I have an all digital home studio, with a multitrack, Mac, synths etc etc. I love the freedom it affords me to make music how I like and when I like. However, if I were putting together a big album project, I would still use a studio for at least some bits. Here's why...
1, You have a nice acoustic space for recording "real" instruments, like Drums, Guitars etc with nice mics (Neumann, AKG, B&K etc) 2, You don't have potential noise abatement issues like you would in an apartment. If I want to crank up that 'ol Mesa Boogie amp, it's much easier in a studio. 3, Studios usually have great monitoring systems and outboard equipment. The rooms are also designed to listen to music in, as opposed to the perfectly rectangular study in my abode. No standing waves!!! 4, You have the expertise of a sound engineer. This has enormous value, IMHO.
All these new tools are wonderful, and I make as much use of them as possible. They don't, however, replace experience and plain old skill. I didn't start playing with my own gear until I had been in a few recording studios and saw how it was done. I do love the fact that the entry cost of recording has come down dramatically with the advent of DAW's (Digital Audio Workstation).
This hits the nail on the head, I think. Make money legally and ethically, then spend and invest! Everyone benefits: more jobs, better standard of living etc etc.
The beauty of this system is *you* get to choose where your money goes. Want to donate to a non-profit? Go ahead, use the money you have earned to do that! Want to boycott a company whose practices you don't agree with? Fine, withhold your cash from them and spend it with someone who is more compatible with your ideaology.
Agreed. The so called "middle classes" of a nation are the engine room of the economy. They are the largest in number in most democratic/market economy nations and will therefore spend the most on goods and services. They also often carry the largest tax burden. Added to that, they also often run small businesses, employ people and add significantly to the wealth of a country. If they are given an incentive to invest (or at least not a disincentive) then they help actually make the economy grow.
I am intrigued by your definition of ethics and morals as they pertain to business. What are they exactly? Can you elaborate on your business method of "a community of people working together". Does this mean all people in the community are compensated for their work equally? Or are they paid according to merit/effort/output etc? The former is communism and the later is capitalism IMHO. I'll take the latter thanks.
I think that business has been unfairly demonised in the wake of the U.S. downturn. There seems to be this mindset among some that businesses are somehow evil and are actively pursuing policies to hurt citizens. What a load of nonsense! If they did adopt said policies, they wouldn't have any customers to supply! I think the people responsible for the Enron/Worldcom debacles etc etc scandals should go to jail, for sure. They committed acts of fraud, after all. But let other businesses get on with it! The only way the world economy will improve is for free enterprises to get profitable and efficient. Everybody benefits from this, as their products are cheaper and they will tend to actually employ more people as their business expands.
As for the offshore thing, why shouldn't those businesses pursue a cheaper way of doing business? It's not illegal, heck, it's not even immoral! If you dislike these practices so much, boycott these companies, buy shares in them and vote against the board at the next shareholder meeting, etc. In other words, act in your own self interest. If a company works out that they are less profitable as a result, then they'll most likely stop doing it. Don't forget also that citizens of those countries who get the outsourcing contracts improve their standard of living.
Finally this pratt is no longer in a position to do any more damage to his portfolio. IMHO he has done more to damage telecoms and tech in Australia than any other communications minister in recent history. Good riddance Richard. Get rid of Tony Abbott while you are at it, John.
Please permit me to laugh sarcastically at Dell. Their ad campaigns champion a non-proprietary approach, but how is Intel hardware any less proprietary than anybody else (or their other buddies Microsoft)? I don't see any open source CPU's on the market. Sure, there are multiple vendors you can get system boards, disks etc etc from for PC's, but the same also applies for SPARC, MIPS, PPC et al. Not to mention that some of the proprietary O/S's run on more than one CPU family (HP-UX, VMS, Solaris).
Let's not be fooled into thinking that if Michael Dell could corner a market and bump up his margins, he wouldn't do so in a heartbeat.
Having said all of that, I love standards (there are so many of them), as it makes interoperability easier and reduces vendor lock in. That's where the real value is for me. He is right, though, about standards opening the door for high volume sales. This is often the case, in my experience.
I agree. Unless the new tech is embraced as a lifestyle choice, the gloss of the gadgets will wear off and they will sit unused in a corner. I also blame the manufacturers for this. So many devices are put on the market that have an initial "gee-whiz" factor, but wear thin after a while. Think of how many supposedly "indepensible" tools/gadgets/toys you have in your house and how many *actually* get used...
One of the things I like about the show "Fresh Gear" on TechTV is that they aren't afraid to say "this sucks" or "this has limited use". Many times I have found myself agreeing with their conclusions.
Having said all of that, some of the stuff installed for the family in the story made me wanna drool:)
I tend to agree. Where I work, we use LDAP extensively as our corporate directory. It works sufficiently well that I can't imagine life in a ~80,000 person company without it. We don't just use it for auth, but all sorts of other useful information about employees (e.g. expertise, cost centres, calendar, training etc etc). There is no reason that a directory server cannot be congruent with Unix. It works with our HP-UX/Solaris infrastructure just fine. And after all, what is NIS/NIS+, if not a (albeit) directory server?
This article could have been talking about outsourcing in general, not just overseas. One thing I have learned over the years is that you don't outsource simply to save money. In some cases, it costs *more* money to outsource than to do things internally.
In my opinion, you outsource to gain expertise you don't otherwise have, focus on your core business or other sound commercial reasons. Reduced costs should be the last reason for doing this. I have seen far too many outsourcing contracts go bad as a result of a failure to factor in the appropriate costs (this is on the providor as well as customer sides). I'm not saying don't ever do it, but be aware of a few things.
One of the comments at the end of the article also raised a good point: intellectual property. Be careful about dealing with *any* outsourcing company whom you suspect might take your brilliant idea and sell it on the open market. The opportunity cost of this happening can be staggering.
Another often forgotten part is the opportunity cost of not having an internal staff who understand and are aligned with the goals of the organisation. That is, those high potential technical and management staff who add more value to the business with their ideas/techniques etc than they cost in terms of total compensation. Do you want to outsource those people? An outsourcing comany has only one goal: to maximise the amount of profit they make per contract. This is not a bad thing, but it may mean that their goals diverge from your own.
The IT market seems to be very cyclical when it comes to outsourcing. It happens to be in favour right now, but who knows if that will be the case in 5 years time.
Compared to the ASCI series put together by the US government, $5M is not a lot of money. Consider that a single Sun SF15k, IBM p or z series or HP SuperDome can cost easily this much (that's just one machine).
I think VT are getting a pretty good deal. Really large clusters/supercomputers *can* cost upwards of $30M, depending on the configuration.
As someone who like photography and technology, I have to say I like the pictures. A reviewer of the G5 put it best when he said that it had a "brutal austerity". I think these photos capture it beautifully.
As other posters have alluded to, the most constructive answer is: "Sure we can do that, but it will take this much time and this much money". I was in your situation once and I dealt with it in a number of ways:
1, I had a white board, with days of the week and hours of the day in a matrix. I filled in available spots when people wanted work done. If someone complained, I invited them to find a free spot in the matrix and write their name in.
2, Make appointments with your users. That is, try and schedule the time required. I used to leave a bit of "fat" in there in case emergencies came up.
3, If somebody gives you a job that you think will take more time/effort than they think, send them and their boss an estimate and see if they still want to do it.
4, Try and put some sort of ticketing system in, where you can track who has been asking you what and how much time it took. The worst thing is to have no proof against people who complain about the service they are getting.
5, Try and build an infrastructure that requires next to no babysitting (easier said than done), which will free up more of your time to deal with one-off end user issues.
6, Send reports about what you have been up to to your boss and/or the big boss. It might get them on your side.
It's not really a matter of saying no, but more a case of managing expectations
Automatic pitch correction has been around for some time. I saw a variation of it over ten years ago with the Eventide Harmoniser series (as used by Steve Vai and others). The difference now is that they are a lot more sophisticated and it is harder to tell when one is being used.
Are they evil? I think they are really not a lot different from all the other tricks you might use on a vocalist's sound: compression (to fatten up the sound), EQ (to notch or boost particular frequencies), reverb (to give it ambience), detuning/chorusing (to thicken the sound) and harmonising (instant vocal harmonies). These techniques have been used live and in recording studios for decades and are considered fair game.
I personally think a crap vocalist who's in tune is still a crap vocalist. There's more than just perfect pitch. There's got to be a bit of personality and power in the voice too. A slightly off tune Freddie Mercury, Sarah McLachlan or Geoff Tate will still sound better than a perfectly in tune Britney or Kelly Osbourne (to my ears).
There's an old saying in the music business: crap gear in the hands of a talented professional will still sound better than great gear in the hands of a no talent.
This might sound like I am stating the bloody obvious, but it's true. I think there are several facets to good benchmarking (based on my own experiences and reading other reports)....
1, Choose a test/workload that is representative of what *you* will be doing. There is no point in looking at SPECINT200 if you are going to be running an I/O intensive application like a RDBMS. Try and run or study tests that are relevant to the intended use of the system/component you are benchmarking.
2, Take note of things like compiler flags etc. These are important in tests like SPEC, as your results can vary wildly according to things like optimisation level. Some compilers produce faster code on certain CPU families and not on others. This is a reason why a lot of vendors will build their own compilers and test with them (e.g. SGI, SUN, DECPAQ).
3, Look at the full disclosure notice in the benchmarks. Take a look at the system configuration used. This is particularly, IMHO, on tests like TPC-C. The score you see might be based on a really whacky config, like most of the figures at the top of the list. For example, look at the Proliant figure (709k) and look at the config: 32 x 8 way servers to run a single database. Then compare it to a 64-way SuperDome or 32-way p690. Which comfig makes more sense? For a database, I would likely go with the single system for simplicity's sake. On another application, maybe the cluster would make sense.
4, Compare apples to apples. This is the hardest part, as CPU's, OS's, I/O, Apps. Compilers etc etc all vary across platforms. I like to to try and compare one variable if possible. To take the TPC-C again, I try to compare DB against DB, Cluster against Cluster, SMP against SMP etc etc. There is nothing to be gained, IMHO, from comparing MS-SQL server in a cluster on Xeon with Win2k3 to Sybase on a SF15k running SPARC Solaris. How do you properly compare these two results? Maybe the solution would be to look at SQLServer on one system against another or Sybase vs Oracle on a similar Unix system.
5, YMMV. Benchmarks are only ever an indicator of performance, not a guarantee. I tell my customers this all the time. They represent a result with a particular system, data set, O/S, tuning settings etc etc at a point in time. Other people's results with a similar config might differ considerably.
We have had at least a couple of Sun representatives come and present to us with PowerBook's, so I wouldn't doubt that plenty of Sun employees have Apple systems. Makes sense when you think about it: they're not running Windows and have Unix underneath. I would think they'd integrate pretty nicely with the Sun internal systems and allow them to do cool things like run Keynote, and Photoshop (for the marketing group) etc etc
If only OpenOffice/Staroffice would run in Aqua mode under OSX, everything would be just perfect:)
I work in the Oil&Gas business and we use Linux clusters (and in the past bloody large Sun, IBM and SGI systems) for seismic processing and reservoir simulation. These particular problems are DSP and FP intensive and also can require a fairly large amount of memory to run. They are exactly the kind of commercial workload either a supercomputer or cluster can chew on.
Some of our customers (I work for a company that writes the software, amongst other things) have upwards of 100TB of 3D Seismic they want to process. These jobs can take weeks or months to run. The simulation jobs can take days as well. Obviously having a big computer or tight cluster of lots of small ones will help decisions get made faster and/or more accurately.
There are other examples too: I met a gentleman who works for the lab that does crash simulation for Porsche, Audi and VW. Another example would be an ex-boss of mine who went to work for an engine manufacturer, who used a couple of SGIs to simulate the bore and stroke in a cylinder. The simulation took several weeks to run. They need large computers to do this too. So there is a market for these machines.
It gets down to what kind of problem you are trying to solve and how you are trying to solve it. Clusters built from common of the shelf components need an interconnect for dividing up the workload. For algorithms that don't require a lot of message passing between nodes, a relatively low bandwidth, relatively high latency interconnect like Gigabit Ethernet or (compared to something NUMA) Myrinet is sufficient. There comes a point, however, where the amount of message passing and/or the speed of the interconnect may flatten out performance. That is, simply adding more nodes to the problem yields little or no performance gain. You are kind of stuck when this happens, unless you replace the processors in the cluster with faster ones or someone comes along with a faster connector.
A supercomputer (and to a lesser extent large SMP systems, like the p690, E15k or SuperDome) is designed from the ground up with an extremely high bandwidth, low latency interconnect. This means that those problems requiring message passing between processors will execute a lot faster than the self made cluster (in theory anyway). In fact, you might find that such a system with lower speed CPU's will outperform a cluster of 3Ghz PC's, just because the machine is designed with "total performance" in mind. There are other smarts built into the system and the O/S is generally engineered specifically for the processor (e.g. Unicos), which is the opposite approach of Linux and *BSDs.
Which approach is right for solving *your* problem gets down to how much money you have, the algorithm you are using, the size of the problem set, your own expertise and a bunch of other factors. It's a bit like deciding whether to use a semi-trailer or a group of Subaru WRX's to move a house full of furniture from one place to another. Either way will work, but one is generally more optimal for a given problem set.
For a more thorough exploration of this topic, I would recommend the O'Reilly book: High Performance Computing by Severance and Dowd, which discusses these kinds of issues in great detail. It's a bit old these days, but much of the information is as valid now as it was then.
I have started doing Yoga *after* work and have already found that it helps me relax and unwind. It has other benefits, such as toning muscles and increasing flexibility. It's not for everybody, granted, but I gotten some benefit out of it.
All benchmarks are only useful up to a point. What really counts is how they run *your* apps. The Intel platform does a good job for single threaded uniprocessor tasks, as evidenced by things like SPECINT and SPECFP. However, the story in other benchmarks might be quite different (e.g. SPECINTRATE, SPECFPRATE, TPC etc etc). What it gets down to is that each configuration, compiler, optimisation flags etc are different enough to make real comparisons difficult. The other thing about benchmarks is that they only give you an *indication* of expected performance, not what it will really be like. The above factors will greatly influence how your workloads will run in the real world...
If I'm running a database, I want something like an SMP UltraSPARC, Power4, Alpha or Itanium2, whereas computational fluid dynamics might lend itself to a cluster of IA-32 boxes or G4/5 with Altivec.
The point is, every system has its sweet spot and Apple most definitely has theirs. I'm just glad to see a) that they have a 64-bit system out there and b) they can compete toe to toe with Intel again. This can't be a bad thing for the industry.
You're right, CD is superior to minidisc in terms of quality. A lot of consumers probably wouldn't be able to hear a huge difference. The MD and old Philips DCC were designed as cassette tape replacement, not usurping CD.
It's not just minidiscs that use optical. My Fostex digital multitrack also has an optical in and out. They are a little passe these days, but 5-10 years ago, many pro or semi-pro audio gear had them. I guess enough of 'em still do:)
I can certainly understand a group of people wanting to keep their jobs. It's what us tech folks have been trying to do for the past couple of years, after all.
Overall, I am not in favour of replacing human umpires with computerised systems. It's a *game* after all, played by people, not robots. We might as well get rid of the players while we're at it. I am in favour of it as a training aid, however. Anything that improves the overall standard of umpiring is a good thing.
The game of cricket has the same issue with the video replay technology. The television networks now have all these whiz bang replays that show from every angle. What they show me, as a fan, is that the umpire actually gets it right 98% of the time. This is certainly a higher strike rate than most of the players out there.
At least it's not as bad as the NFL, where you can *challenge* a call. What a load of crap!!! It slows down an already boring game.
I agree that Sun's new gear looks great. This is basically the reason why I am more hopeful now than before. I think they have responded to the Intel threat and have woken up to the realities of the market somewhat.
I hope to have my hands on some of their new gear very soon. Our local Sun reps seem pretty confident that GeoFrame will perform well on these systems. Their pricing is competitive too.
Some of their new developments in the graphics area looked very nice. They even impressed an SGI bigot like myself:)
Re:Do younger minds absorb quicker?
on
Ageism in IT?
·
· Score: 1
I think generalisations about people's abilities, based on factors like age, are rather silly.
It all comes down to the individuals involved. I know some top flight engineers and some shite ones. Neither group comes exclusively from a particular age group. It's a bit like engineering processes (I am thinking mainly of software here). The best process in the world will not get you any better result if the people implementing it are crap.
It's all about finding the best person for the job, giving them the best tools and letting them get on with it...old or young -- who cares?
Slashdot Mirror
I have an all digital home studio, with a multitrack, Mac, synths etc etc. I love the freedom it affords me to make music how I like and when I like. However, if I were putting together a big album project, I would still use a studio for at least some bits. Here's why...
1, You have a nice acoustic space for recording "real" instruments, like Drums, Guitars etc with nice mics (Neumann, AKG, B&K etc)
2, You don't have potential noise abatement issues like you would in an apartment. If I want to crank up that 'ol Mesa Boogie amp, it's much easier in a studio.
3, Studios usually have great monitoring systems and outboard equipment. The rooms are also designed to listen to music in, as opposed to the perfectly rectangular study in my abode. No standing waves!!!
4, You have the expertise of a sound engineer. This has enormous value, IMHO.
All these new tools are wonderful, and I make as much use of them as possible. They don't, however, replace experience and plain old skill. I didn't start playing with my own gear until I had been in a few recording studios and saw how it was done. I do love the fact that the entry cost of recording has come down dramatically with the advent of DAW's (Digital Audio Workstation).
This hits the nail on the head, I think. Make money legally and ethically, then spend and invest! Everyone benefits: more jobs, better standard of living etc etc.
The beauty of this system is *you* get to choose where your money goes. Want to donate to a non-profit? Go ahead, use the money you have earned to do that! Want to boycott a company whose practices you don't agree with? Fine, withhold your cash from them and spend it with someone who is more compatible with your ideaology.
Agreed. The so called "middle classes" of a nation are the engine room of the economy. They are the largest in number in most democratic/market economy nations and will therefore spend the most on goods and services. They also often carry the largest tax burden. Added to that, they also often run small businesses, employ people and add significantly to the wealth of a country. If they are given an incentive to invest (or at least not a disincentive) then they help actually make the economy grow.
I am intrigued by your definition of ethics and morals as they pertain to business. What are they exactly? Can you elaborate on your business method of "a community of people working together". Does this mean all people in the community are compensated for their work equally? Or are they paid according to merit/effort/output etc? The former is communism and the later is capitalism IMHO. I'll take the latter thanks.
I think that business has been unfairly demonised in the wake of the U.S. downturn. There seems to be this mindset among some that businesses are somehow evil and are actively pursuing policies to hurt citizens. What a load of nonsense! If they did adopt said policies, they wouldn't have any customers to supply! I think the people responsible for the Enron/Worldcom debacles etc etc scandals should go to jail, for sure. They committed acts of fraud, after all. But let other businesses get on with it! The only way the world economy will improve is for free enterprises to get profitable and efficient. Everybody benefits from this, as their products are cheaper and they will tend to actually employ more people as their business expands.
As for the offshore thing, why shouldn't those businesses pursue a cheaper way of doing business? It's not illegal, heck, it's not even immoral! If you dislike these practices so much, boycott these companies, buy shares in them and vote against the board at the next shareholder meeting, etc. In other words, act in your own self interest. If a company works out that they are less profitable as a result, then they'll most likely stop doing it. Don't forget also that citizens of those countries who get the outsourcing contracts improve their standard of living.
Finally this pratt is no longer in a position to do any more damage to his portfolio. IMHO he has done more to damage telecoms and tech in Australia than any other communications minister in recent history. Good riddance Richard. Get rid of Tony Abbott while you are at it, John.
Please permit me to laugh sarcastically at Dell. Their ad campaigns champion a non-proprietary approach, but how is Intel hardware any less proprietary than anybody else (or their other buddies Microsoft)? I don't see any open source CPU's on the market. Sure, there are multiple vendors you can get system boards, disks etc etc from for PC's, but the same also applies for SPARC, MIPS, PPC et al. Not to mention that some of the proprietary O/S's run on more than one CPU family (HP-UX, VMS, Solaris). Let's not be fooled into thinking that if Michael Dell could corner a market and bump up his margins, he wouldn't do so in a heartbeat. Having said all of that, I love standards (there are so many of them), as it makes interoperability easier and reduces vendor lock in. That's where the real value is for me. He is right, though, about standards opening the door for high volume sales. This is often the case, in my experience.
Fingers crossed, no issues on my iBook so far. I noticed a couple of apps launched a little faster (like terminal) but nothing major yet...
I agree. Unless the new tech is embraced as a lifestyle choice, the gloss of the gadgets will wear off and they will sit unused in a corner. I also blame the manufacturers for this. So many devices are put on the market that have an initial "gee-whiz" factor, but wear thin after a while. Think of how many supposedly "indepensible" tools/gadgets/toys you have in your house and how many *actually* get used...
:)
One of the things I like about the show "Fresh Gear" on TechTV is that they aren't afraid to say "this sucks" or "this has limited use". Many times I have found myself agreeing with their conclusions.
Having said all of that, some of the stuff installed for the family in the story made me wanna drool
I tend to agree. Where I work, we use LDAP extensively as our corporate directory. It works sufficiently well that I can't imagine life in a ~80,000 person company without it. We don't just use it for auth, but all sorts of other useful information about employees (e.g. expertise, cost centres, calendar, training etc etc). There is no reason that a directory server cannot be congruent with Unix. It works with our HP-UX/Solaris infrastructure just fine. And after all, what is NIS/NIS+, if not a (albeit) directory server?
This article could have been talking about outsourcing in general, not just overseas. One thing I have learned over the years is that you don't outsource simply to save money. In some cases, it costs *more* money to outsource than to do things internally.
In my opinion, you outsource to gain expertise you don't otherwise have, focus on your core business or other sound commercial reasons. Reduced costs should be the last reason for doing this. I have seen far too many outsourcing contracts go bad as a result of a failure to factor in the appropriate costs (this is on the providor as well as customer sides). I'm not saying don't ever do it, but be aware of a few things.
One of the comments at the end of the article also raised a good point: intellectual property. Be careful about dealing with *any* outsourcing company whom you suspect might take your brilliant idea and sell it on the open market. The opportunity cost of this happening can be staggering.
Another often forgotten part is the opportunity cost of not having an internal staff who understand and are aligned with the goals of the organisation. That is, those high potential technical and management staff who add more value to the business with their ideas/techniques etc than they cost in terms of total compensation. Do you want to outsource those people? An outsourcing comany has only one goal: to maximise the amount of profit they make per contract. This is not a bad thing, but it may mean that their goals diverge from your own.
The IT market seems to be very cyclical when it comes to outsourcing. It happens to be in favour right now, but who knows if that will be the case in 5 years time.
Compared to the ASCI series put together by the US government, $5M is not a lot of money. Consider that a single Sun SF15k, IBM p or z series or HP SuperDome can cost easily this much (that's just one machine).
I think VT are getting a pretty good deal. Really large clusters/supercomputers *can* cost upwards of $30M, depending on the configuration.
As someone who like photography and technology, I have to say I like the pictures. A reviewer of the G5 put it best when he said that it had a "brutal austerity". I think these photos capture it beautifully.
What exactly do you mean by that comment?? Are you implying that we Aussies are somehow *all* racial bigots?
As other posters have alluded to, the most constructive answer is: "Sure we can do that, but it will take this much time and this much money". I was in your situation once and I dealt with it in a number of ways:
1, I had a white board, with days of the week and hours of the day in a matrix. I filled in available spots when people wanted work done. If someone complained, I invited them to find a free spot in the matrix and write their name in.
2, Make appointments with your users. That is, try and schedule the time required. I used to leave a bit of "fat" in there in case emergencies came up.
3, If somebody gives you a job that you think will take more time/effort than they think, send them and their boss an estimate and see if they still want to do it.
4, Try and put some sort of ticketing system in, where you can track who has been asking you what and how much time it took. The worst thing is to have no proof against people who complain about the service they are getting.
5, Try and build an infrastructure that requires next to no babysitting (easier said than done), which will free up more of your time to deal with one-off end user issues.
6, Send reports about what you have been up to to your boss and/or the big boss. It might get them on your side.
It's not really a matter of saying no, but more a case of managing expectations
Automatic pitch correction has been around for some time. I saw a variation of it over ten years ago with the Eventide Harmoniser series (as used by Steve Vai and others). The difference now is that they are a lot more sophisticated and it is harder to tell when one is being used.
Are they evil? I think they are really not a lot different from all the other tricks you might use on a vocalist's sound: compression (to fatten up the sound), EQ (to notch or boost particular frequencies), reverb (to give it ambience), detuning/chorusing (to thicken the sound) and harmonising (instant vocal harmonies). These techniques have been used live and in recording studios for decades and are considered fair game.
I personally think a crap vocalist who's in tune is still a crap vocalist. There's more than just perfect pitch. There's got to be a bit of personality and power in the voice too. A slightly off tune Freddie Mercury, Sarah McLachlan or Geoff Tate will still sound better than a perfectly in tune Britney or Kelly Osbourne (to my ears).
There's an old saying in the music business: crap gear in the hands of a talented professional will still sound better than great gear in the hands of a no talent.
This might sound like I am stating the bloody obvious, but it's true. I think there are several facets to good benchmarking (based on my own experiences and reading other reports)....
1, Choose a test/workload that is representative of what *you* will be doing. There is no point in looking at SPECINT200 if you are going to be running an I/O intensive application like a RDBMS. Try and run or study tests that are relevant to the intended use of the system/component you are benchmarking.
2, Take note of things like compiler flags etc. These are important in tests like SPEC, as your results can vary wildly according to things like optimisation level. Some compilers produce faster code on certain CPU families and not on others. This is a reason why a lot of vendors will build their own compilers and test with them (e.g. SGI, SUN, DECPAQ).
3, Look at the full disclosure notice in the benchmarks. Take a look at the system configuration used. This is particularly, IMHO, on tests like TPC-C. The score you see might be based on a really whacky config, like most of the figures at the top of the list. For example, look at the Proliant figure (709k) and look at the config: 32 x 8 way servers to run a single database. Then compare it to a 64-way SuperDome or 32-way p690. Which comfig makes more sense? For a database, I would likely go with the single system for simplicity's sake. On another application, maybe the cluster would make sense.
4, Compare apples to apples. This is the hardest part, as CPU's, OS's, I/O, Apps. Compilers etc etc all vary across platforms. I like to to try and compare one variable if possible. To take the TPC-C again, I try to compare DB against DB, Cluster against Cluster, SMP against SMP etc etc. There is nothing to be gained, IMHO, from comparing MS-SQL server in a cluster on Xeon with Win2k3 to Sybase on a SF15k running SPARC Solaris. How do you properly compare these two results? Maybe the solution would be to look at SQLServer on one system against another or Sybase vs Oracle on a similar Unix system.
5, YMMV. Benchmarks are only ever an indicator of performance, not a guarantee. I tell my customers this all the time. They represent a result with a particular system, data set, O/S, tuning settings etc etc at a point in time. Other people's results with a similar config might differ considerably.
I could go on forever, but the above are my 2c
We have had at least a couple of Sun representatives come and present to us with PowerBook's, so I wouldn't doubt that plenty of Sun employees have Apple systems. Makes sense when you think about it: they're not running Windows and have Unix underneath. I would think they'd integrate pretty nicely with the Sun internal systems and allow them to do cool things like run Keynote, and Photoshop (for the marketing group) etc etc
:)
If only OpenOffice/Staroffice would run in Aqua mode under OSX, everything would be just perfect
I work in the Oil&Gas business and we use Linux clusters (and in the past bloody large Sun, IBM and SGI systems) for seismic processing and reservoir simulation. These particular problems are DSP and FP intensive and also can require a fairly large amount of memory to run. They are exactly the kind of commercial workload either a supercomputer or cluster can chew on.
Some of our customers (I work for a company that writes the software, amongst other things) have upwards of 100TB of 3D Seismic they want to process. These jobs can take weeks or months to run. The simulation jobs can take days as well. Obviously having a big computer or tight cluster of lots of small ones will help decisions get made faster and/or more accurately.
There are other examples too: I met a gentleman who works for the lab that does crash simulation for Porsche, Audi and VW. Another example would be an ex-boss of mine who went to work for an engine manufacturer, who used a couple of SGIs to simulate the bore and stroke in a cylinder. The simulation took several weeks to run. They need large computers to do this too. So there is a market for these machines.
It gets down to what kind of problem you are trying to solve and how you are trying to solve it. Clusters built from common of the shelf components need an interconnect for dividing up the workload. For algorithms that don't require a lot of message passing between nodes, a relatively low bandwidth, relatively high latency interconnect like Gigabit Ethernet or (compared to something NUMA) Myrinet is sufficient. There comes a point, however, where the amount of message passing and/or the speed of the interconnect may flatten out performance. That is, simply adding more nodes to the problem yields little or no performance gain. You are kind of stuck when this happens, unless you replace the processors in the cluster with faster ones or someone comes along with a faster connector.
A supercomputer (and to a lesser extent large SMP systems, like the p690, E15k or SuperDome) is designed from the ground up with an extremely high bandwidth, low latency interconnect. This means that those problems requiring message passing between processors will execute a lot faster than the self made cluster (in theory anyway). In fact, you might find that such a system with lower speed CPU's will outperform a cluster of 3Ghz PC's, just because the machine is designed with "total performance" in mind. There are other smarts built into the system and the O/S is generally engineered specifically for the processor (e.g. Unicos), which is the opposite approach of Linux and *BSDs.
Which approach is right for solving *your* problem gets down to how much money you have, the algorithm you are using, the size of the problem set, your own expertise and a bunch of other factors. It's a bit like deciding whether to use a semi-trailer or a group of Subaru WRX's to move a house full of furniture from one place to another. Either way will work, but one is generally more optimal for a given problem set.
For a more thorough exploration of this topic, I would recommend the O'Reilly book: High Performance Computing by Severance and Dowd, which discusses these kinds of issues in great detail. It's a bit old these days, but much of the information is as valid now as it was then.
I have started doing Yoga *after* work and have already found that it helps me relax and unwind. It has other benefits, such as toning muscles and increasing flexibility. It's not for everybody, granted, but I gotten some benefit out of it.
Agreed.
All benchmarks are only useful up to a point. What really counts is how they run *your* apps. The Intel platform does a good job for single threaded uniprocessor tasks, as evidenced by things like SPECINT and SPECFP. However, the story in other benchmarks might be quite different (e.g. SPECINTRATE, SPECFPRATE, TPC etc etc). What it gets down to is that each configuration, compiler, optimisation flags etc are different enough to make real comparisons difficult. The other thing about benchmarks is that they only give you an *indication* of expected performance, not what it will really be like. The above factors will greatly influence how your workloads will run in the real world...
If I'm running a database, I want something like an SMP UltraSPARC, Power4, Alpha or Itanium2, whereas computational fluid dynamics might lend itself to a cluster of IA-32 boxes or G4/5 with Altivec.
The point is, every system has its sweet spot and Apple most definitely has theirs. I'm just glad to see a) that they have a 64-bit system out there and b) they can compete toe to toe with Intel again. This can't be a bad thing for the industry.
You're right, CD is superior to minidisc in terms of quality. A lot of consumers probably wouldn't be able to hear a huge difference. The MD and old Philips DCC were designed as cassette tape replacement, not usurping CD.
:)
It's not just minidiscs that use optical. My Fostex digital multitrack also has an optical in and out. They are a little passe these days, but 5-10 years ago, many pro or semi-pro audio gear had them. I guess enough of 'em still do
I can certainly understand a group of people wanting to keep their jobs. It's what us tech folks have been trying to do for the past couple of years, after all. Overall, I am not in favour of replacing human umpires with computerised systems. It's a *game* after all, played by people, not robots. We might as well get rid of the players while we're at it. I am in favour of it as a training aid, however. Anything that improves the overall standard of umpiring is a good thing. The game of cricket has the same issue with the video replay technology. The television networks now have all these whiz bang replays that show from every angle. What they show me, as a fan, is that the umpire actually gets it right 98% of the time. This is certainly a higher strike rate than most of the players out there. At least it's not as bad as the NFL, where you can *challenge* a call. What a load of crap!!! It slows down an already boring game.
I agree that Sun's new gear looks great. This is basically the reason why I am more hopeful now than before. I think they have responded to the Intel threat and have woken up to the realities of the market somewhat.
:)
I hope to have my hands on some of their new gear very soon. Our local Sun reps seem pretty confident that GeoFrame will perform well on these systems. Their pricing is competitive too.
Some of their new developments in the graphics area looked very nice. They even impressed an SGI bigot like myself
I think generalisations about people's abilities, based on factors like age, are rather silly.
It all comes down to the individuals involved. I know some top flight engineers and some shite ones. Neither group comes exclusively from a particular age group. It's a bit like engineering processes (I am thinking mainly of software here). The best process in the world will not get you any better result if the people implementing it are crap.
It's all about finding the best person for the job, giving them the best tools and letting them get on with it...old or young -- who cares?