Overview Features Coding Performance Forum Downloads Products OrderV4 Contact

Welcome to the Apollo Forum

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



All TopicsNewsPerformanceGamesDemosApolloVampireAROSWorkbenchATARIReleases
The team will post updates and news here

Upcoming Release8 Will Allow More MIPSpage  1 2 

Gunnar von Boehn
(Apollo Team Member)
Posts 5581
24 Nov 2021 23:56


how about we make a new thread for video topic?


Dawid 1983

Posts 26
30 Nov 2021 23:17


Gunnar von Boehn wrote:

  The upcoming RELEASE 8 for V4 Standalone and Firebird does allow optional switching to Turbo Mode.
  (This mode will run on number card but not all)
   
 
 

 
  Nice MIPS is like 603e at 166Mhz.I think if chip have good cooling will go much faster then this :) Good radiator and fan can make big differences :) Release 8 software or cards ?


Gunnar von Boehn
(Apollo Team Member)
Posts 5581
01 Dec 2021 07:20


Dawid 1983 wrote:

Nice MIPS is like 603e at 166Mhz.

   
Cross architecture MIPS comparison is often relative difficult or the result not measure the same and are useless.
The code executed by SYSINFO is unique to SYSINFO alone and the SYSINFO code is only available on 68K, there is no other benchmark which does the same code for PowerPC - so you can not compare it.
 
The Code in SYSSPEED for example is totally different to SYSINFO.
It sounds unbelieveable but even the Code in SYSSPEED for 68k and for PowerPC is not the same code! It is not doing the same thing on both 68k and on PowerPC - so even the MIPS comparison in SYSSPEED is actually very much useless.

 
If you just count how many instructions can the CPU do per second how do you want to factor in that for doing the work of a SINGLE 68K

Instruction for example:
  ADD.L #$123456,$876543 that you need on the PowerPC 6 instructions to do the same?

The problem with comparing just instructions between different architectures is that some architectures can do more things than others per instruction. For example the 68K can do memory operations  in each instructions, the PowerPC can not do this and needs always more instructions to do the same effective amount of work.
   
So if architecture 68K can do in 100 instructions "100 work units" but architecture PowerPC needs for example 300 instructions to do the same amount of work - then counting how many instructions both did does not measure the amount of work they can do.

This is like comparing money only counting how many coins you have but not looking at their value.
One having 5 coins and the other 4 coins.
But the 5 coins can be pennies while the other has 4 Pound coins.
So which one is more valuable in reality? 


Dawid 1983

Posts 26
01 Dec 2021 09:13


Nice explained thank you


posts 24page  1 2