Home   News   Concept   AMIGA-Compatible   Hardware   Forum   Questions+Answers   Pictures   Contact & Team

Welcome to the Natami / Amiga Forum

This forum is for AMIGA fans interested in the new NATAMI platform.
Please read the forum usage manual.



All TopicsNewsQAFeaturesTalkTEAMLogin to post    Create account
Do you have ideas and feature wishes? Post them here and discuss your ideas.

Coldfire 54455 Dev System - First Impressions
Gunnar von Boehn
Germany
(Moderator)
Posts 5775
10 May 2008 12:33


Hi,

As you might know we are evaluating the usage of the Coldfire as powerful but also inexpensive option for a very low price NATAMI verison.
 
I've work some time with Freescales Coldfire 54455 development and here are my impressions of it.

Excellent quality of the dev-system
The case is high quality.
The first glance on the board and layout give the same impression. Very quality work. No cheap parts.

Setting up the system was easy
Setting up the system with cross-compiler was easy and simple.
The system comes with good documentation and cross-compiler for MS-Windows on CD.
I've choosen to download the compiler for Linux.

GCC for Coldfire is working good but has room for improvement
Compiling for Coldfire was easy.
But I did look at the ASM that the GCC was producing, and it was obvious that GCC is a bit challenged to create efficient code for Coldfire.
Some may say this is no news.
C is a high language - you can not expect that the code created by GCC will be small or fast.
This might be true.
It was obvious that the created code was both bigger and slower than needed.

The 54455 Coldfire CPU
The 54455 is a V4e Coldfire in a 16 Bit casing, designe dfor very low cost systems.
The V4 Coldfire is a very efficient CPU.
The V4 is much more efficient than all previous Coldfire CPUs.
Under best conditions it can execute up to 2 integer instructions plus 1 branch per clock.

The 54455 is the smallest of the V4e cores.
Its the "embedded" version of the fast embedded CPUs and has only a 16bit data bus.
Because of this the memory performance is a bit limited of course.
All other V4 cores have a 32bit memory interface and if we want to use a Coldfire for a device that needs to have memory performance than all other V4 will have better memory performance.

Cheers

Gunnar von Boehn
Germany
(Moderator)
Posts 5775
10 May 2008 12:37


This tiny 16bit Coldfire is of course not a memory performance monster.

But for comparison here some facts:

Memory-Benchmark
One of the things I like to start with is evaluating the memory performance as it has such a big impact on the overall performance of many tasks.
The 54455 is of course not a memory wonder (with 16bit)
And I just started with benchmarking so here are some early results:

CF 54455 (16bit bus)
Read: 126 MB/sec
Write: 196 MB/sec
Copy: 121 MB/sec

For comparison G3-600 MHz (64Bit Bus) under MorphOS
Read: 228 MB/sec
Write: 87 MB/sec
Copy: 116 MB/sec

As you can see the 16bit Coldfire V4 perfroma quite OK when  compared to a Pegasus I.

Its worth to mind that the all other Coldfire V4 CPUs are 32bit and their memory bandwidth is accordingly higher.

Cheers

Gunnar von Boehn
Germany
(Moderator)
Posts 5775
10 May 2008 12:49


CPU Benchmark
 
  Or test showed the following results.
  These benchmark results of the V4 100% match the perfromance documentation provided by Freescale.
 
  Clock by Clock the Coldfire V4 is a little bit better than a 68040 CPU
 
  Your mileage will vary of course depending on your individual setup.
 
  Over simplified you can say that the current 266 MHz V4 cores are comparable in performance to a 68040 with 300 Mhz or a 68060 with 160 Mhz.
 
  Freescale does produce as well the Coldfire V5 CPU generation.
  We currently have no samples for testing so our estimates are onlöy based on the documentation that we got.
 
  Clock by clock the V5 Coldfire CPU is comparable to the 68060 CPU.
  From what we were told the currently produced
  400 MHz V5 Coldfire CPU
  are comparable to a 68060 running at 400 Mhz
  and comparable to a 68040 running at 600 Mhz
 
 
  Freesclae told us that the currently produces V5 Coldfire  match the performance of the PowerPC CPUs currently used in the SAM or EFIKA designs.
 
 
  Cheers
  Gunnar

Fabian Nunez
USA

Posts 312
10 May 2008 23:11


Gunnar von Boehn wrote:
CF 54455 (16bit bus)
    Read: 126 MB/sec
    Write: 196 MB/sec
    Copy: 121 MB/sec
   
    For comparison G3-600 MHz (64Bit Bus) under MorphOS
    Read: 228 MB/sec
    Write: 87 MB/sec
    Copy: 116 MB/sec

   
  Those numbers look a bit odd to me... why are reads almost 2x as fast on the G3 but writes half the speed?  That 4x difference is very strange, what's the write caching configuration on the Peg?
   
 
Gunnar von Boehn wrote:
Freesclae told us that the currently produces V5 Coldfire  match the performance of the PowerPC CPUs currently used in the SAM or EFIKA designs.

  That's great, the performance of PPC with the familiarity of the 68K family.  Too bad there's no 68K equivalent to Altivec, that would have really rocked.  What's the price difference between comparable V4 and V5 chips?
 

Gunnar von Boehn
Germany
(Moderator)
Posts 5775
12 May 2008 14:40


Fabian Nunez wrote:

Gunnar von Boehn wrote:
CF 54455 (16bit bus)
    Read: 126 MB/sec
    Write: 196 MB/sec
    Copy: 121 MB/sec
   
    For comparison G3-600 MHz (64Bit Bus) under MorphOS
    Read: 228 MB/sec
    Write: 87 MB/sec
    Copy: 116 MB/sec

   
  Those numbers look a bit odd to me... why are reads almost 2x as fast on the G3 but writes half the speed?  That 4x difference is very strange, what's the write caching configuration on the Peg?
   

The result of the Pegasos are typical for the Pegasos 1.
The read and write performance depends on the combination of CPU, the memory protocla of the CPU, the used Northbridge or the used memory.

For having 16bit bus the mini Coldfire performs quite nicely.
I was not able to benchmark the 32bit Coldfire yet but I assume that the performance of the 32bit V4 Coldfire is double of the 16bit one.

I think the 32bit Coldfire should archive around.
read 32          250 MB/s
write 32        500 MB/s

For comparison here are benchmark results of some PowerPC Neo-AMIGAs.

Pegasos 1 G3 @ 600MHz 64bit-DDR memorybus 200
read 32          210 MB/s
write 32        88 MB/s

Pegasos 2 G4 @ 1000MHz 64bit-DDR memorybus 266
read 32          243 MB/s
write 32        496 MB/s

AMIGA-ONE G3 @ 800 MB 64bit-DDR memorybus 266
read 32          200 MB/s
write 32        122 MB/s

AMIGA-ONE G4 @ 933 MHz 64bit-DDR memorybus 266
read 32          205 MB
write 32        268 MB

As you can see different design have different performance.
The AMIGA-ONEe and the Pegaos 1 both use the same Northbridge.
Its often said that the memory interface of the AMIGA-ONE and PEGASOS 1 is somewhat of a limitation of the system.
 

Fabian Nunez wrote:

That's great, the performance of PPC with the familiarity of the 68K family.  Too bad there's no 68K equivalent to Altivec, that would have really rocked.

The Coldfire have the EMAC unit.
EMAC is said to be highly effective for mass multiplication routines needed for various audio and video codecs.

Fabian Nunez wrote:

What's the price difference between comparable V4 and V5 chips?

The goal is to have a V5 with integrated SuperAGA in one chip.
I would assume that this AMIGA in a chip will come to around $20 total.

Cheers
Gunnar

Fabian Nunez
USA

Posts 312
13 May 2008 09:27


Gunnar von Boehn wrote:

  The goal is to have a V5 with integrated SuperAGA in one chip.
  I would assume that this AMIGA in a chip will come to around $20 total.

I remember Bill Buck saying a short while ago that he was going to work with you on this... any news?


Joe M
Norway

Posts 500
27 May 2008 04:52


Fabian Nunez wrote:
   
> I remember Bill Buck saying a short while ago that he was going
> to work with you on this... any news?
   
Wouldn't it be a bad idea to drag Bill Buck into this? As far as I've understood he's not very much interested in the classic Amiga. Everything he seems to want is to use the Amiga name for his own projects. He's already been talking about Amiga OS 5 which is being developed in India. Since it's being developed there, it will probably turn out to be something completely different than Amiga OS 3 and 4.

Fabian Nunez
USA

Posts 312
27 May 2008 05:20


I thought the AmigaOS 5 from India thing was done by Amiga Inc?  Either way, now that Thomas and Gunnar have decided to put the CPU inside the FPGA instead of using a discrete Coldfire I think that Bill Buck will lose interest in the project.


Joe M
Norway

Posts 500
27 May 2008 05:29


Fabian Nunez wrote:

I thought the AmigaOS 5 from India thing was done by Amiga Inc?  Either way, now that Thomas and Gunnar have decided to put the CPU inside the FPGA instead of using a discrete Coldfire I think that Bill Buck will lose interest in the project.

He he, I simply thought you meant Bill in Amiga Inc. ;-) Cool. Thomas and Gunnar's new decision sounds great.

posts 9