Scalability. Runs on a calculator, even on your brain.
No need to upgrade benchmarks every year.
Compiler independent. No easy "cheating" possible.
Speed dependent on problem size is calculated.
Different processors have different cache sizes and
perform by a factor of 10 or more differently on
different problem sizes.
You can check every result, Hint is Open Source and
free.
As most other benchmarks fulfill none of the design goals above, almost all benchmarks are nearly useless.
HINT or Hierarchical INTegration is a computer benchmarking tool developed at the Scalable Computing Laboratory (SCL) of Ames Laboratory, and is funded by the Office of Scientific Computing, U.S. Department of Energy (DOE). Unlike traditional benchmarks, HINT neither fixes the size of the problem nor the calculation time and instead uses a measure called QUIPS (QUality Improvement Per Second).
This enables HINT to display the speed for a given machine specification and problem size. Computers typically start up fast and slow down as they run out of fast memory and start using the main memory, or slow down even more if they have to access the disk. Such changes are easily visible with HINT generated data.
HINT is scalable and easily portable for a variety of architectures. It can be run on anything from a programmable calculator to a supercomputer.
It's not the costs for packaging, shipping and handling. It' the marketing costs only that account for the $10 or $15 for a CD.
From the FAQ of the shop's website:
I've never heard of these artists. Who are they?
FightForRock.com only deals with underground and independant artists. Many of them have probably never performed in your area. The music from these artists is usually not found in large corporate music stores. Usually, the music form these artists is only for sale in specialty shops, at a performance or directly from the artist.
Now you know why marketing is responsible for the costs. The quality of the music does not play any roll at all. If you want to reduce prices, you'll have to do less marketing. Now, you could try to participate from the marketing that already has been done by letting the customers search for the singles they want. What singles do the customers want? Yes, the ones they have heard of, and yes, these are the ones that marketing has pushed. So, you would have to have a large collection of widely known individual singles, as a lot of the other posters suggest.
But the big labels will not give away their marketing efforts for free and thus you will not be allowed to sell widely known singles, except if you pay the $10 for a CD or the $3 for a single (not accounting for the fact that the big labels do not want to support high-quality MP3s floating around the internet). Therefore, the website's approach of reducing marketing and giving away albums of unknown artists for a low price is not this bad. At prices as low as $3 to $5, shipping and packaging comes into play again and the MP3 approach might be feasible.
Here is a suggestion for currency that can not be counterfeited:
The idea is to create banknotes that can not be reproduced, by no
one, be it the original producer or counterfeiters. To achieve that,
take colorless lacquer with tiny pieces of reflecting foil in
it. Print a strip onto each banknote with that lacquer. Make a
digital photograph of the strip and store the photo together with
the serial number of the bankote.
No banknote produced that way can be reproduced, because for doing
so,
one would have to
arrange hundreds or thousands of pieces of reflecting foil such that
they are oriented in the very same way as in the original, so that they
reflect light in the same way.
To check whether a banknote is valid, you'd just have to go to the
website of the national bank, type in the serial number, look at
the stored photo you get back, and compare it to the banknote you
have got. Of
course the check could be automated with the help of a machine that has
an online connection. To make sure you do not get faked photos, you
would just have to use a secure (SSL) connection or check the
cryptographic signatures of the photos you get back from the site.
The method above would be feasible and relatively cheap. It would
take just two more passes in production that can both be
accomplished easily by a machine: printing a strip of lacquer onto
the banknote and taking a photo with a high resolution digital
camera. The costs for storage of the photos would be far less than
the value of the bill and even less than the other production costs,
I assume (about 1 cent for a 100 KB JPEG image). Retrieval of the
photos would be trivial, because they would only have to be accessed via
the serial number of the corresponding banknote.
Slashdot Mirror
Hint is designed to fulfill the following goals.
As most other benchmarks fulfill none of the design goals above, almost all benchmarks are nearly useless.
What is HINT?It's not the costs for packaging, shipping and handling. It' the marketing costs only that account for the $10 or $15 for a CD.
From the FAQ of the shop's website:
Now you know why marketing is responsible for the costs. The quality of the music does not play any roll at all. If you want to reduce prices, you'll have to do less marketing. Now, you could try to participate from the marketing that already has been done by letting the customers search for the singles they want. What singles do the customers want? Yes, the ones they have heard of, and yes, these are the ones that marketing has pushed. So, you would have to have a large collection of widely known individual singles, as a lot of the other posters suggest.
But the big labels will not give away their marketing efforts for free and thus you will not be allowed to sell widely known singles, except if you pay the $10 for a CD or the $3 for a single (not accounting for the fact that the big labels do not want to support high-quality MP3s floating around the internet). Therefore, the website's approach of reducing marketing and giving away albums of unknown artists for a low price is not this bad. At prices as low as $3 to $5, shipping and packaging comes into play again and the MP3 approach might be feasible.
The idea is to create banknotes that can not be reproduced, by no one, be it the original producer or counterfeiters. To achieve that, take colorless lacquer with tiny pieces of reflecting foil in it. Print a strip onto each banknote with that lacquer. Make a digital photograph of the strip and store the photo together with the serial number of the bankote.
No banknote produced that way can be reproduced, because for doing so, one would have to arrange hundreds or thousands of pieces of reflecting foil such that they are oriented in the very same way as in the original, so that they reflect light in the same way.
To check whether a banknote is valid, you'd just have to go to the website of the national bank, type in the serial number, look at the stored photo you get back, and compare it to the banknote you have got. Of course the check could be automated with the help of a machine that has an online connection. To make sure you do not get faked photos, you would just have to use a secure (SSL) connection or check the cryptographic signatures of the photos you get back from the site.
The method above would be feasible and relatively cheap. It would take just two more passes in production that can both be accomplished easily by a machine: printing a strip of lacquer onto the banknote and taking a photo with a high resolution digital camera. The costs for storage of the photos would be far less than the value of the bill and even less than the other production costs, I assume (about 1 cent for a 100 KB JPEG image). Retrieval of the photos would be trivial, because they would only have to be accessed via the serial number of the corresponding banknote.