I'm really interested in "reliably works at my house".
Next in line would be "stop gouging for bandwidth".
After that "better battery life".
Somewhere way down the list is "faster".
Too obtuse for Slashdot?
OK I'll spell it out for you. "the system does feel a lot snappier, more responsive, and just overall a much more pleasant user experience." This is user perception, and while valid, doesn't provide any measurable details. It reminded me of "audiophile" equipment reviews where the entire test is wrapped in unmeasurable, subjective, and arguably useless terms like "wider soundstage" and "more lifelike". So I pulled some text directly from an audio cable site and made it fit Linux. I thought it would be funny that kernels would require a break-in period, just as I think it is absurd that a *wire* requires a break-in period after which it provides unmeasurable but perceptually improved performance.
I am sorry that nobody else found it funny. I'll go away again now.
The Break-In process of the new Linux kernel takes time. There is a significant change in Video Performance as the kernel break in. There is still a perception that Linux kernels have a short break in time or worse yet, don't require break in. Some hackers used a second computer to break in the Linux kernel, and transfer the image to their primary computer. This method will not appreciably reduce the break in time required for the kernel. Linux kernel Break-In must be done in the position where you plan to use it.
The System Performance Stages of the kernel are as follows:
* First Stage of Break-In = The system will feel very open, clear and with good detail resolution and dynamics. The greens and lower reds will have elevated intensity levels. The lower output of the blue and green information is due to reduced bandwidth performance at this Stage. In some systems, especially with aluminum or titanium heatsinks, the greens and blues may appear edgy or even fatiguing. The visual stage will appear OK with some lack of Focus. It will take from 5 to 15 hours of break-in for the kernel to reach the Second Stage of Break-In.
* Second Stage of Break-In = The blues and greens will appear less elevated and up front as the monitor intensity level increases. This is followed by the reds starting fill in. The lack of Focus may become more noticeable and the visual stage will start to widen and have more depth. It will take an additional 15 to 35 hours to reach the Third Stage of Break-In.
* Third Stage of Break-In = The system response time will completely flatten out. The presentation will become very clean and less up front. The lack of Focus is disappearing and the imaging will improve as will the low level detail resolution. The Green comes in and it is very tight and you will see lower Red frequencies than your other kernel provided. In effect the visual signature of the kernel will seem to disappear and the X-window presentation will be very real and non-fatiguing. It will take an additional 30 to 50 hours to reach the Final Stage of Kernel Performance.
* Fourth and Final Stage of Kernel Performance = The Visual Stage will be wider than your monitor with excellent depth, height and precise localization of individual icons on the desktop. The hue of the icons will be very accurate over the entire desktop. Symbolic links have excellent referencing ability. The metallic sound of your hard drive is very lifelike. Rhythm, Pace and Dynamics are effortless. Many users find they are now viewing the X-window system at lower Light Levels due to the effortless presentation. You will start to see subtle visual cues like the programmer turning his head while he is programming. You will find you are viewing the Window Manager and forgetting about evaluating your system.
I used to believe that, until I supported Ralph Nader in 2000, and George W. Bush won the presidential contest as a result of people like me thinking exactly like you. It is not "as simple as that."
In Florida in the year 2000, George W. Bush beat Al Gore by only 537 votes, while Ralph Nader drew 97,488 votes on the Green party ticket. Had Nader not been running, those environmental votes would have overwhelmingly fallen to Al Gore and we would have avoided the entire second Bush administration. I voted(1) Green on environmental issues, of all things, as if Al Gore wasn't green enough.
We got Bush instead.
Now I vote more pragmatically, not for who I think will win, but definitely not for an idealistic vote-splitter when there is a realistic shot of making incremental change with a realistic candidate.
God Bless the Freaks for they make the world more interesting, but don't vote for them in the general election.
(1) not in Florida so ultimately it didn't matter, but people just like me tilted the election to Bush.
And that would be a small problem for the 4% of all men that cannot distinguish between red and green. What I want to know is what makes the other 96% of all men that cannot distinguish between red and green so special, if only 4% of them are affected anyway...
Luxury ISP in my book is one that gives me a dial-in number with plain old ppp and leaves me alone after that.
I gave SBC/Yahoo dial a whirl, and for $9.95/month (the price they gave me when they realized they couldn't sell me DSL) I was pretty pleased. It took a Windows laptop to run the install software and get the dial-in number, but once I had that, I was able to make it work perfectly from Linux using wvdial. It would stay connected for something like 6 hours at a time, and was on pretty much constantly due to dial-on-demand and ntpd running. I did this for at least a year - can't remember exactly when previous ISP turned evil. Can't speak to UDP traffic, but if it is like the rest of the service, it would probably just work.
Don't get me wrong; I hate SBC like any sane person, but their dial-up access worked very well for me. I can only guess as to why they didn't kick a customer like me off - my guess is that the dialup department can't talk to the billing or cancellation departments and there was no way internally to report my, ahm, overuse of the dialup network. I called to cancel last week and they seemed very sorry to lose me as a customer (got wireless innernet now).
Sorry for the US centric post (mid-south is SBC land). Hope this helps someone stuck with a sorry or expensive ISP. YMMV.
I could look it up, or I could dismiss it as meaningless because age and I/Q don't progress linearly, period. Any formula that tries to do so is meaningless, even if you know what the terms are. That was my point; perhaps it was too subtle for you, AC.
For Oracle, in terms of CPU power and memory bandwidth, HP's DL585 with four dual-core Opterons running Linux in 64-bit mode is a very attractive option. Getting it to work with a SAN is a challenge, but if you can get similar Oracle performance with locally attached storage and save many hundreds of thousands of dollars in the processes, it might be worth looking into.
I don't work for HP or AMD or IBM, but the DL585 absolutely makes Oracle fly. Any company still pursuing high performance computing on Itanium or P4 is barking up the wrong tree, HP included.
I bought 4 of them (WRT54GS v2.0) and loaded sveasoft alchemy public version on them in order to create a very extended bridged network using wds. The outer two are connected to separate physical networks, and the inner two only have power (no physical network connection) and are situated in weatherproof boxes in the treetops. I am posting this note over them now!
On hindsight, OpenWRT would have been a better choice, but for the minimally-enhanced functionality I was looking for (wds and a shell), sveasoft is dead easy. Yes, you run a risk and void your warranty. But you only go around once, eh?
Make sure your chosen firmware version is known to work on your particular linksys hardware. Here is a good page outlining the differences in Linksys versions. Sveasoft Alchemy doesn't specifically support my version but I found a reference of someone doing it, and it does work for me.
Good luck / happy hacking, and if you mess up, you want to search for the work "debricking":)
Perhaps the reason these things aren't being brought to market is less the result of "them" and more the result of the fact that they don't work, or are of little practical value.
Your specific example patent, in my unstudied opinion, either violates the 2nd law of thermodynamics or eats rare-earth magnets wholesale. Even if it does in fact work as described, it's not very useful. You will never get out of it more energy than is used in the creation of the magnet.
Assuming this device even works, do you think that rare-earth magnets are more energy dense than, say, lead-acid batteries? Have you ever tried to handle a large rare-earth magnet safely?
Probably so for kleptomaniacs, but you can't draw any general conclusions from that, such as most people in low paying jobs are there because they have some social disorder(*). I've been poor before - poor enough to steal the sandpaper-like toilet paper from a site I was working as a janitor, of all things. Desperation is a funny thing. I haven't been tempted to steal anything since I started drawing a real paycheck.
(*) in fact, that sort of elitist bullshit really pisses me off. You didn't state it, but the intention of your comment seems to slant that way to me. Forgive me if I am wrong.
You missed this gem: "Unfortunately, the Phoenix name was already used by a web browser that ran on top of a BIOS".
Admittedly, that would be really cool (kinda like
LinuxBIOS but with just a web browser?), but I don't think the Phoenix BIOS was a web browser, by pretty much any definition of the term. Is that what the author really meant?
You sound like you know what you are talking about, but your 10 bullets are not silver either. Once you have those down you can move on to areas where they don't apply. Some counterpoints:
(2) learn when to NOT normalize. If most of your access patterns pull data from several tables, you can avoid expensive joins by not over-normalizing. You can build the world's best data model where nothing is replicated and you'll have a system that can't answer anything without multiple table joins. That's OK in OLTP systems - but not OK in the world of DSS.
(4) On low-end hardware, you know of what you speak, but it doesn't hold true if the raid controller's cpu is fast and the cache is large. Sure you can create data files faster on 0+1, but on customer queries (even sort intensive ones), you can purchase raid-5 controllers fast enough that your bottleneck is somewhere else. Does the controller and FC disk cost eat up all your savings over a cheaper raid-0+1 on ATA or SCSI disk kit? probably.
(5) old school. New world order is stripe-and-mirror-everything. Disk controllers are smart enough these days that seek contention on multiple I/O streams is a thing of the past. I wouldn't waste valuable controller bandwidth on something that is going to be utilized 5% of the time. The key is to have *lots* of channels, not necessarily *separate* channels, and that boils down to spend more, get a faster database server.
(6) Damn straight. Can I get an "AMEN". And here you thought I was refuting points.
(9) Learn when to NOT use an index. You've had the "ah-HA" moment when adding an index to a frequently accessed table speeds up queries dramatically, but if your access patterns have you constantly pulling more than a small fraction of the rows, tune for a hash join or a full table scan on a single table... but not without seeing (5) above. Indexes can *kill* your performance if your customers ask the wrong questions. And then you miss your update window due to index maintenance... PostGreSQL has hash joins now, for which I am very happy. Too bad it's still 1/3rd the speed of Oracle*.
Everything say is spot-on in some, maybe most cases, so I know you are a true fellow database geek... just one with a different perspective than mine.
(*) when Oracle is working right. When it doesn't work right, damn near everything is faster than Oracle. If you want to be blown away by a fast database system - go benchmark a Netezza. No, I don't work for them.
Aw, man! You had me goin' there for a minute. I was really hoping to read an informative essay on the topic of cow manure, but you just pulled that out of... your... ass... figuratively speaking.
Or you can just use the case as both a ground and somthing to solder to. Here's a picture of an audio amplifier for boosting weak headphone output to listenable levels for speakers. It sounds better with a couple altoids in it (not really, but maybe we can sell that to the audiophile community?)
I saw it too. Backed up and re-clicked, and it's gone. I really didn't need to see that... is Wikipedia being severly hacked? Maybe one of their servers is only serving the "autofellatio" page?
Well anyway... you learn something new every day it seems.
Parent is modded funny, but this is a very reasonable solution; that's what I do at my house and it works fantastically.
You can get a nice little PLL FM transmitter from ccrane.com for around US$70. With the addition of an antenna wire that is about 6 feet long on the transmitter, I get hiss-free FM stereo that sounds certainly good enough for me, in every room of my house and out in the shop. I've ripped every CD I own and put that on "shuffle" from the main PC, and have either nice stereos (living room and shop) or rinky jam-box radios (bathrooms) about everywhere I can expect to be. Two NI-MH batteries will last for at least a couple days in the thing, or you can use the wall-wort and not mess with transmitter batteries at all.
Implementation notes: (1) PLL synthesized FM is important; don't get a kit without it or your transmitter frequency will drift and your digitally-tuned receivers will get fuzz. (2): There are about 15 radio stations in my local market but I'm several miles out; YMMV if you are in a crowded market with no blank spaces on the FM dial. (3) When you are tired of your digital audio collection, you can always listen to the radio:)
Ha! On CNN I found this: The Huygens probe, about the size of a Volkswagen Beetle,... I *LOVE* it when people use a Volkswagen Beetle as a unit of measurement! Now if Google Calculator would only convert that to cubic meters for me...
What did you do? Post as AC on Slashdot? How helpful.
Do try to remember there are ugly people everywhere, not just in the U.S.A., and conversely there are GOOD people everywhere, even in the U.S.A.
Maybe when you are a little older you will understand. It's your brand of hatred and bigotry that fuels things like, oh, the Iraq war, the attack on Spain's trains, etc.
Nothing much; Redhat AS 2.1 and Oracle on most of the Xeon systems, but idle at time of benchmark. Most likely gzip fits well into the cache of both your CPU and the Xeons, and it could be that your gzip is more optimized than Redhat AS 2.1's gzip. From the numbers it looks to measure clock speed more than anything else, but it's still fun to compare.
I agree, but I figured someone would point out that compressing a stream of zeros is not quite like compressing real data - but at least the stream of zeros is available on (nearly) all platforms, and it is very nearly CPU bound and so a pretty good relative test, for what that is worth.
I hate waiting on gzip so much that I wrote a multi-threaded program to make gzip files faster on machines with >1 CPU; see http://lemley.net/smp.html if you have more than one CPU and hate to wait on compression programs.
Slashdot Mirror
I'm really interested in "reliably works at my house". Next in line would be "stop gouging for bandwidth". After that "better battery life". Somewhere way down the list is "faster".
I've been spending a lot more time on fark.com lately. I'm perhaps not getting as much done...
Too obtuse for Slashdot? OK I'll spell it out for you. "the system does feel a lot snappier, more responsive, and just overall a much more pleasant user experience." This is user perception, and while valid, doesn't provide any measurable details. It reminded me of "audiophile" equipment reviews where the entire test is wrapped in unmeasurable, subjective, and arguably useless terms like "wider soundstage" and "more lifelike". So I pulled some text directly from an audio cable site and made it fit Linux. I thought it would be funny that kernels would require a break-in period, just as I think it is absurd that a *wire* requires a break-in period after which it provides unmeasurable but perceptually improved performance.
I am sorry that nobody else found it funny. I'll go away again now.
I'm glad somebody got it. What is it with kids these days?
The Break-In process of the new Linux kernel takes time. There is a significant change in Video Performance as the kernel break in. There is still a perception that Linux kernels have a short break in time or worse yet, don't require break in. Some hackers used a second computer to break in the Linux kernel, and transfer the image to their primary computer. This method will not appreciably reduce the break in time required for the kernel. Linux kernel Break-In must be done in the position where you plan to use it.
The System Performance Stages of the kernel are as follows:
* First Stage of Break-In = The system will feel very open, clear and with good detail resolution and dynamics. The greens and lower reds will have elevated intensity levels. The lower output of the blue and green information is due to reduced bandwidth performance at this Stage. In some systems, especially with aluminum or titanium heatsinks, the greens and blues may appear edgy or even fatiguing. The visual stage will appear OK with some lack of Focus. It will take from 5 to 15 hours of break-in for the kernel to reach the Second Stage of Break-In.
* Second Stage of Break-In = The blues and greens will appear less elevated and up front as the monitor intensity level increases. This is followed by the reds starting fill in. The lack of Focus may become more noticeable and the visual stage will start to widen and have more depth. It will take an additional 15 to 35 hours to reach the Third Stage of Break-In.
* Third Stage of Break-In = The system response time will completely flatten out. The presentation will become very clean and less up front. The lack of Focus is disappearing and the imaging will improve as will the low level detail resolution. The Green comes in and it is very tight and you will see lower Red frequencies than your other kernel provided. In effect the visual signature of the kernel will seem to disappear and the X-window presentation will be very real and non-fatiguing. It will take an additional 30 to 50 hours to reach the Final Stage of Kernel Performance.
* Fourth and Final Stage of Kernel Performance = The Visual Stage will be wider than your monitor with excellent depth, height and precise localization of individual icons on the desktop. The hue of the icons will be very accurate over the entire desktop. Symbolic links have excellent referencing ability. The metallic sound of your hard drive is very lifelike. Rhythm, Pace and Dynamics are effortless. Many users find they are now viewing the X-window system at lower Light Levels due to the effortless presentation. You will start to see subtle visual cues like the programmer turning his head while he is programming. You will find you are viewing the Window Manager and forgetting about evaluating your system.
I used to believe that, until I supported Ralph Nader in 2000, and George W. Bush won the presidential contest as a result of people like me thinking exactly like you. It is not "as simple as that."
In Florida in the year 2000, George W. Bush beat Al Gore by only 537 votes, while Ralph Nader drew 97,488 votes on the Green party ticket. Had Nader not been running, those environmental votes would have overwhelmingly fallen to Al Gore and we would have avoided the entire second Bush administration. I voted(1) Green on environmental issues, of all things, as if Al Gore wasn't green enough.
We got Bush instead.
Now I vote more pragmatically, not for who I think will win, but definitely not for an idealistic vote-splitter when there is a realistic shot of making incremental change with a realistic candidate.
God Bless the Freaks for they make the world more interesting, but don't vote for them in the general election.
(1) not in Florida so ultimately it didn't matter, but people just like me tilted the election to Bush.
Luxury ISP in my book is one that gives me a dial-in number with plain old ppp and leaves me alone after that.
I gave SBC/Yahoo dial a whirl, and for $9.95/month (the price they gave me when they realized they couldn't sell me DSL) I was pretty pleased. It took a Windows laptop to run the install software and get the dial-in number, but once I had that, I was able to make it work perfectly from Linux using wvdial. It would stay connected for something like 6 hours at a time, and was on pretty much constantly due to dial-on-demand and ntpd running. I did this for at least a year - can't remember exactly when previous ISP turned evil. Can't speak to UDP traffic, but if it is like the rest of the service, it would probably just work.
Don't get me wrong; I hate SBC like any sane person, but their dial-up access worked very well for me. I can only guess as to why they didn't kick a customer like me off - my guess is that the dialup department can't talk to the billing or cancellation departments and there was no way internally to report my, ahm, overuse of the dialup network. I called to cancel last week and they seemed very sorry to lose me as a customer (got wireless innernet now).
Sorry for the US centric post (mid-south is SBC land). Hope this helps someone stuck with a sorry or expensive ISP. YMMV.
I could look it up, or I could dismiss it as meaningless because age and I/Q don't progress linearly, period. Any formula that tries to do so is meaningless, even if you know what the terms are. That was my point; perhaps it was too subtle for you, AC.
l cuate.html. Specifically, Because of these problems, MA is no longer used in calculating IQ scores - instead "deviation IQ" is used.
A reference:
http://www.wilderdom.com/intelligence/IQHistoryCa
3 years loss at age 11 is an IQ of 100*8/11 or 73
Um, no. If that were the case, I would have an IQ of 100*35/11, or 318.
Instead, I am posting on Slashdot.
For Oracle, in terms of CPU power and memory bandwidth, HP's DL585 with four dual-core Opterons running Linux in 64-bit mode is a very attractive option. Getting it to work with a SAN is a challenge, but if you can get similar Oracle performance with locally attached storage and save many hundreds of thousands of dollars in the processes, it might be worth looking into.
I don't work for HP or AMD or IBM, but the DL585 absolutely makes Oracle fly. Any company still pursuing high performance computing on Itanium or P4 is barking up the wrong tree, HP included.
I bought 4 of them (WRT54GS v2.0) and loaded sveasoft alchemy public version on them in order to create a very extended bridged network using wds. The outer two are connected to separate physical networks, and the inner two only have power (no physical network connection) and are situated in weatherproof boxes in the treetops. I am posting this note over them now!
:)
On hindsight, OpenWRT would have been a better choice, but for the minimally-enhanced functionality I was looking for (wds and a shell), sveasoft is dead easy. Yes, you run a risk and void your warranty. But you only go around once, eh?
Make sure your chosen firmware version is known to work on your particular linksys hardware. Here is a good page outlining the differences in Linksys versions. Sveasoft Alchemy doesn't specifically support my version but I found a reference of someone doing it, and it does work for me.
Good luck / happy hacking, and if you mess up, you want to search for the work "debricking"
Perhaps the reason these things aren't being brought to market is less the result of "them" and more the result of the fact that they don't work, or are of little practical value.
Your specific example patent, in my unstudied opinion, either violates the 2nd law of thermodynamics or eats rare-earth magnets wholesale. Even if it does in fact work as described, it's not very useful. You will never get out of it more energy than is used in the creation of the magnet.
Assuming this device even works, do you think that rare-earth magnets are more energy dense than, say, lead-acid batteries? Have you ever tried to handle a large rare-earth magnet safely?
Probably so for kleptomaniacs, but you can't draw any general conclusions from that, such as most people in low paying jobs are there because they have some social disorder(*). I've been poor before - poor enough to steal the sandpaper-like toilet paper from a site I was working as a janitor, of all things. Desperation is a funny thing. I haven't been tempted to steal anything since I started drawing a real paycheck.
(*) in fact, that sort of elitist bullshit really pisses me off. You didn't state it, but the intention of your comment seems to slant that way to me. Forgive me if I am wrong.
Cynical remark about HP's misdirection. Outmoded sentimental longing for superior non-Intel processors. "If-only" scenario. Obvious comment about Itanium. Snarky unsubstantiated armchair prediction.
--
sig
Admittedly, that would be really cool (kinda like LinuxBIOS but with just a web browser?), but I don't think the Phoenix BIOS was a web browser, by pretty much any definition of the term. Is that what the author really meant?
You sound like you know what you are talking about, but your 10 bullets are not silver either. Once you have those down you can move on to areas where they don't apply. Some counterpoints:
(2) learn when to NOT normalize. If most of your access patterns pull data from several tables, you can avoid expensive joins by not over-normalizing. You can build the world's best data model where nothing is replicated and you'll have a system that can't answer anything without multiple table joins. That's OK in OLTP systems - but not OK in the world of DSS.
(4) On low-end hardware, you know of what you speak, but it doesn't hold true if the raid controller's cpu is fast and the cache is large. Sure you can create data files faster on 0+1, but on customer queries (even sort intensive ones), you can purchase raid-5 controllers fast enough that your bottleneck is somewhere else. Does the controller and FC disk cost eat up all your savings over a cheaper raid-0+1 on ATA or SCSI disk kit? probably.
(5) old school. New world order is stripe-and-mirror-everything. Disk controllers are smart enough these days that seek contention on multiple I/O streams is a thing of the past. I wouldn't waste valuable controller bandwidth on something that is going to be utilized 5% of the time. The key is to have *lots* of channels, not necessarily *separate* channels, and that boils down to spend more, get a faster database server.
(6) Damn straight. Can I get an "AMEN". And here you thought I was refuting points.
(9) Learn when to NOT use an index. You've had the "ah-HA" moment when adding an index to a frequently accessed table speeds up queries dramatically, but if your access patterns have you constantly pulling more than a small fraction of the rows, tune for a hash join or a full table scan on a single table... but not without seeing (5) above. Indexes can *kill* your performance if your customers ask the wrong questions. And then you miss your update window due to index maintenance... PostGreSQL has hash joins now, for which I am very happy. Too bad it's still 1/3rd the speed of Oracle*.
Everything say is spot-on in some, maybe most cases, so I know you are a true fellow database geek... just one with a different perspective than mine.
(*) when Oracle is working right. When it doesn't work right, damn near everything is faster than Oracle. If you want to be blown away by a fast database system - go benchmark a Netezza. No, I don't work for them.
Aw, man! You had me goin' there for a minute. I was really hoping to read an informative essay on the topic of cow manure, but you just pulled that out of ... your... ass... figuratively speaking.
Or you can just use the case as both a ground and somthing to solder to. Here's a picture of an audio amplifier for boosting weak headphone output to listenable levels for speakers. It sounds better with a couple altoids in it (not really, but maybe we can sell that to the audiophile community?)
It's one of a kind, that's for sure.
I saw it too. Backed up and re-clicked, and it's gone. I really didn't need to see that... is Wikipedia being severly hacked? Maybe one of their servers is only serving the "autofellatio" page?
Well anyway... you learn something new every day it seems.
Parent is modded funny, but this is a very reasonable solution; that's what I do at my house and it works fantastically.
:)
You can get a nice little PLL FM transmitter from ccrane.com for around US$70.
With the addition of an antenna wire that is about 6 feet long on the transmitter, I get hiss-free FM stereo that sounds certainly good enough for me, in every room of my house and out in the shop. I've ripped every CD I own and put that on "shuffle" from the main PC, and have either nice stereos (living room and shop) or rinky jam-box radios (bathrooms) about everywhere I can expect to be. Two NI-MH batteries will last for at least a couple days in the thing, or you can use the wall-wort and not mess with transmitter batteries at all.
Implementation notes: (1) PLL synthesized FM is important; don't get a kit without it or your transmitter frequency will drift and your digitally-tuned receivers will get fuzz. (2): There are about 15 radio stations in my local market but I'm several miles out; YMMV if you are in a crowded market with no blank spaces on the FM dial. (3) When you are tired of your digital audio collection, you can always listen to the radio
Ha! On CNN I found this: ...
The Huygens probe, about the size of a Volkswagen Beetle,
I *LOVE* it when people use a Volkswagen Beetle as a unit of measurement! Now if Google Calculator would only convert that to cubic meters for me...
What did you do? Post as AC on Slashdot? How helpful.
Do try to remember there are ugly people everywhere, not just in the U.S.A., and conversely there are GOOD people everywhere, even in the U.S.A.
Maybe when you are a little older you will understand. It's your brand of hatred and bigotry that fuels things like, oh, the Iraq war, the attack on Spain's trains, etc.
Nothing much; Redhat AS 2.1 and Oracle on most of the Xeon systems, but idle at time of benchmark. Most likely gzip fits well into the cache of both your CPU and the Xeons, and it could be that your gzip is more optimized than Redhat AS 2.1's gzip. From the numbers it looks to measure clock speed more than anything else, but it's still fun to compare.
I agree, but I figured someone would point out that compressing a stream of zeros is not quite like compressing real data - but at least the stream of zeros is available on (nearly) all platforms, and it is very nearly CPU bound and so a pretty good relative test, for what that is worth.
I hate waiting on gzip so much that I wrote a multi-threaded program to make gzip files faster on machines with >1 CPU; see http://lemley.net/smp.html if you have more than one CPU and hate to wait on compression programs.