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!

A1200 IceDrake Bad Network Performancepage  1 2 

A. Dewart
(Needs Verification)
Posts 12/ 1
09 Aug 2022 18:36


@ferell
  yes and industry now decided to count in mebibyte.
  and will you follow that too ?
  when industry start to do crazy things it's not necessary to follow it. common sense is better. also according to industry pc was always the better choice (even in the 90s when amiga was actually superior). industry have a part of fashion and hype and bias according to where money come from.
  my point is, my file size are in bytes, so my network traffic is in bytes/s. when i download something in ibrowse, the indicated speed is in mb/s (megabyte per seconds). i stick to that.
  every network software i know of do count speed in "b/s". (byte per sec .. byte, kilobyte, megabyte, gigabyte. it never count in bit). so i stick to that. whatever the current fashion says. differenciating between megabit and megabyte just by using capitalisation on letters is not precise enough and lead to confusion. just look at the last robo post, even he couldn't keep the proper capital letter standard he advocates like you do.
  i use mb/s as megabyte/s. all the time. and when i want to talk about bit/s i write it fully for what it is: mbit/s. there is no error possible that way.
  noone is forcing you to adopt my way.
  noone is forcing you to read my post.
  and if you choose to avoid reading my post, you do not really need to tell me about it.
 


Steve Ferrell

Posts 424
09 Aug 2022 20:54


a. dewart wrote:

@ferell
    yes and industry now decided to count in mebibyte.
    and will you follow that too ?
    when industry start to do crazy things it's not necessary to follow it. common sense is better.
   

There's this thing called "industry standards" and they're actually quite nice to have because it gives everyone a common language and standardized metrics to make comparisons.  So, yes, I'll follow  the industry standards because it makes my job easier and it enables others to understand me, especially when I ask for assistance or when they ask for mine.

If you want to go off on your own and make your own standards and metrics, be my guest, but no one is going to follow you or assume that you have common sense.  Not following established rules and standards actually shows that you lack common sense.  Do you disregard traffic laws and standards and make up your own?  No, that's just foolish, as is making up your own IT standards and metrics.

And you're right, no one is forcing me to adopt your standards or metrics or to even read your posts.  I, along with several others here, will also ignore your posts going forward.  Good luck!  You're going to need it with your rude and dismissive attitude.



Steve Hodson

Posts 32
10 Aug 2022 05:50


I have a V4SA and an FTP server on my LAN connected to a new Mac capable of far exceeding the Amiga bandwidth via a 10.100.1000 switch.

Practically, with newer version of the v4 ethernet device and firmware, I can achieve around 3.5 Mega Bytes peer second in large transfers.

For comparison, under the same conditions an 060 Amiga is getting 1-1.5 Mea Bytes per second.

Both are with the current version of ROADSHOW.


Gunnar von Boehn
(Apollo Team Member)
Posts 6207
10 Aug 2022 07:01


a. dewart wrote:

i use mb/s as megabyte/s. all the time. and when
noone is forcing you to adopt my way.
noone is forcing you to read my post.

   
If you would use SHIFT and write it correctly as "MB/s" then people can better understand you.

Why are you so rude to the people answering your questions?

You seem on purpose to misquote others posts. What benefit do you have from this?

You made your forum account only 1 day ago.
I hope you made your account not only for trolling?
 


A. Dewart
(Needs Verification)
Posts 12/ 1
10 Aug 2022 11:10


@hodson

what you report is better than what i'm getting on my v4sa. but it's also slower than what biggun stated (5mb/s on local). and it's also a lot slower than what i see on fsuae (configured as 160mips 68k).
i hope improvement will be made on that front in future updates. in the meantime maybe gunnar will be able to give details of what is planned for improvement in the future, compared to current situation. that would be interesting to read, to understand better what's going on.
i have similar questions on the usb ports too but i'll ask about that in another topic since this one here is about networking performance.



Gunnar von Boehn
(Apollo Team Member)
Posts 6207
10 Aug 2022 14:51


a. dewart wrote:

@hodson
 
  what you report is better than what i'm getting on my v4sa. but it's also slower than what biggun stated (5mb/s on local). and it's also a lot slower than what i see on fsuae (configured as 160mips 68k).

Please stop posting nonsense.
The numbers you posted are even hard to reach for a GHz Pentium.

I understand that you do not know how UAE does ethernet an how much of this is done by the PC. No one can expect that you know this. But please stop making nonsense claims.




Gunnar von Boehn
(Apollo Team Member)
Posts 6207
10 Aug 2022 16:51


I think we better stop this useless discussion now.
This smells to much like trolling now.
 
We had here a user with a real configuration problem.
But his question got hijacked. :(
 
Some users tell you that they get 3.5 Mbyte/sec
I tell you that we got 5 MB/sec in local tests and I tell you that I get 1 MB/sec from the internet. And that the Ethernet stack on Aros is not holding any world records.

Yes, of course the Amitcp/Ethernet stack can be improved.

Lets spend time on coding to improve Amiga software!

 
 


A. Dewart
(Needs Verification)
Posts 12/ 1
10 Aug 2022 17:05


@gunnar
ok, but why? could you explain ?
are you saying the cpu is wasting a lot of time and resource on tcp/ip processing ?
i'm just trying to understand.
also the other day i ran same test on the v4sa, and cpu usage was less than 50% and cpu monitoring app from flype showed 31mips usage while downloading that file (at 1.6mb/s).

how can it be true in same time to say that what makes the throughput so slow is that cpu is wasting resources on tcp/ip stack processing and having cpu at 50% in same time ..

also would the test make more sense if using a protocol that make to reception verification / error control, like udp maybe ?



Allan Versaevel

Posts 111
21 Aug 2022 18:22


With the Vampire V2 Ethernet Port, I usually get an average of 650 KBps (Kilo-Bytes per second) or 0.65 MBps (Mega-Bytes per second), which translates to the infamous industry marketing standard of 5.2 mbps (Mega-bits per second). For my needs, this speed is more than enough for my Amiga tasks including Browsing and File Transfers. Anyone ever use ParNet back in the day? Enough said :)

Allan V.


Brett Eden

Posts 17
09 Feb 2024 12:16


Gunnar von Boehn wrote:

    I think we better stop this useless discussion now.
    This smells to much like trolling now.
     
    We had here a user with a real configuration problem.
    But his question got hijacked. :(
     
    Some users tell you that they get 3.5 Mbyte/sec
    I tell you that we got 5 MB/sec in local tests and I tell you that I get 1 MB/sec from the internet. And that the Ethernet stack on Aros is not holding any world records.
   
    Yes, of course the Amitcp/Ethernet stack can be improved.
   
    Lets spend time on coding to improve Amiga software!
   
   

   
    Hi Gunnar,
   
    I know this is quite an old post, but it is a topic some people may still wonder about.  There is a video on YouTube (I am very sorry I did not bookmark it :-(...) where a user compares the Internet speed of the V4SA, Firebird, Manticore and IceDrake.
   
    The available Internet bandwidth was over 10 Megabytes/sec incoming.
   
    The V4SA scored the highest, downloading over 4.5 Megabytes/sec.  The Firebird came 2nd, at a little over 4 Megabytes/sec, and the Manticore and Icedrake were close behind.
   
    Around 4 Megabytes/sec is what appears to be the speed that people should be able to attain.  This is much faster than what people are able to get from 802.11b PCMCIA card...

posts 30page  1 2