When I was starting my own company a few years back I applied for sponsorship to get it going. As part of the interview process they asked applicants to do a 20 minute presentation for the selection board. I wanted the board to understand what I was doing and to realise I knew my stuff so I produced precisely four slides. They were black and white, plain simple text consisting of a single sentence which served as a banner to each of the points I was making along the lines of "what is it?", "who uses it?", "will people buy it?" and "how much?" I simply stated my case for each of these and then opened for questions.
I got the money.
Often presentations are used as a crutch by people who don't really know what they are talking about and if you get them off track they lose their way which is why they prefer to write everything they want to say down and read from it. I have never believed that this approach will be successful in engaging with an audience and it looks like I was right.
"- create a folder named FordChevyDodge. Search for 'Chevy'--it pops right up. Search for 'hevy'--nada. Oops. "
Its the capitalisation. Change the file to FordCHevyDodge and search for 'hevy' and it shows up just fine. I'll admit that it is odd though. Maybe they are assuming that if a word starts with a capital then that is a proper name and you wouldn't want to find it if you missed the first letter. Possibly it helps the indexing by reducing size. Or something.:-)
Nice one though, I've never run across the problem before and I use spotlight all the time.
Yeah, I was surprised. I had assumed that since 1080p worked through component there would be no problem with HDMI. However, I have an upscaling DVD player with HDMI and it supports 1080p output but whenever I try to select 1080p it switches back to 1080i. Looking in the specs of my projector (InFocus In76) it does say that it only supports 1080i on HDMI but 1080p through DVI and component. I'm actually not bothered that much since I can't see any difference between 1080i and 1080p over component as far as resolution goes although with long cables there is some ghosting on the 1080p which is absent with 1080i so that is what I use. Too many people are fixated on the 'p' part and don't realise that 1080 is 1080. The only time the interlacing would have any impact is if input source was aquired that way. If the source is progressive but transmitted as 1080i and then deinterlaced at the other end it doesn't look any different to progressive all the way since the deinterlacing is perfect. With the Xbox HDDVD drive it forces 1080i over component but the disc is progressive so this isn't a handicap at all and I don't know why they insisted on it. Finally, although my projector has a native resolution of 1280x720 it looks better when driven with the 1080i output from the Xbox because the scaler in the projector does a better job of preserving the detail on HDDVD than the Xbox itself does. With my upscaling DVD player though, I run that at 1280x720 which means there is only one scaler involved. Running at 1080i would mean it is scaling up and then the projector is scaling down again and in my own tests this resulted in significant image degradation.
Oh, and HDDVD blows even the best DVDs out of the water despite the fact that I am still only actually using about half the available resolution on the disc. Don't let anyone tell you that an upscaling DVD player is anything like as good as HDDVD or BluRay. Maybe on a small TV (32" or so) you might not be able to tell the difference but on my 70" front projector the difference is very apparent.
I have a 360 with HD-DVD drive and wireless adapter and a pair of wireless pads. If I was to switch to this just for the sake of HDMI and a bigger drive, bearing in mind that my HDTV doesn't support 1080p over HDMI, only over component and also I can't even see a difference between component and HDMI on the thing, I would also have to replace all the white extras. Oh, and I actually think the 360 looks better in white. Mind you, when I came to buy a new iPod Video I bought the white one because I didn't like the look of the black one.
"MS represents itself as a software company, where you generally get 90% margins. "
Where MS generally gets 90% margins. These profits only exist in a monopoly market. If the market place was really competitive MS would be making 30% at best because most of their profits would be rolled back into development costs. MS doesn't spend anything like 60% of profit on R&D, especially not on the cash cows (Windows and Office). Yes, they do burn a lot of money trying to buy their way into other markets but those are all costing them lots. In real terms, MS has raise the cost of their software substantially over the years in order to fund this profligate spending to try and dominate the entire industry. In reality, MS should be made to charge far less for their software and this would have the effect of reining in their dominance. Of course, like everyone else I am not going to hold my breath on that one. All I can do is avoid MS software where possible and pay them as little as possible where I can't.
Don't these schools typically have Windows site licences anyway? By buying Macs and installing Boot Camp the users get the choice they want and the school can install a copy of Windows as usual due to their site license. MS doesn't get paid twice for each PC which can only be good IMHO. Personally, I wouldn't give the users the choice. I would do a deal with parallels or VMware to allow the Windows software to be run while encouraging the users to run OS X all the time.
"What are they suggesting, that they don't want their prospective employees to be quick-thinking and efficient when it comes to menial and repetitive tasks?"
Unfortunately I think that is exactly the situation. The sort of people who need to do these tests have such an amazingly low capacity to actually function with software like Office that anyone who can use it in the way that is intended is probably far too high up the evolutionary tree to actually want to work in a job where they have to use Office all day long. To many of these people, Office *IS* the computer. The people doing the training also have such a low ability that they can't cope with people who know more about it than they do and this even goes for the teachers in schools too. What is holding computing and computer use back is the way people are trained (poorly) to use (poor) software (poorly) to do their jobs (poorly). I have been in several jobs where I was able to complete the entire days work in the first hour simply because I can use a computer efficiently. The training industry doesn't want people who can sit at a computer they have never used before and function quickly. And the support industry doesn't want a computer that doesn't break constantly. And the managers are too deluded to know that the training and support that people are getting is so second rate that they could actually afford to hire 1/8th of the people if they actually had a reasonable level of training and skill.
A few years back I was between proper jobs and had to do some temping work. The temping agency asked if I could use Office. I didn't know as I had only used OpenOffice as I had been a Linux user since 1994 and before that I was on SunOS. Anyway, I knew OpenOffice pretty well at the time so I figured it would be interesting to see how well I did in their Office proficiency test. They set it up for Office XP and away I went. The funny thing was that they tested my ability to find things in the menu within a couple of tries and didn't let me use keyboard shortcuts. Despite this, the interface of Word was so similar to what I knew from OpenOffice 1.x that I was able to pass the test easily with what they said was a very high score. Some of the really specific functions in Office versus OpenOffice differ in their placement or what they are called but most are close enough that a user of one will be perfectly able to use the other. This is why schools should encourage the use of OpenOffice. This massive discount is a cyncical attempt by MS to get students so used to Office that they won't consider anything else.
On a similar note, I recently bought my Mum a MacBook and just gave it to her. She has never used anything other than Windows but even without training she was able to find her way around but recently she was struggling to get Word to format some pictures properly on the page so I suggested she use the trial copy of Pages. She was amazingly difficult to convince to try and use anything other than Office, even though she happily used OpenOffice on her Windows box but eventually she tried it and a few minutes in she was suddenly very enthusiastic about it. In the end, what MS wants to do is get people scared of trying anything else. Ever. Teaching people only to use Windows and MS applications from an early age is key to this strategy and it is a cycle that needs to be broken if we are ever to have people who can really function in the face of alternative software. MS has been so successful that people often struggle when moving from one Windows machine to another simply because an icon is in a different place. That just sucks.
"Anti-virus should be an included part of the OS along with updates. It's addressing flaws in the product."
Anti-virus shouldn't exist since viruses themselves shouldn't exist. The problem here is that viruses and trojans took hold when MS bolted networking into an operating system that was unprepared for it. This made the virus problem much worse than it ever would have been because they spread like a plague with nothing to prevent the spread. Anti-virus tries to fix this by detecting this spread but this can never be completely secure and the virus/trojan writers know they have a very fertile ground and do all they can to keep it that way. On other platforms, the virus problem never got out of hand because the mechanisms for their propogation were reasonably controlled. It would only have taken a little roadblock here and there to prevent the virus problem entirely but MS chose not to implement them. Macs and Linux/UNIX machines will never have the same problem even if they become as popular as Windows is today because the virus/trojan culture cannot get sufficient traction to be successfully propogated in sufficient numbers to make it worthwhile. On Windows the lax security made a wonderful landscape for malware propogation and now that it exists the writers will keep trying to maintain it so MS has a much harder struggle to kill off the problem than those platforms which never had it in the first place. Even with equivalent security to a Mac I don't believe that Windows would be as secure since there is just so much out there already.
"This guys little application does tons of random memory reads"
If only that was the case but actually it is very linear. The application can hold the whole of its memory requirements in cache these days so it hardly has to touch main memory and it was designed to do all the inner loop code using only registers. Heck, I doubled the size of the inner loop just to avoid a single register copy because it made a significant performance increase.
The reason I like this code is that it shows how many operations you can expect a chip to achieve when it isn't having to wait on main memory. It is an extremely compute intensive application with very little I/O. If it really was about random memory reads then I would be inclined to agree with you but it isn't, it loads blocks of memory into cache and chews through them linearly.
I am pretty processor agnostic. I did really like the Alpha but that is dead now and that is a real shame. I also object to being called a 'fanboi' as I personally don't own any AMD kit since I generally prefer Macs which means regardless of what you might think all my personal machines are PPC or Intel. If anything though, my application is an example of general computing applications in that it doesn't use any SSE tricks to increase performance. It is just code that anyone could write in C and compile on whatever machine they have to hand so the performance is pretty real world. Sure, I've spent a bit of time right in the core making the thing efficient but that is where the program spends 99% of its time so not doing so would be stupid.
How's this for something to make your head spin, we were benchmarking some Java code written by someone else the other day and found that Java under Windows XP Pro on one of these Opterons was no quicker than it was on my G4 1.5Ghz PowerBook but the same app under Linux on the same Opteron was 4x quicker. The guy running the machine under Windows is a Java developer and now wants Linux installing and will use Windows via VMware in future. Also interesting, the 2.66Ghz MacPro was about 30% slower than the Opteron under Linux running the same bytecode but still faster than Windows running the same code. Not my 'little application' but still seems to follow the same trend which I thought was interesting. Apart from the Windows thing. No idea what is wrong with Java under Windows unless Sun did it deliberately which wouldn't surprise me.
Hmmm, abject failure to deliver on Longhorn and the fact that two years in they had to dump it because it wasn't going to work and do a simple retread of Windows 2003 with a bit of flashy OS X ripped off graphics is how I remember it. Blaming XP SP2 is simply trying to change history. They made all these great claims about how wonderful Longhorn was going to be and now they are claiming that Apple has copied all their great ideas and delivered them in a working OS while they have dropped most of them because they couldn't make it work. But Apple could. And Apple is the one doing to copying.
How about this for a prediction. The next version of Windows will be late, more of the same, still insecure and a desperate copy of whatever Apple was shipping in 2007.
I should say that this program was written a very long time ago originally. It implements an efficient but standard Smith and Waterman dynamic programming algorithm. I have done vectorisation of this algorithm in the past and the performance improvement was dramatic (about x20). With this test program though, it hasn't really benefited from extreme compiler optimisations. I do remember running it after compiling with ccc on an Alpha and seeing a 30% speedup so there is definitely room for improvement but most of these benchmarks are with gcc on x86 which isn't too bad. OK, not icc but I don't have access to that.
As you say, the G5 was never really good at integer work but its floating point performance isn't too bad. If only HPaq hadn't killed Alpha. *sigh* What a chip.
"I hope the benchmarks don't take get advantage out of using 64-bit arithmetic."
Nope, straight 32 bit. If it had been 64 bit then the Core 2 Duo would also have seen a more significant boost versus its 32 bit predecessor not to mention the G5 should have been better than the G4 which it wasn't.
"Any numbers for an Athlon 64? I just bought a 3800+ single core and would like to be made really excited about it.:-P"
Pretty much the same as the Opteron in this case. The program doesn't really hammer cache or main memory, just the CPU. Work out your clock speed as a percentage of 2Ghz and do the sums and that should be the number.
The Opteron, Core 2 Duo and Core Duo are all dual core chips in this test, the others single core although the G5 was a dual processor system. Since the program is single threaded there isn't any benefit to extra cores in this case but if you multithread your program you can utilise multiple CPUs and improve performance substantially. At the moment, the sweet spot seems to be dual core as modern operating systems are happy working with multiple CPUs. A quad system is really only necessary if you are maxing out a dual core system and need to run more tasks at once.
"Actually I was thinking more about benchmarking/coding flaws than lying from your part."
Certainly a possibility. In my defense I would like to point out that all benchmarks are open to question. I know my own code, I know what it does and it doesn't do much but it does a lot of it so the performance figures are what they are. I originally wrote this code on an SGI, ported it to Linux on a 486, SPARC, Alpha, PPC and so on. Its old and simple but does real work. While I could make it faster using SSE and have done so with other code, that wasn't the purpose of these numbers. It was simply to see what the processors do using the same code, the same compiler and similar OSs (Linux v OSX in this case). Anyway, the PPC code from gcc is likely not particularly well optimised, especially for the G5, but for the x86 based chips it isn't too bad. All code was compiled 32 bit with just the basic optimisation a C programmer would put. Compiling with -m64 doesn't really help it much and on the Intel chips has been known to reduce performance so I stuck with 32 bit.
"Well, until you show us your source code those numbers are as believable as anything else one might randomly type here..."
I can't because the program is really large and it doesn't entirely belong to me (you know, work for people, they own your code).
You're right, I could just be making these numbers up and if you prefer to believe that then there is nothing I can do to change your mind. All I can say is that this is my own (admittedly anecodatal) experience.
"The P3 you list looks a Coppermine, I suspect a P3 Tualatin would perform much better."
Pretty sure it is a Tualatin since it is a 1Ghz PIII Mobile which I bought in early 2002 (http://www.theregister.co.uk/2001/01/31/chipzilla _readies_1ghz_mobile_piii would seem to support this).
Given that it is a Tualatin, then the peformance of the Core Duo at 2Ghz looks about right. The Core 2 Duo gets about 10% better performance clock for clock from all the blurb I have read except when it comes to SSE where it is about twice as fast so the performance figure of 146 million also looks pretty much on the mark too as a 2Ghz Core 2 Duo should be able to manage about 110 million if you scale the figure for clock speed and that is (surprise) ~10% quicker than the Core Duo at 2Ghz (94 million) so the basic integer performance of the Core 2 Duo is better than the Core Duo but doesn't compare with the 205 million the 2.0Ghz Opteron manages.
"perhaps you need to write some more cache efficient code to test with. goto BLAS can feed the beast like no other."
goto BLAS uses SSE so doesn't count. It has already been acknowledged that the SSE implementation of Core 2 Duo is very good. The new AMD chips may address this but we won't know until we see the benchmarks. For non-SSE the Core 2 Duo is a little better than the Core Duo which was similar to the PIII/PII/PentiumPro clock for clock. The current Opteron is much quicker clock for clock for non-SSE work. Also, my test code is really just integer code and it works mainly in registers and uses very little memory so it is quite a good test of the raw performance of a CPU which is why I like it.
"Care to publish your numbers that debunk all the other hardware sites that are typically AMD-biased anyways?"
OK. I can't give you the code but it is my own implementation of a pretty standard bioinformatics sequence comparison program which doesn't use SSE/MMX type instructions and is single threaded. On all platforms it was compiled using gcc with -O3 optimisation. I have tried adding other optimisations but it doesn't really make much difference to these numbers (no more than a couple of percent at best).
AMD Opteron 2.0Ghz (HP wx9300) - 205 Million calculations per second Intel Core 2 Duo 2.66Ghz (Mac Pro) - 146 Million Intel Core Duo 2.0 Ghz (MacBook Pro) - 94 Million IBM G5 PPC 2.3 Ghz (Apple Xserve) - 81 Million Motorola G4 PPC 1.42 Ghz (Mac mini) - 72 Million Intel P4 2.0 Ghz (Dell desktop) - 61 Million Intel PIII 1.0 Ghz (Toshiba laptop) - 45 Million
Interesting things about these numbers. The Core Duo is clearly a close relative of the PIII since the performance at 2Ghz is roughly twice that of the PIII at 1Ghz. The P4 at 2Ghz is really very poor indeed which isn't a huge surprise as it was never very efficient. The G4 PPC puts in a reasonable result easily beating the much higher clocked P4 (what, the Mac people were right? Shock!) although I have to say that the performance of the G5 is disappointing. The Core 2 Duo isn't a bad performer although it does have the highest clock speed of any processor in this set but it is seriously beaten by the Opteron. From these numbers, a Core 2 Duo at 2Ghz would be about half as quick as an Opteron at the same speed.
In my own benchmarks (generic C integer and floating point scientific code) I have found that the Core Duo and Core 2 Duo aren't all that quick compared with an AMD64. Clock for clock the AMD64 Opterons we have are about 50% quicker than an equivalent Core 2 Duo for integer work. I know this doesn't agree with all the usual magazine benchmarks but they are heavily biased towards using SSE instructions where possible and it is SSE where the Core 2 Duo has been a real improvement over previous Intel designs and also bests the AMD chips. Hopefully, AMD has recognised this and the new SSE implementation will bring them back on par with Intel for these benchmarks but even today an AMD64 processor is a beast and more than a match for anything Intel produces.
The problem with Dell is that they have a reputation for the cheapest machines around. Yes, they have the XPS line and Alienware but those are not where the volume is. Most of their traffic is in the cheap end of the scale where they make pretty much no money and people are always trying to game the system to get even cheaper kit. This is not sustainable. They need to up the price on their lowest cost systems and improve the quality as well as reduce the number of options.
Oh, and stop putting stupid blue LEDs in everything, they are tacky and annoying (looking at the horrid XPS 17" laptop on another table that is just a horrible fright).
We had a problem with people harassing the sysadmin (me) to do stuff and I was having trouble with time management and documenting my workload. We already had a trac system (http://trac.edgewall.org/) in place for other reasons and we used this to implement a sysadmin request system where people could enter their problem in trac and their request would be sent to the sysadmin (or a list of people in our case) who would then resolve the issue and report it as such. This produced a nice audit trail showing requests and their resolutions as well as any outstanding issues. Of course, it is all open source and free which is also nice.
Slashdot Mirror
When I was starting my own company a few years back I applied for sponsorship to get it going. As part of the interview process they asked applicants to do a 20 minute presentation for the selection board. I wanted the board to understand what I was doing and to realise I knew my stuff so I produced precisely four slides. They were black and white, plain simple text consisting of a single sentence which served as a banner to each of the points I was making along the lines of "what is it?", "who uses it?", "will people buy it?" and "how much?" I simply stated my case for each of these and then opened for questions.
I got the money.
Often presentations are used as a crutch by people who don't really know what they are talking about and if you get them off track they lose their way which is why they prefer to write everything they want to say down and read from it. I have never believed that this approach will be successful in engaging with an audience and it looks like I was right.
"- create a folder named FordChevyDodge. Search for 'Chevy'--it pops right up. Search for 'hevy'--nada. Oops. "
:-)
Its the capitalisation. Change the file to FordCHevyDodge and search for 'hevy' and it shows up just fine. I'll admit that it is odd though. Maybe they are assuming that if a word starts with a capital then that is a proper name and you wouldn't want to find it if you missed the first letter. Possibly it helps the indexing by reducing size. Or something.
Nice one though, I've never run across the problem before and I use spotlight all the time.
"I can't be the only person still running 10.3.9 (on 2 boxes). Spotlight just wasn't that killer of an app to me."
The download page says you need 10.4+ to run Google Desktop so you're still SOL.
"I'd be very curious to know your TV!"
Yeah, I was surprised. I had assumed that since 1080p worked through component there would be no problem with HDMI. However, I have an upscaling DVD player with HDMI and it supports 1080p output but whenever I try to select 1080p it switches back to 1080i. Looking in the specs of my projector (InFocus In76) it does say that it only supports 1080i on HDMI but 1080p through DVI and component. I'm actually not bothered that much since I can't see any difference between 1080i and 1080p over component as far as resolution goes although with long cables there is some ghosting on the 1080p which is absent with 1080i so that is what I use. Too many people are fixated on the 'p' part and don't realise that 1080 is 1080. The only time the interlacing would have any impact is if input source was aquired that way. If the source is progressive but transmitted as 1080i and then deinterlaced at the other end it doesn't look any different to progressive all the way since the deinterlacing is perfect. With the Xbox HDDVD drive it forces 1080i over component but the disc is progressive so this isn't a handicap at all and I don't know why they insisted on it. Finally, although my projector has a native resolution of 1280x720 it looks better when driven with the 1080i output from the Xbox because the scaler in the projector does a better job of preserving the detail on HDDVD than the Xbox itself does. With my upscaling DVD player though, I run that at 1280x720 which means there is only one scaler involved. Running at 1080i would mean it is scaling up and then the projector is scaling down again and in my own tests this resulted in significant image degradation.
Oh, and HDDVD blows even the best DVDs out of the water despite the fact that I am still only actually using about half the available resolution on the disc. Don't let anyone tell you that an upscaling DVD player is anything like as good as HDDVD or BluRay. Maybe on a small TV (32" or so) you might not be able to tell the difference but on my 70" front projector the difference is very apparent.
I have a 360 with HD-DVD drive and wireless adapter and a pair of wireless pads. If I was to switch to this just for the sake of HDMI and a bigger drive, bearing in mind that my HDTV doesn't support 1080p over HDMI, only over component and also I can't even see a difference between component and HDMI on the thing, I would also have to replace all the white extras. Oh, and I actually think the 360 looks better in white. Mind you, when I came to buy a new iPod Video I bought the white one because I didn't like the look of the black one.
"MS represents itself as a software company, where you generally get 90% margins. "
Where MS generally gets 90% margins. These profits only exist in a monopoly market. If the market place was really competitive MS would be making 30% at best because most of their profits would be rolled back into development costs. MS doesn't spend anything like 60% of profit on R&D, especially not on the cash cows (Windows and Office). Yes, they do burn a lot of money trying to buy their way into other markets but those are all costing them lots. In real terms, MS has raise the cost of their software substantially over the years in order to fund this profligate spending to try and dominate the entire industry. In reality, MS should be made to charge far less for their software and this would have the effect of reining in their dominance. Of course, like everyone else I am not going to hold my breath on that one. All I can do is avoid MS software where possible and pay them as little as possible where I can't.
Don't these schools typically have Windows site licences anyway? By buying Macs and installing Boot Camp the users get the choice they want and the school can install a copy of Windows as usual due to their site license. MS doesn't get paid twice for each PC which can only be good IMHO. Personally, I wouldn't give the users the choice. I would do a deal with parallels or VMware to allow the Windows software to be run while encouraging the users to run OS X all the time.
"What are they suggesting, that they don't want their prospective employees to be quick-thinking and efficient when it comes to menial and repetitive tasks?"
Unfortunately I think that is exactly the situation. The sort of people who need to do these tests have such an amazingly low capacity to actually function with software like Office that anyone who can use it in the way that is intended is probably far too high up the evolutionary tree to actually want to work in a job where they have to use Office all day long. To many of these people, Office *IS* the computer. The people doing the training also have such a low ability that they can't cope with people who know more about it than they do and this even goes for the teachers in schools too. What is holding computing and computer use back is the way people are trained (poorly) to use (poor) software (poorly) to do their jobs (poorly). I have been in several jobs where I was able to complete the entire days work in the first hour simply because I can use a computer efficiently. The training industry doesn't want people who can sit at a computer they have never used before and function quickly. And the support industry doesn't want a computer that doesn't break constantly. And the managers are too deluded to know that the training and support that people are getting is so second rate that they could actually afford to hire 1/8th of the people if they actually had a reasonable level of training and skill.
A few years back I was between proper jobs and had to do some temping work. The temping agency asked if I could use Office. I didn't know as I had only used OpenOffice as I had been a Linux user since 1994 and before that I was on SunOS. Anyway, I knew OpenOffice pretty well at the time so I figured it would be interesting to see how well I did in their Office proficiency test. They set it up for Office XP and away I went. The funny thing was that they tested my ability to find things in the menu within a couple of tries and didn't let me use keyboard shortcuts. Despite this, the interface of Word was so similar to what I knew from OpenOffice 1.x that I was able to pass the test easily with what they said was a very high score. Some of the really specific functions in Office versus OpenOffice differ in their placement or what they are called but most are close enough that a user of one will be perfectly able to use the other. This is why schools should encourage the use of OpenOffice. This massive discount is a cyncical attempt by MS to get students so used to Office that they won't consider anything else.
On a similar note, I recently bought my Mum a MacBook and just gave it to her. She has never used anything other than Windows but even without training she was able to find her way around but recently she was struggling to get Word to format some pictures properly on the page so I suggested she use the trial copy of Pages. She was amazingly difficult to convince to try and use anything other than Office, even though she happily used OpenOffice on her Windows box but eventually she tried it and a few minutes in she was suddenly very enthusiastic about it. In the end, what MS wants to do is get people scared of trying anything else. Ever. Teaching people only to use Windows and MS applications from an early age is key to this strategy and it is a cycle that needs to be broken if we are ever to have people who can really function in the face of alternative software. MS has been so successful that people often struggle when moving from one Windows machine to another simply because an icon is in a different place. That just sucks.
"Now they'll totally be worth $600!"
But will it be worth $832?
FYI the PS3 is £425 in the UK which is how I got to the above number.
"Anti-virus should be an included part of the OS along with updates. It's addressing flaws in the product."
Anti-virus shouldn't exist since viruses themselves shouldn't exist. The problem here is that viruses and trojans took hold when MS bolted networking into an operating system that was unprepared for it. This made the virus problem much worse than it ever would have been because they spread like a plague with nothing to prevent the spread. Anti-virus tries to fix this by detecting this spread but this can never be completely secure and the virus/trojan writers know they have a very fertile ground and do all they can to keep it that way. On other platforms, the virus problem never got out of hand because the mechanisms for their propogation were reasonably controlled. It would only have taken a little roadblock here and there to prevent the virus problem entirely but MS chose not to implement them. Macs and Linux/UNIX machines will never have the same problem even if they become as popular as Windows is today because the virus/trojan culture cannot get sufficient traction to be successfully propogated in sufficient numbers to make it worthwhile. On Windows the lax security made a wonderful landscape for malware propogation and now that it exists the writers will keep trying to maintain it so MS has a much harder struggle to kill off the problem than those platforms which never had it in the first place. Even with equivalent security to a Mac I don't believe that Windows would be as secure since there is just so much out there already.
"This guys little application does tons of random memory reads"
If only that was the case but actually it is very linear. The application can hold the whole of its memory requirements in cache these days so it hardly has to touch main memory and it was designed to do all the inner loop code using only registers. Heck, I doubled the size of the inner loop just to avoid a single register copy because it made a significant performance increase.
The reason I like this code is that it shows how many operations you can expect a chip to achieve when it isn't having to wait on main memory. It is an extremely compute intensive application with very little I/O. If it really was about random memory reads then I would be inclined to agree with you but it isn't, it loads blocks of memory into cache and chews through them linearly.
I am pretty processor agnostic. I did really like the Alpha but that is dead now and that is a real shame. I also object to being called a 'fanboi' as I personally don't own any AMD kit since I generally prefer Macs which means regardless of what you might think all my personal machines are PPC or Intel. If anything though, my application is an example of general computing applications in that it doesn't use any SSE tricks to increase performance. It is just code that anyone could write in C and compile on whatever machine they have to hand so the performance is pretty real world. Sure, I've spent a bit of time right in the core making the thing efficient but that is where the program spends 99% of its time so not doing so would be stupid.
How's this for something to make your head spin, we were benchmarking some Java code written by someone else the other day and found that Java under Windows XP Pro on one of these Opterons was no quicker than it was on my G4 1.5Ghz PowerBook but the same app under Linux on the same Opteron was 4x quicker. The guy running the machine under Windows is a Java developer and now wants Linux installing and will use Windows via VMware in future. Also interesting, the 2.66Ghz MacPro was about 30% slower than the Opteron under Linux running the same bytecode but still faster than Windows running the same code. Not my 'little application' but still seems to follow the same trend which I thought was interesting. Apart from the Windows thing. No idea what is wrong with Java under Windows unless Sun did it deliberately which wouldn't surprise me.
Hmmm, abject failure to deliver on Longhorn and the fact that two years in they had to dump it because it wasn't going to work and do a simple retread of Windows 2003 with a bit of flashy OS X ripped off graphics is how I remember it. Blaming XP SP2 is simply trying to change history. They made all these great claims about how wonderful Longhorn was going to be and now they are claiming that Apple has copied all their great ideas and delivered them in a working OS while they have dropped most of them because they couldn't make it work. But Apple could. And Apple is the one doing to copying.
How about this for a prediction. The next version of Windows will be late, more of the same, still insecure and a desperate copy of whatever Apple was shipping in 2007.
I should say that this program was written a very long time ago originally. It implements an efficient but standard Smith and Waterman dynamic programming algorithm. I have done vectorisation of this algorithm in the past and the performance improvement was dramatic (about x20). With this test program though, it hasn't really benefited from extreme compiler optimisations. I do remember running it after compiling with ccc on an Alpha and seeing a 30% speedup so there is definitely room for improvement but most of these benchmarks are with gcc on x86 which isn't too bad. OK, not icc but I don't have access to that.
As you say, the G5 was never really good at integer work but its floating point performance isn't too bad. If only HPaq hadn't killed Alpha. *sigh* What a chip.
"I hope the benchmarks don't take get advantage out of using 64-bit arithmetic."
Nope, straight 32 bit. If it had been 64 bit then the Core 2 Duo would also have seen a more significant boost versus its 32 bit predecessor not to mention the G5 should have been better than the G4 which it wasn't.
"Any numbers for an Athlon 64? I just bought a 3800+ single core and would like to be made really excited about it. :-P"
Pretty much the same as the Opteron in this case. The program doesn't really hammer cache or main memory, just the CPU. Work out your clock speed as a percentage of 2Ghz and do the sums and that should be the number.
The Opteron, Core 2 Duo and Core Duo are all dual core chips in this test, the others single core although the G5 was a dual processor system. Since the program is single threaded there isn't any benefit to extra cores in this case but if you multithread your program you can utilise multiple CPUs and improve performance substantially. At the moment, the sweet spot seems to be dual core as modern operating systems are happy working with multiple CPUs. A quad system is really only necessary if you are maxing out a dual core system and need to run more tasks at once.
"Actually I was thinking more about benchmarking/coding flaws than lying from your part."
Certainly a possibility. In my defense I would like to point out that all benchmarks are open to question. I know my own code, I know what it does and it doesn't do much but it does a lot of it so the performance figures are what they are. I originally wrote this code on an SGI, ported it to Linux on a 486, SPARC, Alpha, PPC and so on. Its old and simple but does real work. While I could make it faster using SSE and have done so with other code, that wasn't the purpose of these numbers. It was simply to see what the processors do using the same code, the same compiler and similar OSs (Linux v OSX in this case). Anyway, the PPC code from gcc is likely not particularly well optimised, especially for the G5, but for the x86 based chips it isn't too bad. All code was compiled 32 bit with just the basic optimisation a C programmer would put. Compiling with -m64 doesn't really help it much and on the Intel chips has been known to reduce performance so I stuck with 32 bit.
"Well, until you show us your source code those numbers are as believable as anything else one might randomly type here..."
I can't because the program is really large and it doesn't entirely belong to me (you know, work for people, they own your code).
You're right, I could just be making these numbers up and if you prefer to believe that then there is nothing I can do to change your mind. All I can say is that this is my own (admittedly anecodatal) experience.
"The P3 you list looks a Coppermine, I suspect a P3 Tualatin would perform much better."
a _readies_1ghz_mobile_piii would seem to support this).
Pretty sure it is a Tualatin since it is a 1Ghz PIII Mobile which I bought in early 2002 (http://www.theregister.co.uk/2001/01/31/chipzill
Given that it is a Tualatin, then the peformance of the Core Duo at 2Ghz looks about right. The Core 2 Duo gets about 10% better performance clock for clock from all the blurb I have read except when it comes to SSE where it is about twice as fast so the performance figure of 146 million also looks pretty much on the mark too as a 2Ghz Core 2 Duo should be able to manage about 110 million if you scale the figure for clock speed and that is (surprise) ~10% quicker than the Core Duo at 2Ghz (94 million) so the basic integer performance of the Core 2 Duo is better than the Core Duo but doesn't compare with the 205 million the 2.0Ghz Opteron manages.
"perhaps you need to write some more cache efficient code to test with. goto BLAS can feed the beast like no other."
goto BLAS uses SSE so doesn't count. It has already been acknowledged that the SSE implementation of Core 2 Duo is very good. The new AMD chips may address this but we won't know until we see the benchmarks. For non-SSE the Core 2 Duo is a little better than the Core Duo which was similar to the PIII/PII/PentiumPro clock for clock. The current Opteron is much quicker clock for clock for non-SSE work. Also, my test code is really just integer code and it works mainly in registers and uses very little memory so it is quite a good test of the raw performance of a CPU which is why I like it.
"Care to publish your numbers that debunk all the other hardware sites that are typically AMD-biased anyways?"
OK. I can't give you the code but it is my own implementation of a pretty standard bioinformatics sequence comparison program which doesn't use SSE/MMX type instructions and is single threaded. On all platforms it was compiled using gcc with -O3 optimisation. I have tried adding other optimisations but it doesn't really make much difference to these numbers (no more than a couple of percent at best).
AMD Opteron 2.0Ghz (HP wx9300) - 205 Million calculations per second
Intel Core 2 Duo 2.66Ghz (Mac Pro) - 146 Million
Intel Core Duo 2.0 Ghz (MacBook Pro) - 94 Million
IBM G5 PPC 2.3 Ghz (Apple Xserve) - 81 Million
Motorola G4 PPC 1.42 Ghz (Mac mini) - 72 Million
Intel P4 2.0 Ghz (Dell desktop) - 61 Million
Intel PIII 1.0 Ghz (Toshiba laptop) - 45 Million
Interesting things about these numbers. The Core Duo is clearly a close relative of the PIII since the performance at 2Ghz is roughly twice that of the PIII at 1Ghz. The P4 at 2Ghz is really very poor indeed which isn't a huge surprise as it was never very efficient. The G4 PPC puts in a reasonable result easily beating the much higher clocked P4 (what, the Mac people were right? Shock!) although I have to say that the performance of the G5 is disappointing. The Core 2 Duo isn't a bad performer although it does have the highest clock speed of any processor in this set but it is seriously beaten by the Opteron. From these numbers, a Core 2 Duo at 2Ghz would be about half as quick as an Opteron at the same speed.
In my own benchmarks (generic C integer and floating point scientific code) I have found that the Core Duo and Core 2 Duo aren't all that quick compared with an AMD64. Clock for clock the AMD64 Opterons we have are about 50% quicker than an equivalent Core 2 Duo for integer work. I know this doesn't agree with all the usual magazine benchmarks but they are heavily biased towards using SSE instructions where possible and it is SSE where the Core 2 Duo has been a real improvement over previous Intel designs and also bests the AMD chips. Hopefully, AMD has recognised this and the new SSE implementation will bring them back on par with Intel for these benchmarks but even today an AMD64 processor is a beast and more than a match for anything Intel produces.
F.E.A.R. made it and Far Cry didn't? Give me a break. I bought both and Far Cry was amazing while F.E.A.R...... wasn't.
The problem with Dell is that they have a reputation for the cheapest machines around. Yes, they have the XPS line and Alienware but those are not where the volume is. Most of their traffic is in the cheap end of the scale where they make pretty much no money and people are always trying to game the system to get even cheaper kit. This is not sustainable. They need to up the price on their lowest cost systems and improve the quality as well as reduce the number of options.
Oh, and stop putting stupid blue LEDs in everything, they are tacky and annoying (looking at the horrid XPS 17" laptop on another table that is just a horrible fright).
We had a problem with people harassing the sysadmin (me) to do stuff and I was having trouble with time management and documenting my workload. We already had a trac system (http://trac.edgewall.org/) in place for other reasons and we used this to implement a sysadmin request system where people could enter their problem in trac and their request would be sent to the sysadmin (or a list of people in our case) who would then resolve the issue and report it as such. This produced a nice audit trail showing requests and their resolutions as well as any outstanding issues. Of course, it is all open source and free which is also nice.