Thread: ATOM prcessors & Programming...

  1. #1
    Registered User
    Join Date
    Aug 2008
    Posts
    33

    Lightbulb ATOM prcessors & Programming...

    Hey,

    Ive been looking at those ASUS netbooks for a couple of days now and cant decide on it.
    Basically, I plan to take it to work on my bike and spend lunch programming...

    But I wasnt sure if the ATOM processor, that these netbooks use, would be efficent for GNU?

    Your thoughts, if you please...

  2. #2
    Registered User
    Join Date
    Aug 2008
    Posts
    33
    Im guessing two things:

    a) Depending on the size/speed of the developed application, ATOM could prove slow...

    and

    b) Nobody (so far) has tried writting decent applications on something so small...

  3. #3
    Kernel hacker
    Join Date
    Jul 2007
    Location
    Farncombe, Surrey, England
    Posts
    15,677
    Yes, for all intents and purposes, it's like a slower model of a Pentium4 or Core2 - but it should work find for compiling and such - most of the "slow" stuff is the disk IO anyways.

    --
    Mats
    Compilers can produce warnings - make the compiler programmers happy: Use them!
    Please don't PM me for help - and no, I don't do help over instant messengers.

  4. #4
    Kernel hacker
    Join Date
    Jul 2007
    Location
    Farncombe, Surrey, England
    Posts
    15,677
    Quote Originally Posted by Else View Post
    Im guessing two things:

    a) Depending on the size/speed of the developed application, ATOM could prove slow...

    and

    b) Nobody (so far) has tried writting decent applications on something so small...
    It should be a lot faster than what you find in your average mobile phone, and they still do 20 fps of decent size graphics in software (and 50 fps with hardware).

    --
    Mats
    Compilers can produce warnings - make the compiler programmers happy: Use them!
    Please don't PM me for help - and no, I don't do help over instant messengers.

  5. #5
    Registered User
    Join Date
    Aug 2008
    Posts
    33
    Thats to be expected though...

    Cheers for the input...

    I dont mind the wait for compilation, within reason, as its a small netbook.
    SSD card storage should mean its speedier than HDD though right...?

    ...Running real-time applications shouldnt be a problem though eh?

  6. #6
    Registered User
    Join Date
    Aug 2008
    Posts
    33
    ...Alright, so its a naff idea...

  7. #7
    Kernel hacker
    Join Date
    Jul 2007
    Location
    Farncombe, Surrey, England
    Posts
    15,677
    Depends on what "real-time apps" it is. As I said, it should be a lot faster (like 2-8 times faster) than what you find in for example a Nokia N95 or a Apple iPhone.

    But it will also be about 3-6 times slower than a modern laptop machine, as it's running at a lower frequency and is a single core processor.

    --
    Mats
    Compilers can produce warnings - make the compiler programmers happy: Use them!
    Please don't PM me for help - and no, I don't do help over instant messengers.

  8. #8
    Registered User khdani's Avatar
    Join Date
    Oct 2007
    Posts
    42
    i've programmed on ATOM proccessor on ASP.NET and Java, and it was absolutely fine.

  9. #9
    Kernel hacker
    Join Date
    Jul 2007
    Location
    Farncombe, Surrey, England
    Posts
    15,677
    Yes, I would expect that just about anything you do on a PC today is doable on an Atom-based machine.

    An interesting experiment would be to do soemthing like this:
    Code:
    #include <stdio.h>
    
    int main()
    {
       int n, i, s = 0;
       
       printf("enter number of loops (1000s):");
       fflush(stdout);
       scanf("%d", &n);
       n *= 1000;
       for(i = 0; i < n; i++)
          s += i;
       printf("Done: %d\n", s);
       return 0;
    }
    How many loops do you need before the "done" doesn't come out "immediately"?

    On my machine, I need to enter 1000000 to make it noticable. That is 1000 million iterations of the loop. That's some 4 billion instructions at least - probably more like 8 billion. Ok, so my machine should be about 2-4x faster than the Atom, but I doubt that any simpler applications will actually run slow enough that it's noticable other than by measuring the time it takes.

    --
    Mats
    Compilers can produce warnings - make the compiler programmers happy: Use them!
    Please don't PM me for help - and no, I don't do help over instant messengers.

  10. #10
    Cat without Hat CornedBee's Avatar
    Join Date
    Apr 2003
    Posts
    8,895
    If you're compiling, I'd worry about RAM, not CPU speed.
    All the buzzt!
    CornedBee

    "There is not now, nor has there ever been, nor will there ever be, any programming language in which it is the least bit difficult to write bad code."
    - Flon's Law

  11. #11
    Woof, woof! zacs7's Avatar
    Join Date
    Mar 2007
    Location
    Australia
    Posts
    3,459
    I've got an Intel Atom N270 and it's more than ample for my needs.

    Granted I don't run X unless I need it, but it runs pretty fast and handles the MySQL daemon and Java running together just fine.

    To give you an idea of how fast (or slow) the Atom is, here is a list of games the Msi Wind (Intel Atom N270 @ 1.6GHz) can run http://wiki.msiwind.net/index.php/Windows_Games . Some videos are supplied too.

    > Nobody (so far) has tried writting decent applications on something so small...
    I am / have...

    > If you're compiling, I'd worry about RAM, not CPU speed.
    True, that's why I upgraded my netbook to 2GB as soon as I got it.

    Running matsp's experiment on my N270:
    Code:
    [zac@breeze ~]$ uname -a
    Linux breeze 2.6.27-ARCH #1 SMP PREEMPT Mon Dec 8 22:01:01 UTC 2008 i686 Intel(R) Atom(TM) CPU N270 @ 1.60GHz GenuineIntel GNU/Linux
    
    [zac@breeze ~]$ cat atombench.c
    #include <stdio.h>
    
    int main()
    {
       int n, i, s = 0;
       
       printf("enter number of loops (1000s):");
       fflush(stdout);
       scanf("%d", &n);
       n *= 1000;
       for(i = 0; i < n; i++)
          s += i;
       printf("Done: %d\n", s);
       return 0;
    }
    
    [zac@breeze ~]$ time gcc -W -Wall -pedantic -std=c89 -O0 atombench.c -o atombenchO0
    
    real	0m0.113s
    user	0m0.090s
    sys	0m0.020s
    [zac@breeze ~]$ time gcc -W -Wall -pedantic -std=c89 -O3 atombench.c -o atombenchO3
    
    real	0m0.140s
    user	0m0.113s
    sys	0m0.030s
    
    [zac@breeze ~]$ time ./atombenchO0 < `echo 10000000`
    enter number of loops (1000s):Done: -2021519872
    
    real	0m4.453s
    user	0m4.453s
    sys	0m0.003s
    [zac@breeze ~]$ time ./atombenchO3 < `echo 10000000`
    enter number of loops (1000s):Done: -2021519872
    
    real	0m2.673s
    user	0m2.670s
    sys	0m0.003s
    Compared with my main PC:
    Code:
    [zac@neux ~]$ uname -a
    Linux neux 2.6.26-ARCH #1 SMP PREEMPT Tue Aug 26 20:53:58 UTC 2008 x86_64 AMD Athlon(tm) 64 X2 Dual Core Processor 3800+ AuthenticAMD GNU/Linux
    [zac@neux ~]$ time gcc -W -Wall -pedantic -std=c89 -O0 atombench.c -o atombenchO0
    
    real    0m0.062s
    user    0m0.037s
    sys     0m0.027s
    [zac@neux ~]$ time gcc -W -Wall -pedantic -std=c89 -O3 atombench.c -o atombenchO3
    
    real    0m0.076s
    user    0m0.067s
    sys     0m0.013s
    [zac@neux ~]$ time ./atombenchO0 < `echo 10000000`
    enter number of loops (1000s):Done: -2021519872
    
    real    0m7.047s
    user    0m7.036s
    sys     0m0.000s
    [zac@neux ~]$ time ./atombenchO3 < `echo 10000000`
    enter number of loops (1000s):Done: -2021519872
    
    real	0m0.540s
    user	0m0.540s
    sys	0m0.000s
    Sorry for the spam! It's all related to the OP.

    Notice it runs slower non-optimized on my 3800+ for some reason (both tests were not cached)!?! Both were compiled with gcc version 4.3.1.
    Last edited by zacs7; 12-22-2008 at 06:17 PM. Reason: Removed /proc/cpuinfo (not required)

  12. #12
    Kernel hacker
    Join Date
    Jul 2007
    Location
    Farncombe, Surrey, England
    Posts
    15,677
    Quote Originally Posted by Else View Post
    b) Nobody (so far) has tried writting decent applications on something so small...
    if you mean "decent applications" such as MS Office type applications, I don't see why that wouldn't work just straight out of the box. Likewise for Visual Studio 2008.

    Biggest drawback of the Netbook here would be screen resolution, particularly for VS 2008.

    So strictly speaking, there is no need to write application specifically for this type of processor.

    --
    Mats
    Compilers can produce warnings - make the compiler programmers happy: Use them!
    Please don't PM me for help - and no, I don't do help over instant messengers.

  13. #13
    Woof, woof! zacs7's Avatar
    Join Date
    Mar 2007
    Location
    Australia
    Posts
    3,459
    I can vouch for the screen size, I have a 10" screen. Which I guess you could say it almost a laptop and I find some things annoying to use. Mainly applications that pop-up dialog boxes that aren't resizeable that go off the top and bottom of the screen.

    It's generally fine for web-browsing, although some websites are annoying.

Popular pages Recent additions subscribe to a feed

Similar Threads

  1. vector contents mysteriously vanishing
    By Loduwijk in forum C++ Programming
    Replies: 5
    Last Post: 03-24-2006, 11:09 AM
  2. Odd Problem with Linking
    By jamez05 in forum C++ Programming
    Replies: 6
    Last Post: 09-21-2005, 01:49 PM
  3. Need assistance with program
    By DJ_Mittens in forum C Programming
    Replies: 4
    Last Post: 04-19-2005, 08:16 PM
  4. ""every atom is everywhere..."
    By axon in forum A Brief History of Cprogramming.com
    Replies: 18
    Last Post: 03-15-2004, 09:27 AM