Yep, I bet Sacramento is a real slum. I mean it can't compare with most of the "progressive" cities in the world like Detroit, LA and NY. You probably have a much worse time with drugs, gangs,teen pregnancy and homelessness than any of those cities. I can see why you hate it so much there. Heck, I have been to Sacramento a few times and I feel lucky to get out with my life in hand:-)
It is kind of amusing that people are so worried about offending a very very very small minority of people, so much so that they want to change one of the founding principles that the U.S was built on and is still being built on. All for the fact that someone is offended? Well I guess I am offended at the fact that children can't pray in school anymore. Hmmm now we have two sides of people that are offended. How did the U.S. generally decide such issues? By the laws. Those laws were generally made by their representatives who (mostly) acted on behalf of their people. You will find that almost NO congressman will take this issue on. Why?
Now we have a very small majority of people that hate those decisions and laws. Those people see no way to change the system except through the very liberal court system. (I.E. ACLU) Some of these judges tend to make up laws instead of ruling on existing law, and that is where the problems come in.
The good news is that the majority of people are starting to wake up in the U.S.
Most of our databases fit in to RAM. I do agree with what you are saying in general, however in our case it isn't true. The two biggest factors that we have is the amount of on die cache and the FSB speed. Now, as mentioned by other posters, the Opteron handles dual processor to RAM far more efficently than a dual Xeon. Also the Xeons that it is being compared against only has 512k of on die cache.
The core question that has to be asked is if you need dual processors at all. If not then a Xeon or new P4 would probably be your best bet, but if you do then the dual Opteron is a great buy for processors. Take this plus the fact that you can easily get one of these boxes with 4 GIG of RAM now, and disk i/o becomes less of a factor. Heck, if you want to throw some more money at it, you can actually get one of these boxes with 12 GIG of RAM (If someone makes 333FSB 2GB DDR chips). With 12 GB or RAM that would hide some serious disk i/o issues.
We have a dual Opteron 1.8GH and a 2GH on order. We found that with RedHat and Oracle the Opteron in 32 bit mode beat the crap out of a quad Xeon for the stuff we do.
Just an F.Y.I. Again this was with 32 bit code. I tried the RedHat BETA and it wouldn't even boot up without locking.
So given that Oracle cost us over $20k a processor, we saved over 40 grand!
From what I see slashdot is nothing more than a news server. Yes it does very simple writes, but mostly reads from a database. Yes it does maintain state for a user, but it is a VERY simple application. It behaves the same way (with preferences) for every user. It also has some significant hardware behind it. I would be willing to bet that 90+% of Slashdot is reads from a DB. That fits perfectly with MySQL, again this is not indicitive of most businiess web apps. You mentioned 8 web servers.
I would like to know the people that check their connection speed with slashdot, because it's performance varies greatly throughout the day, and they should be informed that there are far better ways to test their connection speed.
Back to the original posted comment..... Sun has given a lot to the open source community. They have continued to develop OpenOffice, GNOME and other packages. I see this as a huge win for open-source fans. Lets say sun somehow gets 5-15% marketshare for this product. It will then make other 3rd party developers start developing drivers and applications for their crap to work with Linux. Then people like me could more easily suggest that we replace some of our business with an open source model.
I could see call centers loving this type of product. Specifically Cendant call centers.
Well... I am a AMD fan as well, and just purchased all new AMD systems for our x86 users at work, so this is somewhat difficult for me to write, but if AMD's bet on 64bit technology doesn't pay off, and Windows doesn't support it well, then AMD will be dead. Their financial report has sucked the last year, and they cannot afford to make the mistakes that Intel has made. Intel, like Microsoft can afford to make some very stupid mistakes (Rambus, Itanic ect..) and still recover, but AMD cannot. I hope and pray that the Opteron does great for servers and desktops, but I won't be shocked if AMD is out of business in two years.
Now for PPC. The PPC has done very well in routers and other non PC stuff. It has kinda sucked in PC's in comparision. Motorolla (another company I like, from my Amiga days), has decided to focus off the PC market and on other segments. They are just going to focus more on battery life and other things, than all out power. This left IBM and Apple to take up the lions share of development and fabs for PPC based PC chips. Luckly for Apple, IBM has decided to do it for now. I hope that IBM continues to produce great desktop and server PPC chips for years to come, but the development dollars that it is costing them must be returned some day, and Apples marketshare + AIX systems don't seem to add up to the cost of R&D + fabs for them. Take that and add that Sony and IBM are working on this "Cell" technology, and it appears that the PPC chips days may be numbered. So if you were Apple and IBM was moving to "Cell" chips AND you had to migrate your system again, what would you migrate to? Cell or X86? Or would you decide to take over the fabs, and continue the chip production and design yourself?
Man, I am about one of the biggest Apple fans around. Just because I say Apple may switch to Intel or any x86 chip doesn't make me a troll.
Some of the developers I know, have said that Apple has working hardware on x86 already. They would be foolish not to be prepared to switch, if the PPC chip can't compete. Having said that, I am also a fan of the PPC chip. I hope that it beats the crap out of Intel and AMD, but I just don't see that happening. It is hard to compete with 2 BILLION in R&D EVERY year.
I (like many others) would still buy a mac, with an X86 chip in it. What I pay for is the integration of hardware and software (kinda like Sun/Solaris). Their stuff just works. As long as they continue that mode of operation, I will be a fan.
Ok, I have a Athlon 3000+ and a new G5 Mac, and also have a dual processor Opteron system. The previous poster is correct. The heat syncs on the new G5 are at least 5X as big as any heat sync I have EVER seen on a Intel/AMD system. I honestly don't know what the processor temprature is when it runs, but the noise level of this new G5 is as loud as any modern PC that I have seen. This is a 1.8GH single processor Mac system.
From what I have heard this the current G4 at 1.1 GH takes around 14 watts of power, and the current G5's take more than double that amount at 2GH. Some believe that they could release a G5 based system running at 1.4GH and get the power down to around 16 watts, but why? Speed isn't what people are screaming for with laptops and you would loose more battery life.
I personally think it will be at least a year and a half before you will see a G5 laptop. I think you will see Apple running on Intel (Itanium) or AMD(Opteron) first.
My overall point stands true: It cost far more time and money to take "solid" manufactured goods and bring them in to the U.S.
Some of those goods are taxed, some at high levels.
On a side note: NAFTA has been horrible for the U.S. I still consider myself a conservative, but with idiots like Rush out there saying how great this was, and that smoking isn't bad for you.... Well it just sounds rediculous. There was a giant sucking sound of manufacturing jobs to Mexico, and the U.S. is not any better for it.
Put that job on Monster and see how many resume's you get of people that are out of work. It will shock you.
I live in the midwest, and I can tell you that it is a bloodbath around here. Multiple startups have died, almost every software development company has folded, and there are only a few companies that still have I.T. shops internal. So yes, some smaller companies are hireing one person, but when 4 or 5 HUGE companies cut their entire I.T. staff (hundreds), that sucks bad.
Most companies that I know are "getting by" and they need to hire someone else, but actually doing it is another thing...
Just to give you an example of what I know. Three years ago we had to pay over $100/hour for Java developers. We now pay $20-$25 and have little trouble getting them.
We are getting ready to start looking for a partime Oracle DBA, and I know that we will be able to get one for a very low amount of money, again we are talking in the $20-$25 range.
I feel bad even offering this kind of money for those skills, but as most of these people put it "I make nothing now, so this is better than that".
I will admit that we do still pay $100+/hour for "specific" needs, like a "4 hour" network job, but those are VERY rare.
There is a HUGE difference between manufactured goods and software development. It takes effort and money to bring goods in to the U.S. and those goods are generally taxed. Software development isn't, and it takes little effort to "move" code.
The other core difference is that the other jobs took more than 10 years to move offshore, this has taken around 2.
I do find it ironic that hardly ANY open source development gets done by Indian programmers though...
Also, this will just speed up the use of unions for those remaining I.T. workers here in the U.S. Most computer science people are conservatives by nature , and I look for this one issue to drive a large percentage to vote against the Republican party in the next election.
Well I have had my dual Opteron system for about three weeks now, AND I ordered a G5 over a month ago.... The G5 still hasn't shipped to me (I ordered the 1.8GH model). Even if it did the Opteron was out more than a month ago.
Now granted the article should say something like, this is the first somewhat affordable x86 chip to provide 64 bit computing.
There are many other links that have excellent benchmark results of the Opteron. Specifically, toms hardware has a great review, but to sum most of the review up... The dual Opteron 244 kicks the crap out of a dual Xeon in most linux server type benchmarks. This is WITH 32 bit code. The next step is to get code that is still 32 bit but compiled for the Opteron. i.e. Oracle wich is out now. This gives you another small gain in performance and then lastly you can test again with pure 64 bit compiled programs. The last "MAY" give you some performance increase over 32 bit programs recompiled, but that depends on the program. An example would be a simple "hello world" would not run any faster, but a database that needs more than 4GB of memory should perform better.
So for our company we didn't compare the Opteron to a Itanium, but to a Xeon and the dual Opteron smokes it.
Note, that I always say dual Opteron, because a single Opteron doesn't perform any better than the current 32bit X86 chips. In most cases it performs worse. One of the key factors with the Opteron is the way the chips talk to the front side bus. This is currently the bottleneck in most dual/quad processor x86 systems. AMD has somewhat handled this bottleneck by allowing both chips to have a separate path to "some" RAM and providing a larger amount of data to pass from processor to RAM. I can't speak for the quad and eight way systems, but for the dual processor systems this makes a HUGE difference.
In our environment, a dual Opteron 1.8GH beats a QUAD 2.8GH Xeon in performance (with Oracle). Given that Oracle charges us 15k per CPU + 4k a year, that is a huge savings, and we are not forced to go to a pure 64 bit system today.
Intel addresses the front side bus issue to RAM by increasing the on die cache of the Xeon. So they released a 1MB version of the Xeon 3GH (they had a 512MB version before), to combat the Opteron. The problem is that it didn't help that much. Intel's solution would be to include X86 code, "done well, not the crappy way it is done now" in the Itanium, and lower the price to under a grand. I doubt they will do that any time soon, so it looks like this will give AMD a good path to the small server market. Factor that in with IBM and SUN wanting to use their stuff in new server, things look ok for the Opteron.
I still don't know how well the AMD Athlon 64 will be though. That chip isn't due to be released until next month. That is the replacement for the current Athlons, and for AMD's sake it better beat a P4 in most Windows stuff.
Well, this has been dead for a while, but I just saw it. Notice that I said "at that time" Telnet and FTP where developed WAY before 1995. Their lack of security can be somewhat forgiven because at that time people had ssh and secure ftp if they wanted to use it.
Many people sign wavers not to sue a company and legally get to sue them later. The license agreement issue has NOT been tested in court YET, and if it ever does the software industry will probably have to change the way they do EULA. To run Microsoft software you have NO option but to agree in their terms to their contract. Yes, you could (like me) refuse that agreement and run another OS, but to use their product you must agree without any negotiations on your part.
On the last part of your comment, I guess the gun and tobacco industry will start having EVERYONE sign a EULA agreement now.:-)
That isn't the point I was making. In brief here they are. 1. Microsoft should assume some of the responsibility for this isse. The shear number of 'crititcal' patches they release shows the lack of quality of their software and in my opinion they care little about security because it will effect their bottom line if they spend development effort on it.
2. Gartner and company NEVER include all the extra time it takes to constantly patch and reboot NT servers in their reports.
Kinda funny, I am a code monkey, and have written a ton of code. It is also funny that I our company had to accept responsibility for a flaw in some of the code I wrote (had a critical bug). Granted I will NEVER hear the end of it, and was forced to update a production system with just unit testing, but our company did the correct thing and paid the client for the error. Granted that error was very small and only occured in one test case, thank God!
I currently have 4, with each well over 10k lines (granted you can't count lines as actuall work, but at least this tells you it isn't 4 hello worlds). Also with those apps comes the support I have to provide. How many applications do you have in a production environment now?
As for my analogy... I guess your motto should be: Software development, where quality is job.... well we really don't give a rats ass about quality, because there is nothing you can do about it. Kinda sounds like Ford of the 70's.
The same arguments were used by the automotive companies for years on airbags and other safey elements of a car. They use (and probably still do) have bean counters that calculate how many people will die and what the lawsuites will be if they skimp on a design. If the numbers are less than the increase in engineering, some would do it. Now in their case they still get sent to court. Microsoft doesn't. They get to cut corners and well... because of their illegal monopoly use the users as testers.
But to answer your question about why don't I pay more for a "better and more stable" version of Windows??? Well, this virus also effected their "data center edition". That version is high $$$$. Nuff said!
My first point is still true. It cost more to run Microsoft Windows software than Gartner and company report in their TCO because of the ridiculous amount of patches that must be tested and applied. This by no means is bullshit. So when they report that the servers running other NOS'es require the same number of admins as Windows servers they are dead wrong!
You seem to agree with me that Microsoft didn't spend any money on security because people would buy their software without it. Thats kinda like the car companies cutting corners on safety because the customer won't notice that at the sales floor. Well those automotive companies are held responsible for their actions, just like Microsoft should be.
We do agree also that they are mutually exclusive (security, features), however a defective product in either of those items can cause lawsuits to flow.
Come to think of it, mabe that is why they stockpiled so much cash... they new that someday they were going to get sued.
Fist, the EULA is kinda null and void on this issue. Do you EVER get to send a counter contract to Microsoft? Now if you negotiated a site licence with them, that is another matter.
Next using your argument motorcycle helmet, gun, tobacco, and chemical companies have nothing to EVER worry about. Heck I bet the next box of Camel's will have an EULA on it:-)
AS for using Ford as an example, look at the FireStone tires they had on their Explorers. They got sued for that one, and they finally admitted fault, but yet even after they had the recall people still sued them.
Why is Microsoft any different? This was not a trivial security problem, it was a catastrophic one. They hold some of the responsibility and should pay.
I will change it, so you will hopefully understand it better.
What if someone figured out to hack in to the OnStar system by GM and shut every GM car down? At first people would be mad at the person who did it and some would probably sue GM at this point (they may or may not win, my guess is that they would get some money from GM). Now picture if this happened every month or so, and GM said that ALL people who had this OnStar SOFTWARE installed needs to press an update button on their car every so often to load security patches. Now to keep with this analogy, some people that press the button have such major problems that they have to have their car towed back to a service station just to get it running again. This makes people nervous about pressing the button... Then a big attack comes and because a large percentage of people were too lazy or to nervous to press that button, their cars get shut down while driving. Some die and some GM cars kill other people. Now GM would be put out of business so fast it would make the breast implant thing look like a small claims case. This ALL would be because of a software error.
Am I saying that some admins don't share some of the blame? No! But this isn't the first or even the 50th severe security patch that Microsoft has released.
Now I will not keep saying this. Gartner and company MUST start factoring this in to the TOTAL COST OF OWNERSHIP of running Microsoft software. If you need an admin that must do nothing but patch your companies OS (testing included) all the time, then that hard cost MUST be included in TCO!!!!
As for your comment about Microsoft owning the lions share of the market and that is why they are attacked, I disagree. The real issue is that Microsoft has, from the beginning, tried to make their applications easy to integrate. They developed DDE and OLE. Both with NO real security built in. This was great for their users because it enabled Microsoft to get products out faster and made it far easier to work with them (from an API viewpoint). But look at Java Applets VS Active X controls. Java has the lions share of that type of market and yet you don't hear about security problems with it. Now ActiveX.... well lets just say it is an abomination with little to no security design. You got to love the fact that stuff can get installed on your Windows system without your approval over the Net!
Also look at some of their more recent products. Windows95 sent passwords over the network in clear text!!! No other operating system at that time even thought about doing that!
The last issue is that Microsoft is the worlds largest software maker with around 40 BILLION in the bank. What if they took around say 5 BILLION and really focused on their products security. I guess the question is why didn't they do that? The answer is that it would have taken cash off of their bottom line....
Kind of like an automotive company that cuts corners on stuff to save money... but then they get sued if things go bad and people get hurt....
Actually I believe the real reason for this tracking system is soley for a census. The more homeless and "poor" you show in an area the more the Democrats can try and re-district.
I was told just the opposite by two different sales people at RedHat. It isn't clear on the website. If someone from RedHat reads this, please comment.
Needless to say, I hope that you are correct, because then we would not use RedHat 9 for our other servers AND RedHat would get $350 a server as opposed to almost nothing.
Slashdot Mirror
Yep, I bet Sacramento is a real slum. I mean it can't compare with most of the "progressive" cities in the world like Detroit, LA and NY. You probably have a much worse time with drugs, gangs,teen pregnancy and homelessness than any of those cities. I can see why you hate it so much there. Heck, I have been to Sacramento a few times and I feel lucky to get out with my life in hand :-)
It is kind of amusing that people are so worried about offending a very very very small minority of people, so much so that they want to change one of the founding principles that the U.S was built on and is still being built on. All for the fact that someone is offended? Well I guess I am offended at the fact that children can't pray in school anymore. Hmmm now we have two sides of people that are offended. How did the U.S. generally decide such issues? By the laws. Those laws were generally made by their representatives who (mostly) acted on behalf of their people. You will find that almost NO congressman will take this issue on. Why?
Now we have a very small majority of people that hate those decisions and laws. Those people see no way to change the system except through the very liberal court system. (I.E. ACLU) Some of these judges tend to make up laws instead of ruling on existing law, and that is where the problems come in.
The good news is that the majority of people are starting to wake up in the U.S.
Hello fellow hoosier...
I wonder where this guys company is located?
Most of our databases fit in to RAM. I do agree with what you are saying in general, however in our case it isn't true. The two biggest factors that we have is the amount of on die cache and the FSB speed. Now, as mentioned by other posters, the Opteron handles dual processor to RAM far more efficently than a dual Xeon. Also the Xeons that it is being compared against only has 512k of on die cache.
The core question that has to be asked is if you need dual processors at all. If not then a Xeon or new P4 would probably be your best bet, but if you do then the dual Opteron is a great buy for processors. Take this plus the fact that you can easily get one of these boxes with 4 GIG of RAM now, and disk i/o becomes less of a factor. Heck, if you want to throw some more money at it, you can actually get one of these boxes with 12 GIG of RAM (If someone makes 333FSB 2GB DDR chips). With 12 GB or RAM that would hide some serious disk i/o issues.
We have a dual Opteron 1.8GH and a 2GH on order. We found that with RedHat and Oracle the Opteron in 32 bit mode beat the crap out of a quad Xeon for the stuff we do.
Just an F.Y.I. Again this was with 32 bit code. I tried the RedHat BETA and it wouldn't even boot up without locking.
So given that Oracle cost us over $20k a processor, we saved over 40 grand!
From what I see slashdot is nothing more than a news server. Yes it does very simple writes, but mostly reads from a database. Yes it does maintain state for a user, but it is a VERY simple application. It behaves the same way (with preferences) for every user. It also has some significant hardware behind it. I would be willing to bet that 90+% of Slashdot is reads from a DB. That fits perfectly with MySQL, again this is not indicitive of most businiess web apps. You mentioned 8 web servers.
I would like to know the people that check their connection speed with slashdot, because it's performance varies greatly throughout the day, and they should be informed that there are far better ways to test their connection speed.
Back to the original posted comment.....
Sun has given a lot to the open source community. They have continued to develop OpenOffice, GNOME and other packages. I see this as a huge win for open-source fans. Lets say sun somehow gets 5-15% marketshare for this product. It will then make other 3rd party developers start developing drivers and applications for their crap to work with Linux. Then people like me could more easily suggest that we replace some of our business with an open source model.
I could see call centers loving this type of product. Specifically Cendant call centers.
Well... I am a AMD fan as well, and just purchased all new AMD systems for our x86 users at work, so this is somewhat difficult for me to write, but if AMD's bet on 64bit technology doesn't pay off, and Windows doesn't support it well, then AMD will be dead. Their financial report has sucked the last year, and they cannot afford to make the mistakes that Intel has made. Intel, like Microsoft can afford to make some very stupid mistakes (Rambus, Itanic ect..) and still recover, but AMD cannot. I hope and pray that the Opteron does great for servers and desktops, but I won't be shocked if AMD is out of business in two years.
Now for PPC. The PPC has done very well in routers and other non PC stuff. It has kinda sucked in PC's in comparision. Motorolla (another company I like, from my Amiga days), has decided to focus off the PC market and on other segments. They are just going to focus more on battery life and other things, than all out power. This left IBM and Apple to take up the lions share of development and fabs for PPC based PC chips. Luckly for Apple, IBM has decided to do it for now. I hope that IBM continues to produce great desktop and server PPC chips for years to come, but the development dollars that it is costing them must be returned some day, and Apples marketshare + AIX systems don't seem to add up to the cost of R&D + fabs for them. Take that and add that Sony and IBM are working on this "Cell" technology, and it appears that the PPC chips days may be numbered. So if you were Apple and IBM was moving to "Cell" chips AND you had to migrate your system again, what would you migrate to? Cell or X86? Or would you decide to take over the fabs, and continue the chip production and design yourself?
Man, I am about one of the biggest Apple fans around. Just because I say Apple may switch to Intel or any x86 chip doesn't make me a troll.
Some of the developers I know, have said that Apple has working hardware on x86 already. They would be foolish not to be prepared to switch, if the PPC chip can't compete. Having said that, I am also a fan of the PPC chip. I hope that it beats the crap out of Intel and AMD, but I just don't see that happening. It is hard to compete with 2 BILLION in R&D EVERY year.
I (like many others) would still buy a mac, with an X86 chip in it. What I pay for is the integration of hardware and software (kinda like Sun/Solaris). Their stuff just works. As long as they continue that mode of operation, I will be a fan.
Ok, I have a Athlon 3000+ and a new G5 Mac, and also have a dual processor Opteron system. The previous poster is correct. The heat syncs on the new G5 are at least 5X as big as any heat sync I have EVER seen on a Intel/AMD system. I honestly don't know what the processor temprature is when it runs, but the noise level of this new G5 is as loud as any modern PC that I have seen. This is a 1.8GH single processor Mac system.
From what I have heard this the current G4 at 1.1 GH takes around 14 watts of power, and the current G5's take more than double that amount at 2GH. Some believe that they could release a G5 based system running at 1.4GH and get the power down to around 16 watts, but why? Speed isn't what people are screaming for with laptops and you would loose more battery life.
I personally think it will be at least a year and a half before you will see a G5 laptop. I think you will see Apple running on Intel (Itanium) or AMD(Opteron) first.
My overall point stands true:
It cost far more time and money to take "solid" manufactured goods and bring them in to the U.S.
Some of those goods are taxed, some at high levels.
On a side note:
NAFTA has been horrible for the U.S. I still consider myself a conservative, but with idiots like Rush out there saying how great this was, and that smoking isn't bad for you.... Well it just sounds rediculous. There was a giant sucking sound of manufacturing jobs to Mexico, and the U.S. is not any better for it.
Notice that you said potentially.
Put that job on Monster and see how many resume's you get of people that are out of work. It will shock you.
I live in the midwest, and I can tell you that it is a bloodbath around here. Multiple startups have died, almost every software development company has folded, and there are only a few companies that still have I.T. shops internal. So yes, some smaller companies are hireing one person, but when 4 or 5 HUGE companies cut their entire I.T. staff (hundreds), that sucks bad.
Most companies that I know are "getting by" and they need to hire someone else, but actually doing it is another thing...
Just to give you an example of what I know.
Three years ago we had to pay over $100/hour for Java developers. We now pay $20-$25 and have little trouble getting them.
We are getting ready to start looking for a partime Oracle DBA, and I know that we will be able to get one for a very low amount of money, again we are talking in the $20-$25 range.
I feel bad even offering this kind of money for those skills, but as most of these people put it
"I make nothing now, so this is better than that".
I will admit that we do still pay $100+/hour for "specific" needs, like a "4 hour" network job, but those are VERY rare.
There is a HUGE difference between manufactured goods and software development. It takes effort and money to bring goods in to the U.S. and those goods are generally taxed. Software development isn't, and it takes little effort to "move" code.
The other core difference is that the other jobs took more than 10 years to move offshore, this has taken around 2.
I do find it ironic that hardly ANY open source development gets done by Indian programmers though...
Also, this will just speed up the use of unions for those remaining I.T. workers here in the U.S. Most computer science people are conservatives by nature , and I look for this one issue to drive a large percentage to vote against the Republican party in the next election.
Man if only they were mine... They are our companies. But at least I get to use them.
:-(
The BETA of RedHat ES 3.0 for the Opteron won't load. It locks after reboot.
Yes. It came out a few weeks ago. You have to download it.
I am currently loading RedHat ES 3.x BETA and hope to try it out.
I don't expect it to work with the BETA of RedHat, but I want to test it out.
Oh yeah, the RedHat version is also for the Opteron.
I started the load last night...
Well I have had my dual Opteron system for about three weeks now, AND I ordered a G5 over a month ago.... The G5 still hasn't shipped to me (I ordered the 1.8GH model). Even if it did the Opteron was out more than a month ago.
Now granted the article should say something like, this is the first somewhat affordable x86 chip to provide 64 bit computing.
There are many other links that have excellent benchmark results of the Opteron. Specifically, toms hardware has a great review, but to sum most of the review up...
The dual Opteron 244 kicks the crap out of a dual Xeon in most linux server type benchmarks. This is WITH 32 bit code. The next step is to get code that is still 32 bit but compiled for the Opteron. i.e. Oracle wich is out now. This gives you another small gain in performance and then lastly you can test again with pure 64 bit compiled programs. The last "MAY" give you some performance increase over 32 bit programs recompiled, but that depends on the program. An example would be a simple "hello world" would not run any faster, but a database that needs more than 4GB of memory should perform better.
So for our company we didn't compare the Opteron to a Itanium, but to a Xeon and the dual Opteron smokes it.
Note, that I always say dual Opteron, because a single Opteron doesn't perform any better than the current 32bit X86 chips. In most cases it performs worse. One of the key factors with the Opteron is the way the chips talk to the front side bus. This is currently the bottleneck in most dual/quad processor x86 systems. AMD has somewhat handled this bottleneck by allowing both chips to have a separate path to "some" RAM and providing a larger amount of data to pass from processor to RAM. I can't speak for the quad and eight way systems, but for the dual processor systems this makes a HUGE difference.
In our environment, a dual Opteron 1.8GH beats a QUAD 2.8GH Xeon in performance (with Oracle). Given that Oracle charges us 15k per CPU + 4k a year, that is a huge savings, and we are not forced to go to a pure 64 bit system today.
Intel addresses the front side bus issue to RAM by increasing the on die cache of the Xeon. So they released a 1MB version of the Xeon 3GH (they had a 512MB version before), to combat the Opteron. The problem is that it didn't help that much. Intel's solution would be to include X86 code, "done well, not the crappy way it is done now" in the Itanium, and lower the price to under a grand. I doubt they will do that any time soon, so it looks like this will give AMD a good path to the small server market. Factor that in with IBM and SUN wanting to use their stuff in new server, things look ok for the Opteron.
I still don't know how well the AMD Athlon 64 will be though. That chip isn't due to be released until next month. That is the replacement for the current Athlons, and for AMD's sake it better beat a P4 in most Windows stuff.
Well, this has been dead for a while, but I just saw it. Notice that I said "at that time" Telnet and FTP where developed WAY before 1995. Their lack of security can be somewhat forgiven because at that time people had ssh and secure ftp if they wanted to use it.
Many people sign wavers not to sue a company and legally get to sue them later. The license agreement issue has NOT been tested in court YET, and if it ever does the software industry will probably have to change the way they do EULA. To run Microsoft software you have NO option but to agree in their terms to their contract. Yes, you could (like me) refuse that agreement and run another OS, but to use their product you must agree without any negotiations on your part.
:-)
On the last part of your comment, I guess the gun and tobacco industry will start having EVERYONE sign a EULA agreement now.
That isn't the point I was making. In brief here they are.
1. Microsoft should assume some of the responsibility for this isse. The shear number of 'crititcal' patches they release shows the lack of quality of their software and in my opinion they care little about security because it will effect their bottom line if they spend development effort on it.
2. Gartner and company NEVER include all the extra time it takes to constantly patch and reboot NT servers in their reports.
Kinda funny, I am a code monkey, and have written a ton of code. It is also funny that I our company had to accept responsibility for a flaw in some of the code I wrote (had a critical bug). Granted I will NEVER hear the end of it, and was forced to update a production system with just unit testing, but our company did the correct thing and paid the client for the error. Granted that error was very small and only occured in one test case, thank God!
I currently have 4, with each well over 10k lines (granted you can't count lines as actuall work, but at least this tells you it isn't 4 hello worlds). Also with those apps comes the support I have to provide. How many applications do you have in a production environment now?
As for my analogy...
I guess your motto should be:
Software development, where quality is job.... well we really don't give a rats ass about quality, because there is nothing you can do about it. Kinda sounds like Ford of the 70's.
The same arguments were used by the automotive companies for years on airbags and other safey elements of a car. They use (and probably still do) have bean counters that calculate how many people will die and what the lawsuites will be if they skimp on a design. If the numbers are less than the increase in engineering, some would do it. Now in their case they still get sent to court. Microsoft doesn't. They get to cut corners and well... because of their illegal monopoly use the users as testers.
But to answer your question about why don't I pay more for a "better and more stable" version of Windows??? Well, this virus also effected their "data center edition". That version is high $$$$. Nuff said!
By the way, we don't use Windows servers.
My first point is still true. It cost more to run Microsoft Windows software than Gartner and company report in their TCO because of the ridiculous amount of patches that must be tested and applied. This by no means is bullshit. So when they report that the servers running other NOS'es require the same number of admins as Windows servers they are dead wrong!
You seem to agree with me that Microsoft didn't spend any money on security because people would buy their software without it. Thats kinda like the car companies cutting corners on safety because the customer won't notice that at the sales floor. Well those automotive companies are held responsible for their actions, just like Microsoft should be.
We do agree also that they are mutually exclusive (security, features), however a defective product in either of those items can cause lawsuits to flow.
Come to think of it, mabe that is why they stockpiled so much cash... they new that someday they were going to get sued.
Fist, the EULA is kinda null and void on this issue. Do you EVER get to send a counter contract to Microsoft? Now if you negotiated a site licence with them, that is another matter.
:-)
Next using your argument motorcycle helmet, gun, tobacco, and chemical companies have nothing to EVER worry about. Heck I bet the next box of Camel's will have an EULA on it
AS for using Ford as an example, look at the FireStone tires they had on their Explorers. They got sued for that one, and they finally admitted fault, but yet even after they had the recall people still sued them.
Why is Microsoft any different? This was not a trivial security problem, it was a catastrophic one. They hold some of the responsibility and should pay.
His example isn't void.
I will change it, so you will hopefully understand it better.
What if someone figured out to hack in to the OnStar system by GM and shut every GM car down? At first people would be mad at the person who did it and some would probably sue GM at this point (they may or may not win, my guess is that they would get some money from GM). Now picture if this happened every month or so, and GM said that ALL people who had this OnStar SOFTWARE installed needs to press an update button on their car every so often to load security patches. Now to keep with this analogy, some people that press the button have such major problems that they have to have their car towed back to a service station just to get it running again. This makes people nervous about pressing the button... Then a big attack comes and because a large percentage of people were too lazy or to nervous to press that button, their cars get shut down while driving. Some die and some GM cars kill other people. Now GM would be put out of business so fast it would make the breast implant thing look like a small claims case. This ALL would be because of a software error.
Am I saying that some admins don't share some of the blame? No! But this isn't the first or even the 50th severe security patch that Microsoft has released.
Now I will not keep saying this.
Gartner and company MUST start factoring this in to the TOTAL COST OF OWNERSHIP of running Microsoft software. If you need an admin that must do nothing but patch your companies OS (testing included) all the time, then that hard cost MUST be included in TCO!!!!
As for your comment about Microsoft owning the lions share of the market and that is why they are attacked, I disagree. The real issue is that Microsoft has, from the beginning, tried to make their applications easy to integrate. They developed DDE and OLE. Both with NO real security built in. This was great for their users because it enabled Microsoft to get products out faster and made it far easier to work with them (from an API viewpoint). But look at Java Applets VS Active X controls. Java has the lions share of that type of market and yet you don't hear about security problems with it. Now ActiveX.... well lets just say it is an abomination with little to no security design. You got to love the fact that stuff can get installed on your Windows system without your approval over the Net!
Also look at some of their more recent products. Windows95 sent passwords over the network in clear text!!! No other operating system at that time even thought about doing that!
The last issue is that Microsoft is the worlds largest software maker with around 40 BILLION in the bank. What if they took around say 5 BILLION and really focused on their products security. I guess the question is why didn't they do that? The answer is that it would have taken cash off of their bottom line....
Kind of like an automotive company that cuts corners on stuff to save money... but then they get sued if things go bad and people get hurt....
According to their sales reps this is not accurate. I was shocked!
Again, if someone from RedHat knows please answer.
Actually I believe the real reason for this tracking system is soley for a census. The more homeless and "poor" you show in an area the more the Democrats can try and re-district.
I was told just the opposite by two different sales people at RedHat. It isn't clear on the website. If someone from RedHat reads this, please comment.
Needless to say, I hope that you are correct, because then we would not use RedHat 9 for our other servers AND RedHat would get $350 a server as opposed to almost nothing.