I'm regularly shocked by what I hear here on slashdot about US university CS education. I took a BSc in CS at Durham University in England, which is certainly not considered one of the best CS departments, but we took nothing but theory. Even the OO course expected us to teach ourselves C++ - we were taught the underlying concepts. For our databases course, SQL was the last thing we learned - first we were taught how to implement a database, and the history of it, and how query optimisers work. We were taught how to write a compiler, how to write an OS, we read books that are now 30 years old, but still relevant today... I still pick up every now and then some of those books - "Fundamentals of Database Systems" and "Principals of Concurrent and Distributed Programming" - the issues remain relevant in languages like Java and now Perl.
The times change but the underlying implementation is just 1's and 0's...
Wow... I just hope I don't end up educating my children in the US... especially given the price you pay for your University education.
perl -e 'print scalar reverse q(\)-:,hacker Perl another Just)'
Wrapping variables up in other objects to implement a pseudo pass by reference has to be the single most irritating thing about Java ever... Especially when working with strings.
And java's string classes are a) crap, b) confusing and c) slow.
I went right back to Perl... faster than most languages for complex string processing, and fast enough for most other tasks.
Matt.
perl -e 'print scalar reverse q(\)-:,hacker Perl another Just)'
Perhaps you need to read up some on RAID - RAID1 is the fastest and safest of all the RAID standards. It's what you use if cost is not an issue. For a server, that tends to be the case.
For some good introductory docs on RAID try the Software-RAID mini HOWTO.
Matt.
perl -e 'print scalar reverse q(\)-:,hacker Perl another Just)'
If you're looking for optimal performance you have to a) know what you're doing. and b) compile the packages yourself, including only the options you need.
I also don't know where you got your ideas about RAID. RAID 1 is what you want for higher performance on a web server. RAID 5 shows little thought went into the choice.
perl -e 'print scalar reverse q(\)-:,hacker Perl another Just)'
If you're looking for optimal performance you have to a) know what you're doing. and b) compile the packages yourself, including only the options you need.
I also don't know where you got your ideas about RAID. RAID 1 is what you want for higher performance on a web server. RAID 5 shows little thought went into the choice.
perl -e 'print scalar reverse q(\)-:,hacker Perl another Just)'
Microsoft made a competing product more successful - not because it was better, but because they didn't want people to use a standard that would allow them to port their apps to other platforms. That's all I'm saying - I think this was good for Microsoft and bad for Corba.
But the victims _are_ doing something about it. I was just lamenting that it would have been nice if we could interoperate with Win32 easily, like we could have if they'd adopted Corba.
It's a real shame that Corba didn't take off in the way it should have. Yes I know it's used in a lot of high end systems - and it performs very very well in those systems. Unfortunately only 1 thing contributed to its failure: Microsoft. Sad really. Why MS felt that they had to provide a proprietary system is known - they had to get vendor lock in. That attitude sucks. Corba is a more powerful and easier to use object model than COM. Developing for COM is not nearly as easy.
I suppose it'll end up in the great bit-bucket of marginally used but superior technologies...
This is about the 250th comment on this subject, so most people probably won't even read it. But there's something missing...
Here's all these geeks - hardcore geeks, defending the distribution of movies - hundreds of megabytes of movies - over the internet and internal university networks.
Well guess what you selfish bastards: I pay for that network. Not you the student - you'll pay later - right now it's the working people paying for your fun through our taxes.
And here's these hardcore geeks, that complain about any extra bandwidth, like HTML email, or spam. And yet they don't even blink about clogging up the network with the equivalent of a million spams. A MILLION SPAMS is what your movie costs the rest of your country. Thank you very much - whoever you are - you're the lowest of the low.
Prior to going public, Yahoo certainly didn't have a commercial focus. It certainly does now. When you do a search for cars it puts up ads about cars. I'm finding it difficult to think of an exact parallel for Red Hat there, but I wouldn't want that level of commercialisation for RH Linux.
Something like 0.1% of all injuries by gun use are from people defending themselves.
i.e. people having access to guns isn't helping anyone except the criminals.
That DOES happens, but don't ask the liberal media
on
Why Kids Kill
·
· Score: 1
Is there someplace we can find these numbers?
I doubt it - they look made up to me. 500 school drive by incidents of mass knifings - that's over 5,000 school kids killed by drive by knifings. Bwahahahaha.... Don't make me laugh.
If you read what I wrote you'll realise I did drop Apache and recompile - I just used RPM's. I don't really think the problem is the compiler used for Apache - I think most of the work is in the kernel, managing processes and file caching. I guess a pgcc compiled kernel would be a better option - but that's not something I'm desperate to get into - this system took me 1/2 a day to build - I don't have the time to really increase that.
A better option would be to just use Apache for mod_perl processes, and use thttpd for static content, and use squid to proxy requests to the right port. I think we'd probably blow away even our own estimations with thttpd.
While our box is not SMP, I very much doubt that SMP on 4 processors (which even 2.0.36 was known to cope quite well with, especially for simple tasks like web serving) would cripple a box to the extent provided by the mindcraft survey.
I can see an SMP system to only improve performance up to a point, eventually hitting a limit (e.g. 8 processors won't be twice as good as 4 because of time spent scheduling), but to see the effects that mindcraft saw you would have to do something pretty crafty...
Please note that the dejanews reference that ESR links to is quite wrong. The presence of %h does _not_ cause host name lookups under Apache - only the directive "HostNameLookups on" causes that to occur. I don't believe this to be the case.
I strongly believe however that their httpd was running under inetd, and that would cause the effect they saw.
You're only right on the kernel front - that hasn't really been used enough on high end machines AFAIK.
But both Samba and Apache have been heavily tested on very high end servers. The Samba crew have even been heavily involved in making Samba fast on high end servers.
At a large company I'm working with we're trying to prove to the phb's that Linux is a good thing. The mindcraft study set us back a ways. So what did we do? We did our own tests.
Server: - Hand built by our best hardware guy - PIII 500 (single CPU) - Adaptec 2940U2W SCSI Adapter - 10,000 rpm LRW drive. 1 drive only. - 100Mb/s network card - 256Mb PC100 RAM. - Linux 2.2.6, upgraded from stock Linux-Mandrake box - Apache 1.3.6, configured for best performance.
No changes to the/proc fs to speed things up. Stock kernel options selected from "make xconfig". Apache was the apache+mod_perl srpm found on redhat/contrib, compiled with no configuration changes. We didn't test NT on this box - we were trying to compare against Mindcraft's results.
Want to know the results so far?
Well, we can get about 2200 requests per second out of that box. The Quad Xeon NT box that mindcraft tested got 3700 requests per second at its maximum rate. We are at very early stages so far, and I think I can squeeze more out of the box by dumping Apache and using thttpd or something else that uses a threaded model. But since this is to be a pure mod_perl box I don't think that's important.
Things to remember:
The mindcraft server had 1Gb of RAM. The mindcraft server had RAID (RAID/0 I believe). The mindcraft server had 4 10/100 network cards.
We're so far pretty pleased with our little Linux box... It was a fair bit cheaper than Mindcraft's server....
So should weapons of mass desctruction, and smart bombs be made available in your grocery store too? The "people kill people" argument is the gun lobby's oldest argument. IMHO it's a dumb one, because it means you don't have to stop at guns. Why not make it legal to have 200,000 volt electric wire fencing around your house too so that you can "protect yourself". Please.
The only reason you need to protect yourself in the first place is because you made guns available. I don't know why you don't see the error in that. It's only the US that has this problem of children slaying children with guns.
Articles about Linux are now quoting the DH Brown report, just because it's the first one of its kind. How soon before they start quoting the Mindcrafty study? I bet not long.
If Red Hat or VAR don't do some serious Linux benchmarking soon, and prove its SMP capabilities, then Linux will simply get this information perpetuated. Hurry up with those benchmarks!!!
(Oh, and someone needs to write some performance tuning docs, including SMP + Large memory info...)
"Some people apparently want to do IRC in their web browser, allthough I can't understand why..."
Because it's a whole lot easier to tell my mother-in-law or father-in-law or wife to "Just click on that link on my home page" - like they can do with IE4 (loads up Comic-Chat - ugh!). Personally I'll stick with x-chat thought...
The problem I see with this is that in general Linux people don't want to do a benchmark that will "give them the results they want to see" (It says something like that on Mindcraft's web page). We want to see real hard evidence. We want to know what the exact parameters are to set up a fast server. We want to know if an NT box can be beaten even when tweaked to the max. This isn't something most testing companies will have the time or money to do - and if they do, the report will cost you - big time. The iron alone would be a hellish price - 50 PCs + a quad P3 is going to set you back a hell of a lot.
Take for example the web server benchmarks. No Linux sysadmin is stupid enough (I hope) to just go by static HTML results - the key to benchmarking results is to see results for dynamic page implementations, since they are where your server slows down. For some benchmarking started in that area, see http://perl.apache.org, where we started to benchmark mod_perl vs ASP vs CGI and some other stuff. Of course these tests need more work too.
Slashdot Mirror
I'm regularly shocked by what I hear here on slashdot about US university CS education. I took a BSc in CS at Durham University in England, which is certainly not considered one of the best CS departments, but we took nothing but theory. Even the OO course expected us to teach ourselves C++ - we were taught the underlying concepts. For our databases course, SQL was the last thing we learned - first we were taught how to implement a database, and the history of it, and how query optimisers work. We were taught how to write a compiler, how to write an OS, we read books that are now 30 years old, but still relevant today... I still pick up every now and then some of those books - "Fundamentals of Database Systems" and "Principals of Concurrent and Distributed Programming" - the issues remain relevant in languages like Java and now Perl.
,hacker Perl another Just)'
The times change but the underlying implementation is just 1's and 0's...
Wow... I just hope I don't end up educating my children in the US... especially given the price you pay for your University education.
perl -e 'print scalar reverse q(\)-:
No pass by reference!!!
,hacker Perl another Just)'
Wrapping variables up in other objects to implement a pseudo pass by reference has to be the single most irritating thing about Java ever... Especially when working with strings.
And java's string classes are a) crap, b) confusing and c) slow.
I went right back to Perl... faster than most languages for complex string processing, and fast enough for most other tasks.
Matt.
perl -e 'print scalar reverse q(\)-:
I think they were aiming with telescopic sights... We managed to get twice the performance they did out of a SINGLE PIII500...
,hacker Perl another Just)'
perl -e 'print scalar reverse q(\)-:
I'd run RAID1 on a real server - in a second.
,hacker Perl another Just)'
Perhaps you need to read up some on RAID - RAID1 is the fastest and safest of all the RAID standards. It's what you use if cost is not an issue. For a server, that tends to be the case.
For some good introductory docs on RAID try the Software-RAID mini HOWTO.
Matt.
perl -e 'print scalar reverse q(\)-:
Not really. They weren't really being as productive as they could have been, whereas the Perl Mongers have been extremely successful.
,hacker Perl another Just)'
I just hope they keep the news page going - it was the only thing I ever got out of TPI.
perl -e 'print scalar reverse q(\)-:
I don't know where you got your ideas from.
,hacker Perl another Just)'
If you're looking for optimal performance you have to a) know what you're doing. and b) compile the packages yourself, including only the options you need.
I also don't know where you got your ideas about RAID. RAID 1 is what you want for higher performance on a web server. RAID 5 shows little thought went into the choice.
perl -e 'print scalar reverse q(\)-:
I don't know where you got your ideas from.
,hacker Perl another Just)'
If you're looking for optimal performance you have to a) know what you're doing. and b) compile the packages yourself, including only the options you need.
I also don't know where you got your ideas about RAID. RAID 1 is what you want for higher performance on a web server. RAID 5 shows little thought went into the choice.
perl -e 'print scalar reverse q(\)-:
Corba _has_ taken off on Unix. I was lamenting that it hasn't taken off on 99% of the worlds computers...
Microsoft made a competing product more successful - not because it was better, but because they didn't want people to use a standard that would allow them to port their apps to other platforms. That's all I'm saying - I think this was good for Microsoft and bad for Corba.
But the victims _are_ doing something about it. I was just lamenting that it would have been nice if we could interoperate with Win32 easily, like we could have if they'd adopted Corba.
It's a real shame that Corba didn't take off in the way it should have. Yes I know it's used in a lot of high end systems - and it performs very very well in those systems. Unfortunately only 1 thing contributed to its failure: Microsoft. Sad really. Why MS felt that they had to provide a proprietary system is known - they had to get vendor lock in. That attitude sucks. Corba is a more powerful and easier to use object model than COM. Developing for COM is not nearly as easy.
I suppose it'll end up in the great bit-bucket of marginally used but superior technologies...
Those networks are put there for academia you lowlife. I suppose you think it's really "cool" to do what you do. Wise up and grow up.
This is about the 250th comment on this subject, so most people probably won't even read it. But there's something missing...
Here's all these geeks - hardcore geeks, defending the distribution of movies - hundreds of megabytes of movies - over the internet and internal university networks.
Well guess what you selfish bastards: I pay for that network. Not you the student - you'll pay later - right now it's the working people paying for your fun through our taxes.
And here's these hardcore geeks, that complain about any extra bandwidth, like HTML email, or spam. And yet they don't even blink about clogging up the network with the equivalent of a million spams. A MILLION SPAMS is what your movie costs the rest of your country. Thank you very much - whoever you are - you're the lowest of the low.
Matt.
Prior to going public, Yahoo certainly didn't have a commercial focus. It certainly does now. When you do a search for cars it puts up ads about cars. I'm finding it difficult to think of an exact parallel for Red Hat there, but I wouldn't want that level of commercialisation for RH Linux.
Something like 0.1% of all injuries by gun use are from people defending themselves.
i.e. people having access to guns isn't helping anyone except the criminals.
Is there someplace we can find these numbers?
I doubt it - they look made up to me. 500 school drive by incidents of mass knifings - that's over 5,000 school kids killed by drive by knifings. Bwahahahaha.... Don't make me laugh.
I think the total cost was ~£3500 UK.
If you read what I wrote you'll realise I did drop Apache and recompile - I just used RPM's. I don't really think the problem is the compiler used for Apache - I think most of the work is in the kernel, managing processes and file caching. I guess a pgcc compiled kernel would be a better option - but that's not something I'm desperate to get into - this system took me 1/2 a day to build - I don't have the time to really increase that.
A better option would be to just use Apache for mod_perl processes, and use thttpd for static content, and use squid to proxy requests to the right port. I think we'd probably blow away even our own estimations with thttpd.
Matt.
While our box is not SMP, I very much doubt that SMP on 4 processors (which even 2.0.36 was known to cope quite well with, especially for simple tasks like web serving) would cripple a box to the extent provided by the mindcraft survey.
I can see an SMP system to only improve performance up to a point, eventually hitting a limit (e.g. 8 processors won't be twice as good as 4 because of time spent scheduling), but to see the effects that mindcraft saw you would have to do something pretty crafty...
Matt.
Please note that the dejanews reference that ESR links to is quite wrong. The presence of %h does _not_ cause host name lookups under Apache - only the directive "HostNameLookups on" causes that to occur. I don't believe this to be the case.
I strongly believe however that their httpd was running under inetd, and that would cause the effect they saw.
You're only right on the kernel front - that hasn't really been used enough on high end machines AFAIK.
But both Samba and Apache have been heavily tested on very high end servers. The Samba crew have even been heavily involved in making Samba fast on high end servers.
At a large company I'm working with we're trying to prove to the phb's that Linux is a good thing. The mindcraft study set us back a ways. So what did we do? We did our own tests.
/proc fs to speed things up. Stock kernel options selected from "make xconfig". Apache was the apache+mod_perl srpm found on redhat/contrib, compiled with no configuration changes. We didn't test NT on this box - we were trying to compare against Mindcraft's results.
Server:
- Hand built by our best hardware guy
- PIII 500 (single CPU)
- Adaptec 2940U2W SCSI Adapter
- 10,000 rpm LRW drive. 1 drive only.
- 100Mb/s network card
- 256Mb PC100 RAM.
- Linux 2.2.6, upgraded from stock Linux-Mandrake box
- Apache 1.3.6, configured for best performance.
No changes to the
Want to know the results so far?
Well, we can get about 2200 requests per second out of that box. The Quad Xeon NT box that mindcraft tested got 3700 requests per second at its maximum rate. We are at very early stages so far, and I think I can squeeze more out of the box by dumping Apache and using thttpd or something else that uses a threaded model. But since this is to be a pure mod_perl box I don't think that's important.
Things to remember:
The mindcraft server had 1Gb of RAM.
The mindcraft server had RAID (RAID/0 I believe).
The mindcraft server had 4 10/100 network cards.
We're so far pretty pleased with our little Linux box... It was a fair bit cheaper than Mindcraft's server....
So should weapons of mass desctruction, and smart bombs be made available in your grocery store too? The "people kill people" argument is the gun lobby's oldest argument. IMHO it's a dumb one, because it means you don't have to stop at guns. Why not make it legal to have 200,000 volt electric wire fencing around your house too so that you can "protect yourself". Please.
The only reason you need to protect yourself in the first place is because you made guns available. I don't know why you don't see the error in that. It's only the US that has this problem of children slaying children with guns.
Matt.
Articles about Linux are now quoting the DH Brown report, just because it's the first one of its kind. How soon before they start quoting the Mindcrafty study? I bet not long.
If Red Hat or VAR don't do some serious Linux benchmarking soon, and prove its SMP capabilities, then Linux will simply get this information perpetuated. Hurry up with those benchmarks!!!
(Oh, and someone needs to write some performance tuning docs, including SMP + Large memory info...)
Matt
Because it's a whole lot easier to tell my mother-in-law or father-in-law or wife to "Just click on that link on my home page" - like they can do with IE4 (loads up Comic-Chat - ugh!). Personally I'll stick with x-chat thought...
Matt.
The problem I see with this is that in general Linux people don't want to do a benchmark that will "give them the results they want to see" (It says something like that on Mindcraft's web page). We want to see real hard evidence. We want to know what the exact parameters are to set up a fast server. We want to know if an NT box can be beaten even when tweaked to the max. This isn't something most testing companies will have the time or money to do - and if they do, the report will cost you - big time. The iron alone would be a hellish price - 50 PCs + a quad P3 is going to set you back a hell of a lot.
Take for example the web server benchmarks. No Linux sysadmin is stupid enough (I hope) to just go by static HTML results - the key to benchmarking results is to see results for dynamic page implementations, since they are where your server slows down. For some benchmarking started in that area, see http://perl.apache.org, where we started to benchmark mod_perl vs ASP vs CGI and some other stuff. Of course these tests need more work too.
Matt.
It was "Javascript:" - not javascript. Try typing javascript: into the url box and you get a javascript console.