jeeesus, I was wondering how many fucking thousands of comments I would have to read through before someone mentioned usenet. As I was reading (hehee), I got to thinking: to what extent does slashdot want to retain control over the distribution of their (user-contributed!) content, like some other organizations which control distribution channels of artist-contributed content?
Java, in the form of EJBs and servlets, is becoming fairly common on "big iron". One reason is indeed the cross-platformability of it. But, frankly, most often a change in platform accompanies a change in framework. For example, a switch from UNIX servers to a mainframe may accompany a switch from CGIs to an application server (such as Websphere).
The dramatic switch here is probably not from UNIX to OS/390 (after all, OS/390 has a POSIX layer), but from CGIs to Websphere.
So, as you said, it's not really cross-platformability which is driving Java in the server realm.
I think the principal reason is that Java provides (at least some semblance of:) a language-level security model. By security, I don't mean crypto and virii and external attacks. I mean an application is secure from itself, and its (all too human) programmers.
You should edit/etc/fstab to specify your ext3 partitions as type "auto" rather than ext3. That way, your/etc/fstab is now independent of whether you're running an ext3-enabled kernel, and fsck will know to do log replays (rather than a full fsck) -- however, this latter issue may have been resolved in your distribution, or in later versions of e2fsprogs.
hiya,
ev4 = 21064
ev45 = 21064A
ev5 = 21164
ev56 = 21164A
ev6 = 21264
ev67 = 21264A
and so on. The ev's are the nicknames, the 21's are the release names. The "two-digit" ev's are minor modifications on the base architecture. For example, the 21064A increased the on-chip cache from 8kB to 16kB. The 21164A I think was just a process shrink, but not sure.
nick
Also, with Eclipse, the application you're debugging runs in a separate VM from the development environment. With VisualAge, they ran in the same VM, which has the obvious unfortunate consequences:)
a VLIW architecture needs good compiler support: you gotta file in all the slots of each VLIW instruction ahead of time (at compile time)! How much parallelism can you detect at compile time? Depends on the application, how much code specialization and duplication you wanna.
The idea, though, is that this will be a win in the end (over purely dynamic scheduling) because, among other things, it vastly simplifies the instruction decode stage (and dispatch as well, I think) of the CPU. For certain applications of interest, the instruction decode is the primary bottleneck: whether because you're missing instruction cache, or because you just hafta do so much work to determine the data dependencies, register renaming, etc. that an out-of-order issue processor requires.
What seems strange to me is that the Crusoe is x86 ISA compatible. THis must mean it's doing all the VLIW instruction packing on the fly. My guess is that's not gonna fly, ehhe. What's VLIW buying you in this case?
I applied the preempt patch to 2.4.10. Reboot. Now the IBM 1.2.2 and 1.3 JDKs hang (every time). They work fine on a vanilla 2.4.9, and on 2.4.10-ac4 and 2.4.10-ac10.
Notice that I haven't yet determined whether the bug was in the AA VMs or in the preempt patch:), but thought I'd point this out.
nick
trading link time for run time?
on
KDE 2.2.1 Up
·
· Score: 1
Ok, I'm sure nobody's gonna read this, ehhe, but doesn't this trick trade reduced load time for increased run time?
If I got it correctly, it replaces one indirect function call (to the shared library) with two calls: one direct call (or possibly indirect, if the stub is "far away"), and one indirect call (to the shared library?
the pages you referenced had experiments which showed reductions in link time, but I couldn't find any experiments on how it affects run time.
You assume uniform density of stars. So the chance of star collision on the periphery is much smaller than the probability you gave, but the chance is much higher if center collides with periphery, or even higher if center collides with center. What's the chance of star collision if center collides with center?
don't get cube or espresso if you're want quiet
on
Cappuccino PC Round 2
·
· Score: 1
hey y'all. Fan noise isn't everything. The Mac Cube has a really loud hard drive. The Espresso does, too (both are louder than a typical laptop). What you really want are the designed "quiet pcs" that run at 16.5db or 21.5db (10db means roughly twice as loud, human whispering is around 12-13db, a typical computer is at least 40db).
What about <a href=http://flac.sourceforge.net/>FLAC</a>? Lossless audio compression, that way I'm not beholden to the latest lossy compression fad, and get true CD quality. And, whenever I want to compress further, I can downsample to MP3, or whatever. XMMS plugin, it's got it all? Or am I missing something?
Don't know how either, but <a href=http://www.dealcatcher.com/frame.asp?m=611&am p;c=4876>this screenshot</a> shows eth0 being setup, and grabbing an IP over BOOTP.
What I want to know, and haven't been able to find on any of the screenshots, is what resolution X runs at!?
Remember about, what was it, 4 years back? when monitor manufacturers began inflating monitor sizes. What was once advertised as a 16" monitor became a 17" monitor, despite only having a viewable size of 15.8". As you said, not far off a 15.1" LCD.
Hi. Our site's supported distibution is Redhat. I'm a Debian user of days yore, and am a little lost: is there an equivalent of "apt-get dist-upgrade" on Redhat? In other words, how do I go about upgrading from 7.0 to 7.1?
Regarding yanking out the batteries, you should check out the Journaling Flash File System. Regarding squeezing the kernel onto a PDA, check out the Cram File System (in the 2.4 kernel), which compresses data on a per-page basis.
It's not selfish to want to be able to afford to live decently
Sure it is. As long as there are people with less than you and you don't share, you are being selfish, by the very definition of the word. However, is it "bad" to be selfish? Well, it's probably a matter of degrees, right?
Bill Gates is giving away most of his money. Yet he still lives in a 100 million dollar home. Is that selfish? dunno.
Slashdot Mirror
jeeesus, I was wondering how many fucking thousands of comments I would have to read through before someone mentioned usenet. As I was reading (hehee), I got to thinking: to what extent does slashdot want to retain control over the distribution of their (user-contributed!) content, like some other organizations which control distribution channels of artist-contributed content?
Java, in the form of EJBs and servlets, is becoming fairly common on "big iron". One reason is indeed the cross-platformability of it. But, frankly, most often a change in platform accompanies a change in framework. For example, a switch from UNIX servers to a mainframe may accompany a switch from CGIs to an application server (such as Websphere). The dramatic switch here is probably not from UNIX to OS/390 (after all, OS/390 has a POSIX layer), but from CGIs to Websphere. So, as you said, it's not really cross-platformability which is driving Java in the server realm. I think the principal reason is that Java provides (at least some semblance of :) a language-level security model. By security, I don't mean crypto and virii and external attacks. I mean an application is secure from itself, and its (all too human) programmers.
You should edit /etc/fstab to specify your ext3 partitions as type "auto" rather than ext3. That way, your /etc/fstab is now independent of whether you're running an ext3-enabled kernel, and fsck will know to do log replays (rather than a full fsck) -- however, this latter issue may have been resolved in your distribution, or in later versions of e2fsprogs.
hey AC, does the ev7 use SMT? How many thread contexts? or was that ev8?
hiya, ev4 = 21064 ev45 = 21064A ev5 = 21164 ev56 = 21164A ev6 = 21264 ev67 = 21264A and so on. The ev's are the nicknames, the 21's are the release names. The "two-digit" ev's are minor modifications on the base architecture. For example, the 21064A increased the on-chip cache from 8kB to 16kB. The 21164A I think was just a process shrink, but not sure. nick
Also, with Eclipse, the application you're debugging runs in a separate VM from the development environment. With VisualAge, they ran in the same VM, which has the obvious unfortunate consequences :)
a VLIW architecture needs good compiler support: you gotta file in all the slots of each VLIW instruction ahead of time (at compile time)! How much parallelism can you detect at compile time? Depends on the application, how much code specialization and duplication you wanna.
The idea, though, is that this will be a win in the end (over purely dynamic scheduling) because, among other things, it vastly simplifies the instruction decode stage (and dispatch as well, I think) of the CPU. For certain applications of interest, the instruction decode is the primary bottleneck: whether because you're missing instruction cache, or because you just hafta do so much work to determine the data dependencies, register renaming, etc. that an out-of-order issue processor requires.
What seems strange to me is that the Crusoe is x86 ISA compatible. THis must mean it's doing all the VLIW instruction packing on the fly. My guess is that's not gonna fly, ehhe. What's VLIW buying you in this case?
I'm not an expert on VLIW, but that's my figurin.
when the Levy breaks?
if disk space isn't any issue, why not use FLAC?
I applied the preempt patch to 2.4.10. Reboot. Now the IBM 1.2.2 and 1.3 JDKs hang (every time). They work fine on a vanilla 2.4.9, and on 2.4.10-ac4 and 2.4.10-ac10.
:), but thought I'd point this out.
Notice that I haven't yet determined whether the bug was in the AA VMs or in the preempt patch
nick
Ok, I'm sure nobody's gonna read this, ehhe, but doesn't this trick trade reduced load time for increased run time?
If I got it correctly, it replaces one indirect function call (to the shared library) with two calls: one direct call (or possibly indirect, if the stub is "far away"), and one indirect call (to the shared library?
the pages you referenced had experiments which showed reductions in link time, but I couldn't find any experiments on how it affects run time.
nick
You assume uniform density of stars. So the chance of star collision on the periphery is much smaller than the probability you gave, but the chance is much higher if center collides with periphery, or even higher if center collides with center. What's the chance of star collision if center collides with center?
hey y'all. Fan noise isn't everything. The Mac Cube has a really loud hard drive. The Espresso does, too (both are louder than a typical laptop). What you really want are the designed "quiet pcs" that run at 16.5db or 21.5db (10db means roughly twice as loud, human whispering is around 12-13db, a typical computer is at least 40db).
What about <a href=http://flac.sourceforge.net/>FLAC</a> ;? Lossless audio compression, that way I'm not beholden to the latest lossy compression fad, and get true CD quality. And, whenever I want to compress further, I can downsample to MP3, or whatever. XMMS plugin, it's got it all? Or am I missing something?
hehe, ok that was a pretty lame attempt to add a link. Here's a better link to the screenshot.
Don't know how either, but <a href=http://www.dealcatcher.com/frame.asp?m=611&am p;c=4876>this screenshot</a> shows eth0 being setup, and grabbing an IP over BOOTP.
What I want to know, and haven't been able to find on any of the screenshots, is what resolution X runs at!?
Remember about, what was it, 4 years back? when monitor manufacturers began inflating monitor sizes. What was once advertised as a 16" monitor became a 17" monitor, despite only having a viewable size of 15.8". As you said, not far off a 15.1" LCD.
Hi. Our site's supported distibution is Redhat. I'm a Debian user of days yore, and am a little lost: is there an equivalent of "apt-get dist-upgrade" on Redhat? In other words, how do I go about upgrading from 7.0 to 7.1?
thanks!
nick
shouldn't you just point your disk cache to /tmp or /var/tmp/username?
Trouble indeed, trouble on triton. Hehe, well, at least the book was good.
Regarding yanking out the batteries, you should check out the Journaling Flash File System. Regarding squeezing the kernel onto a PDA, check out the Cram File System (in the 2.4 kernel), which compresses data on a per-page basis.
check it out
so the play Oedipus isn't violent, because in it, he pokes his eyes out offstage?