 Jun 2019 15:22


Nice! great work everyone. now make it say 68080!!
 


Przemyslaw Tkaczyk

Posts 142
25 Jun 2019 09:15


Apollo Core GOLD 2.12 RC beta (x13 / 92MHz) on V600 V2




Peter K

Posts 1
25 Jun 2019 15:27


Apollo Core GOLD 2.12 RC beta (x13 / 92MHz) on V600 V2
 



Gunnar von Boehn
(Apollo Team Member)
Posts 4899
25 Jun 2019 15:34


awesome scores guys!


Gunnar von Boehn
(Apollo Team Member)
Posts 4899
25 Jun 2019 16:27





Andy Hearn

Posts 328
25 Jun 2019 16:49


very very nice! :D


Stefano Briccolani

Posts 501
25 Jun 2019 17:41


Please release that v600 x13 core!!!


Simo Koivukoski
(Apollo Team Member)
Posts 587
08 Jul 2019 21:06


080 was pushed to do a little more work per clock cycle and made an another world record!



posts 49page  1 2 3