Okay, that's nice. Now how about real-life performance where the disk is doing more than just a linear read? This is where you will see that the benchmark numbers for your disks crumble as the tire gets down to the road...
If your event is public, just make sure that you've complied with their licensing. This usually entails one copy of per game per PC if it's being held in a Lab. If it's BYO-PC, then people will have to worry about their own licenses and games.
Now, for what BGSU is known for... If you plan to attend, you best pack a rubber.
Don't dump two tons of rubbish on the projects which are the foundation of your operating system. As far as innovation, user applications are more an exercise in UI design than sheer computer science. Consider the following:
OSX's kernel is a PPC port of the Mach microkernel.
OSX's VM is tailored after the likes of FreeBSD 4.
OSX's networking stack implements *BSD* sockets.
OSX's programming documentation prominently features "FreeBSD" at the bottom of nearly every manpage
JKH, one of the main founders (and long-time innovator) of FreeBSD is working at Apple.
Apple's own "revolutionary" UI has borrowed concepts from NeXT.
Apple has felt the need to make their offerings more appealing by supporting the very window system that you're crapping all over. Yes, I'm talking about X11.
If you knew your history, you would know that Apple's System 6, 7, 8 and 9 did not have protected memory hence any process could overwrite any location in memory, including kernel space. Guess what they adopted to fix that problem...:) As you can imagine, the PC UNIX world has had protected memory since the introduction of the Intel 386 in the mid-eighties.
You knock the very process (and projects) that brought you OSX. If it weren't for Mach, you wouldn't have a kernel for Darwin. If it weren't for FreeBSD, you wouldn't have a lot of OSX. After all, a quick look at the binaries in/sbin on an OSX machine (10.3.2 Build 7D24) reveals the following:
The command run was none other than "strings/sbin/* |grep FreeBSD | sort". Try it on/bin and you'll get 34 more FreeBSD CVS $Id$ strings. Surely FreeBSD doesn't suck so that bad if the almighty OSX incorporates it's code!
A number of people have reported this while using the ULE scheduler and/or switching from UP to SMP.
You want to cvsup your source tree, compile world and your kernel, install kernel, reboot in single user mode, make installworld then run mergemaster.
Re:VM system (Re:BSD is designed. Linux is grown?)
on
BSD For Linux Users
·
· Score: 1
You're forthcoming with words, now let's see the proof to back them. Such a blanket statement is bound to be incorrect. I'm sure there has been a lot of tuning in the Linux kernel to aleviate performance issues that existed in the past, but you cannot say that X is better than Y unless it is so in every category. From my lab tests, Linux has seriously improved it's throughput and beats FreeBSD and NetBSD in some tests, but there are areas where Linux does less than shine (Handling of thousands of fds, for example).
You'll want to go grab a sandwhich at this point while it goes through all of the ports and creates their associated readme files. You'll then be able to use your favorite browser to list the ports and their descriptions. The URL you want file:///usr/ports/README.html .
Jesus Christ, wtf are you smoking? The SiI3112A is the Silicon Image 3112A Serial ATA chip. That said, I will admit that I got a good chuckle out of your post...;-)
AMD K6 300 Mhz overclocked to 333 Mhz so that it can use a 66 Mhz bus? With 256MB of PC100 RAM, this baby is doing production dns on a rather busy network and it's running along just fine!
Well, after all of the costs associated (including staff training), that's still quite a sizeable chunk of cash for a small business (We're just under 40 employees).
Amen! We used to be an all Sun shop and have migrated to FreeBSD on all of servers. The mailing lists have answered all the questions that the manpages couldn't, the hardware is readily available and we've saved over twenty grand worth of misc equipment in the last 5 years. We couldn't possibly be happier with our decision to migrate.
Well, it's been known that Windows disintegrates over time with casual use, even when no updates to the system are made. Having to support a LAN full of Windows machines has always been more of time and resource intensive than a UNIX counterpart (Assuming competant admins); Anyone that's spent 20 minutes in the real world (academia or professionally) could tell you that.
There are other infered costs other than the hardware. Paying a tech to reload Windows (or fix whatever Win32 software package which broke for whatever reason) is what the poster was talking about.
Here's a list of things that you might want to try out:
- Rethink your slicing, see tuning(7) for more on this. - Enable softupdates on the data filesystems of choice. - Compile a custom kernel with only the drivers that you *really* need. - Tune the VFS sysctls. This item cannot be stressed enough. The values selected depend upon the role of the server. - Tune your IP stack. There's a number of sysctls that will quadruple your network throughput. Google has a number of guides which provide good insight.
Remember: Copying a file from a directory to another is by no means a scientific test, as it depends upon the layout of the disk, sector size, disk cache, and a laundry list of other variables. The output of bonnie++ is what you should use to benchmark your IO subsystem.
Slashdot Mirror
Okay, that's nice. Now how about real-life performance where the disk is doing more than just a linear read? This is where you will see that the benchmark numbers for your disks crumble as the tire gets down to the road...
That's pure bollocks. There's a PCI 4-port switch on thinkgeek for $46.99 + S&H.
They've sold these at CompUSA for a long time. I've owned one for almost a year already.
If your event is public, just make sure that you've complied with their licensing. This usually entails one copy of per game per PC if it's being held in a Lab. If it's BYO-PC, then people will have to worry about their own licenses and games.
Now, for what BGSU is known for... If you plan to attend, you best pack a rubber.
Did you know that 129263.376 rods per cubic yards is 2 miles per gallon? :-)
OSX's kernel is a PPC port of the Mach microkernel.
OSX's VM is tailored after the likes of FreeBSD 4.
OSX's networking stack implements *BSD* sockets.
OSX's programming documentation prominently features "FreeBSD" at the bottom of nearly every manpage
JKH, one of the main founders (and long-time innovator) of FreeBSD is working at Apple.
Apple's own "revolutionary" UI has borrowed concepts from NeXT.
Apple has felt the need to make their offerings more appealing by supporting the very window system that you're crapping all over. Yes, I'm talking about X11.
:) As you can imagine, the PC UNIX world has had protected memory since the introduction of the Intel 386 in the mid-eighties.
If you knew your history, you would know that Apple's System 6, 7, 8 and 9 did not have protected memory hence any process could overwrite any location in memory, including kernel space. Guess what they adopted to fix that problem...
Don't get me wrong, I am the proud owner of Dual 2.0Ghz G5 but please, don't throw garbage out there when you have absolutely no idea of what you're talking about.
You knock the very process (and projects) that brought you OSX. If it weren't for Mach, you wouldn't have a kernel for Darwin. If it weren't for FreeBSD, you wouldn't have a lot of OSX. After all, a quick look at the binaries in /sbin on an OSX machine (10.3.2 Build 7D24) reveals the following:
/sbin/* |grep FreeBSD | sort". Try it on /bin and you'll get 34 more FreeBSD CVS $Id$ strings. Surely FreeBSD doesn't suck so that bad if the almighty OSX incorporates it's code!
$FreeBSD: src/sbin/md5/md5.c,v 1.20.2.5 2001/12/26 09:44:56 phk Exp $
$FreeBSD: src/sbin/mount_msdos/mount_msdos.c,v 1.19 2000/01/08 16:47:55 ache Exp $
$FreeBSD: src/sbin/ping6/ping6.c,v 1.4.2.6 2001/07/06 08:56:47 ume Exp $
$FreeBSD: src/sbin/reboot/reboot.c,v 1.17 2002/10/06 16:24:36 thomas Exp $
$FreeBSD: src/sbin/reboot/reboot.c,v 1.17 2002/10/06 16:24:36 thomas Exp $
$FreeBSD: src/sbin/shutdown/shutdown.c,v 1.23 2002/03/21 13:20:48 imp Exp $
The command run was none other than "strings
A number of people have reported this while using the ULE scheduler and/or switching from UP to SMP.
You want to cvsup your source tree, compile world and your kernel, install kernel, reboot in single user mode, make installworld then run mergemaster.
You're forthcoming with words, now let's see the proof to back them. Such a blanket statement is bound to be incorrect. I'm sure there has been a lot of tuning in the Linux kernel to aleviate performance issues that existed in the past, but you cannot say that X is better than Y unless it is so in every category. From my lab tests, Linux has seriously improved it's throughput and beats FreeBSD and NetBSD in some tests, but there are areas where Linux does less than shine (Handling of thousands of fds, for example).
Jesus Christ, wtf are you smoking? The SiI3112A is the Silicon Image 3112A Serial ATA chip. That said, I will admit that I got a good chuckle out of your post... ;-)
Vas te faire foutre. :)
"The nice thing about standards is that there's so many to choose from."
Snail mail requires no such thing. You either use iso9660 or UDF to encode the disc.
AMD K6 300 Mhz overclocked to 333 Mhz so that it can use a 66 Mhz bus? With 256MB of PC100 RAM, this baby is doing production dns on a rather busy network and it's running along just fine!
Break their knees.
It just doesn't make sense for a company like Microsoft, which is part of the BSA to go ahead and steal another company's intellectual property.
How can these people claim recourse against software pirates when they, themselves are pirating other people's ideas, software and concepts?
The stronger the beer, the better the code!
Well, after all of the costs associated (including staff training), that's still quite a sizeable chunk of cash for a small business (We're just under 40 employees).
Amen! We used to be an all Sun shop and have migrated to FreeBSD on all of servers. The mailing lists have answered all the questions that the manpages couldn't, the hardware is readily available and we've saved over twenty grand worth of misc equipment in the last 5 years. We couldn't possibly be happier with our decision to migrate.
Well, it's been known that Windows disintegrates over time with casual use, even when no updates to the system are made. Having to support a LAN full of Windows machines has always been more of time and resource intensive than a UNIX counterpart (Assuming competant admins); Anyone that's spent 20 minutes in the real world (academia or professionally) could tell you that.
There are other infered costs other than the hardware. Paying a tech to reload Windows (or fix whatever Win32 software package which broke for whatever reason) is what the poster was talking about.
Read the tunefs(8) and tuning(7) manpages.
Here's a list of things that you might want to try out:
- Rethink your slicing, see tuning(7) for more on this.
- Enable softupdates on the data filesystems of choice.
- Compile a custom kernel with only the drivers that you *really* need.
- Tune the VFS sysctls. This item cannot be stressed enough. The values selected depend upon the role of the server.
- Tune your IP stack. There's a number of sysctls that will quadruple your network throughput. Google has a number of guides which provide good insight.
Remember: Copying a file from a directory to another is by no means a scientific test, as it depends upon the layout of the disk, sector size, disk cache, and a laundry list of other variables. The output of bonnie++ is what you should use to benchmark your IO subsystem.
It's interesting to note that you need to have shell access to do anything to the FreeBSD systems affected.
It's nice to see NetBSD updating their compiler suite to something newer. Compiling and testing for so many archs and procs is quite an undertaking!