Slashdot Mirror


FreeBSD Looking for People with Lots of RAM

drdink writes "A few weeks ago, PAE (Physical Address Extension) support was added to FreeBSD 5-CURRENT. This allows memory above 4GB to be used normally by the kernel and userland on the x86 platform. Jake Burkholder, the man behind PAE, is now looking for users to help him test this new feature. In his message to the freebsd-current mailing list, Jake describes the current caveats to PAE and also says 'We'd like this feature to be solid for 5.1-RELEASE, so I'm hoping there are people out there with systems with more than 4G of ram that are willing to test it.' This, along with other features make FreeBSD 5-STABLE look very promising."

23 of 271 comments (clear)

  1. ARgghh by cholo54alpha · · Score: 3, Interesting

    MY dilemma is I have a lot of Ram but half of it is flakey!!! so I just tell linux to skip over it. It's really like having a regular amount of good ram. Hey, can BSD map my bad ram out too? Anyone?

  2. Volunteer... by addaon · · Score: 4, Interesting

    I have an HP LXR 8500 with four processors (currently) and 4GB of ram. I've been considering upgrading to 6GB for a while anyway. I'm currently using Windows 2000 advanced server on it, after being somewhat frustrated with Linux support a couple of years ago. I'd be more than willing to try out BSD, although I never have before. Is there anything I should know about this? I presume that BSD would run Mathematica fine under Linux emulation mode, as my main use of the box is just Mathematica crunching. Does FreeBSD make reasonable use of four processors? Anything else I should beware of? And anyone know a good source for cheap lxr-ready ram?

    --

    I've had this sig for three days.
    1. Re:Volunteer... by UniverseIsADoughnut · · Score: 5, Informative

      Freebsd does SMP quit well. They have/are redoing the whole SMP system. It was slated for 5.0 but i don't know if it did or did not make it in. When finished Freebsd will have and extremely good SMP, if not the best.

      So what are you crunching with that thing?

    2. Re:Volunteer... by Brooks+Davis · · Score: 5, Interesting

      Mathematica runs find under Linux emulation. We're currently working on getting Grid Mathematica into production on our FreeBSD cluster at work. It runs, but we haven't really done much with it yet because the real users are still working on understanding the programming model. FreeBSD should work well for you in this mode as long as you have the four licenses you'll need to keep the CPUs busy. The way Mathematica handles parallelism (seperate processes) should be able to take advantage of PAE.

      The one gotcha is that PAE is a bit bleeding edge at this point so moving to it may be intresting.

      -- Brooks

      --
      -- Any statement of the form "X is the one, true Y" is FALSE.
    3. Re:Volunteer... by drdink · · Score: 4, Informative
      One thing you need to remember is that FreeBSD 5.x is currently still not -STABLE. This means it is the current development line. There is no guarentee or illusion of stability. That is why such big features as PAE, SMPng, x86-64, etc are being done there. If you are seriously interested in running a development version of FreeBSD, be ready to play a role in debugging, testing, and possibly watching things explode. That said, it has been pretty stable for me lately. If you are still interested, then please do the following:

      For those who are curious about what is new in -CURRENT compared to 4-STABLE, you can read the 5.0-RELEASE release notes for the bits that were new at the time of 5.0-RELEASE. More has come since.

      --
      Beware, Nugget is watching... See?
    4. Re:Volunteer... by drdink · · Score: 5, Informative
      FreeBSD/i386 5.0-RELEASE Release Notes: Processor/Motherboard Support:
      SMP support has been largely reworked, incorporating code from BSD/OS 5.0. One of the main features of SMPng (``SMP Next Generation'') is to allow more processes to run in kernel, without the need for spin locks that can dramatically reduce the efficiency of multiple processors. Interrupt handlers now have contexts associated with them that allow them to be blocked, which reduces the need to lock out interrupts.
      Yes. This is in 5.0 now.
      --
      Beware, Nugget is watching... See?
    5. Re:Volunteer... by Billly+Gates · · Score: 4, Interesting

      Yes but its unbenchmarked

      FreeBSD 5 also includes Linux like threading. This has been one of the traditional weaknesses that are being addressed. Great java support as well. Since java uses threading heavily FreeBSD 5 will make it have server/workstation performance. Yahoo wanted to go with Java for their next generation portal software but Freebsd 4.x series had mediocre thread support.

      Freebsd 5.0 rocks! The only downside is that my Microsoft USB keyboard does not work with FreeBSD 5 on certain motherboards. I think its a bug and I hope its fixed soon.

    6. Re:Volunteer... by Billly+Gates · · Score: 3, Insightful

      Bologna. Its not beta and its considered stable.

      Current != Release. I looked at there ftp site and only found -current or -Release versions. The only one mentioned as stable I found reading the docs are 4.0.

      Current = beta, and Release = stable. Stable= superstable or enterprise class stability.

      FreeBSD 5 is ready for %95 of user and server use. Its just as stable as FreeBSD 4.6 or 4.7 since they are also Release versions. Only 4.0 is considered STABLE at this point.

      However I would not bet my job on it with a server that needs to stay up 24x7 but FreeBSD 5 is as stable if not more out of the box as Redhat8 or Mandrake. FreeBSD hackers obsess about stability more then most linux hackers with the exception of Debian users. I would be cautious of course but to be release quality it needs to be %99.9 stable as opposed to %99.999 stable as 4.0 stable.

    7. Re:Volunteer... by addaon · · Score: 4, Interesting

      Mathematica is slow, inefficient, and a total joy to work with. Matlab is about 10x faster, and at least 100x faster for matrix stuff (mathematica does matrices as linked lists of linked lists!)... but if I want speed, I'll use Fortran 90, really. Mathematica is just lisp + pretty printing... but it undeniably works, and I've never used a more productive tool in my life.

      My basic method is to take a 3D solid (that is, a polygonal 3D model that encloses a volume). I apply a few standard smoothing operations to this model to get a very, very rough shape (in particular, I smooth it until it is fully convex, the first time). I then do simulated annealing based on a handful of hardcoded starting conditions to find a good approximation of that shape -- metropolis might work better for precision, but I decidedly don't want precision now. I then take the model and resmooth it, but one step less. I use the previous foldset as a starting point, and anneal from there. I repeat this for each level of smoothing that was originally needed.

      Normally it takes around a million attempts to approximate each smoothing level, although this varies by a factor of at least one hundred, where the swan, for instance, takes about fifty levels of smoothing.

      Make sense? Not saying it works wonderfully, but I think it's the correct approach and just needs tuning. Amount of work is, to a first approximation, linear with the geometric complexity of the model, and more or less independent of the number of folds... certainly not exponential in the number of folds!

      --

      I've had this sig for three days.
  3. Uh... by JanusFury · · Score: 3, Interesting

    How would you test that? I can't think of any easy way to actually test that much RAM. What would you do, load 8GB of random data into RAM and compare it byte-by-byte with the original data?

    --
    using namespace slashdot;
    troll::post();
    1. Re:Uh... by Anonymous Coward · · Score: 5, Funny

      Just load up a Java application... Voila!

  4. Re:64 + 128 + 256 + 16 + 224 + 64 + 512k.... by AgentPhunk · · Score: 5, Funny
    I've got 8 gigs of RAM in my old 286 running a hacked version of FreeDOS, so this isn't really news

    Umm, that'd be 8 MEGs you got in there, Sparky.

    I'll bet it has a 5.25 TB floppy drive, and a 20" LCD green-screen monitor.

  5. This just in... by Znonymous+Coward · · Score: 4, Funny

    From the office of Iraqi Information Minister Mohammed Saeed al-Sahhaf (aka Baghdad Bob):

    "BSD isn't dead! The infidel Linux coaliation will soon pay the price for descriating BSD!"

    More at 11.

    --

    Karma: The shiznight, mostly because I am the Drizzle.

  6. Re:Wow! by Anonymous+Struct · · Score: 5, Funny

    Jesus... I could even put the swap space ON THE RAMDISK! Think about how fast that'd be!

  7. Re:Sweet! by Guy+Harris · · Score: 4, Informative
    is support for more than 4GB of memory a first for 32-bit x86 operating systems?
    No. The Linux 2.4.x kernel has it, and Unixware 7.1.3 has it (I don't know what release first supported it), and Solaris 7 and later has it as well.

    As others have noted, Windows NT 5.0^H^H^H^H^H^H2000 also supports it.

  8. Re:Sweet! by lpontiac · · Score: 4, Funny
    is support for more than 4GB of memory a first for 32-bit x86 operating systems?

    No. The Linux 2.4.x kernel has it [com.com], and Unixware 7.1.3 has it [sco.com]

    So the Linux kernel's support was obviously stolen from SCO, and therefore doesn't count.

  9. Jesus Tap-dancing Christ! by Anonymous Coward · · Score: 5, Funny

    Talk about a fucking bloated OS!? 4 gigs of RAM? Not even XP Pro requires that much memory!

  10. Re:Wow! by Anonymous+Struct · · Score: 5, Funny

    I guess I need to work on my delivery... :)

  11. Re:I have just the box for this... by drdink · · Score: 4, Informative

    FreeBSD supports Hyperthreading in 5.0-CURRENT. There is a sysctl variable called "machdep.hlt_cpus". You can use this variable to control which logical CPUs should be taken out of the idle loop and used by the kernel. This, of course, requires a kernel built with the APIC_IO and SMP kernel options. Lacking a SMP system, I haven't tested this. This is just what I see on the mailing lists and in CVSWeb

    --
    Beware, Nugget is watching... See?
  12. What happens if you have 6 Gigabytes? by Anonymous Coward · · Score: 3, Funny

    Looking at a full rack of DELL PowerEdge 2650s Dual Proc 2.8Ghz w/ 6 gigs of Ram and smiling. But they are already running Windows 2003 Enterprise Server. I wonder if the boss will let me take one of these $10,000 babies offline...YEA I WISH!

  13. Re:My current setup (12GB RAM) by FyRE666 · · Score: 3, Funny

    I presume you're one of the Doom 3 beta testers then?

  14. Re:Why... by smash · · Score: 3, Insightful
    4GB is nothing these days... I've got mates who have a couple of gig in their desktop boxes.

    Hardware support? Never had an issue with it under FreeBSD myself, and if you're planning on running it, you can always pick your hardware properly.

    Now as to WHY you'd run it?

    Its reliable, quick, sensibly laid out, and works very much like commercial unix.

    Just because you're too shortsighted to see a use for it, doesn't mean that no one else has uses for it.

    smash.

    --
    I run: Windows, OS X, Linux, FreeBSD. Just because you have a hammer, doesn't mean everything is a nail.
  15. Re:Rembember the LIM standard? by julesh · · Score: 3, Informative

    Years ago, there was the "LIM" (Lotus-Intel-Microsoft) for adding more than 640KB of RAM to a PC, by "windowing in" a section of RAM in a certain area.

    It seems that, 20 years later, we're back to doing essentially the same thing.


    No, this system doesn't work like that.

    Intel processors since the Pentium have supported a system that allows you to use a larger page size than standard so that you can have more physical address space. You specify the start address of each page as 24 bits which are assumed to align to a 4K boundary which gives you 4M*4K = 16Gb of physical RAM. Each page is 2Mb in length. You can mix 4K and 2Mb pages in the same system, although not in the same quarter of the process adress space. So you get more actual physical memory, although each process is limited to 4G at once (whereas with LIM EMS the entire system was limited to 640K + 64K of 'banked' memory)