Yes, but the point is that most developers will change the value of x anyway.
That wouldn't help. The optimization is applied only to the benchmark program. In this case, x represents the direction the camera is facing at a particular time. In a game, this is unpredictable and non-optimizable. In the demo, it was set. The optimization does not translate to any gains in any game, no matter what the game developers do.
They optimized for the equivalent of Q3 Level 1 using player model X on a sunny day with a BFG and 13 bots, none of which have any random behavior, and the player follows a pixel-perfect path with no deviation and doesn't look around at all. This optimization will NOT translate to the improvement of the speed of ANY games.
It's not like saying 1+1 = 3. It's more like saying what's 7+7+7+7+7+7? Well, it's the same as 7*6, but guess which one is faster to calculate?
It's more like saying "What's 7+7+7+x+7+7?" For the benchmark program, x happens to be 7, leading to 7*6, but the general case is actually 7*5+x. No attempt is made to check an arbitrary game to see if x is 7. It only applies the optimization if the executable is a particular benchmark.
I REMOVED THE RE: IN THE SUBJECT. THERE IS A POSSIBLE SPOILER BELOW:
I heard an internet rumor that Cypher may return in the third one (Grains of salt at the ready). Perhaps there is an 'afterlife' Matrix to hold people who die before their body does. People stolen by agents could go there, too.
Downloading a program into a brain doesn't make the brain electronic. However, this raises the question of the implants. If I understand it corectly, an EMP induces high voltage (or amperage?) current through any nearby circuits, melting them. Wouldn't this fry the implants in all of the people? I guess they might not be powered...
1. Why lose the 1% energy/processing power/whatever if you can easily keep them in a second matrix? Especially if you can study them in order to improve the primary matrix.
2. I don't think that the Matrix people are, or should be, paying any attention to what was done in MIB one way or the other.
Buffy is not longer than an hour per week. But to get full enjoyment, you have to watch the show for several weeks. There are a lot of shows like that (Buffy, Angel, 24, Farscape, Alias, Smallville and Babylon 5 are some that I like). Of course, the majority of what's on TV is short, mindless drivel, but does that mean that the good stuff should be cancelled?
How many shows longer then an hour are on TV that are not live or talk shows?
Not many, but when they happen, it is for ratings purposes, and usually during sweeps. They are called miniseries, and are generally much harder to manage than a show like Buffy, since there may be 4-10 hours to watch in a single week. That's when the "not enough free time" issue comes into play much more than getting into Buffy or 24.
Should all TV shows be short, mindless entertainment and all books should be long and require deep introspection? Is there no place for long stories on TV or short stories in books?
Hamlet made his share of bad puns and quips while contemplating suicide and murder. What's your point? The show has its own language and style. The cookie analogy was appropriate for the characters and situation.
To clarify: Both "the Constitution" and "or laws" apply to "of any State". So neither state laws nor state constitutions have precedence over federal treaties, but the US Constitution does.
Taxcut is NOT just as easy to use. I switched to it because of the DRM, and it was torture. The interface is awful if you are doing anything even slightly outside the expected path, and there is extremely poor communication between the Federal and State products. Changes made to a box are not always automatically reflected in boxes that rely on that box.
I am quite happy that Intuit has caved. The free market occasionally does work.
As a programmer, I develop a test plan before I even start writing code. This is similar to someone giving me a requirement, and then changing the requirement after I've built a test plan and developed code toward that test. . . it's not really fair to the driver developers.
When you develop your test plan, do you include particular inputs to test? And when you develop code, do you check if the input is one of those values? If so, do you bypass the calculation of the output and instead check a pregenerated table?
Of course, putting common inputs into a hash table to improve performance is a perfectly valid practice. However, benchmarks are NOT common inputs for Nvidia's customers. The vast majority of customers will never run a benchmark. Therefore, this "performance enhancement" is clearly aimed only at reviewers, in order to trick customers into thinking they are getting more for their money, if ExtremeTech's suspicions are correct.
Note that the article is not saying that this is what is happening. They are just showing the results, which are consistent with this possibility. Nvidia is "currently investigating this issue, and will be letting us know what they discover", according to the article.
Slashdot Mirror
Yes, but the point is that most developers will change the value of x anyway.
That wouldn't help. The optimization is applied only to the benchmark program. In this case, x represents the direction the camera is facing at a particular time. In a game, this is unpredictable and non-optimizable. In the demo, it was set. The optimization does not translate to any gains in any game, no matter what the game developers do.
They optimized for the equivalent of Q3 Level 1 using player model X on a sunny day with a BFG and 13 bots, none of which have any random behavior, and the player follows a pixel-perfect path with no deviation and doesn't look around at all. This optimization will NOT translate to the improvement of the speed of ANY games.
It's not like saying 1+1 = 3. It's more like saying what's 7+7+7+7+7+7? Well, it's the same as 7*6, but guess which one is faster to calculate?
It's more like saying "What's 7+7+7+x+7+7?" For the benchmark program, x happens to be 7, leading to 7*6, but the general case is actually 7*5+x. No attempt is made to check an arbitrary game to see if x is 7. It only applies the optimization if the executable is a particular benchmark.
"Your mother was a hamster and your father smelt of elderberries."
Eeeeewwwww.
I REMOVED THE RE: IN THE SUBJECT. THERE IS A POSSIBLE SPOILER BELOW:
I heard an internet rumor that Cypher may return in the third one (Grains of salt at the ready). Perhaps there is an 'afterlife' Matrix to hold people who die before their body does. People stolen by agents could go there, too.
Downloading a program into a brain doesn't make the brain electronic. However, this raises the question of the implants. If I understand it corectly, an EMP induces high voltage (or amperage?) current through any nearby circuits, melting them. Wouldn't this fry the implants in all of the people? I guess they might not be powered...
The Architect says that 1% (or .1%) of humans reject the Matrix. If the human population is reasonably large, it's easy enough to get to 250,000.
1. Why lose the 1% energy/processing power/whatever if you can easily keep them in a second matrix? Especially if you can study them in order to improve the primary matrix.
2. I don't think that the Matrix people are, or should be, paying any attention to what was done in MIB one way or the other.
Buffy is not longer than an hour per week. But to get full enjoyment, you have to watch the show for several weeks. There are a lot of shows like that (Buffy, Angel, 24, Farscape, Alias, Smallville and Babylon 5 are some that I like). Of course, the majority of what's on TV is short, mindless drivel, but does that mean that the good stuff should be cancelled?
How many shows longer then an hour are on TV that are not live or talk shows?
Not many, but when they happen, it is for ratings purposes, and usually during sweeps. They are called miniseries, and are generally much harder to manage than a show like Buffy, since there may be 4-10 hours to watch in a single week. That's when the "not enough free time" issue comes into play much more than getting into Buffy or 24.
Ross? Isn't he Aiesha Tyler's boyfriend or something like that?
Not an XOR?
Charmed is a fetish show. Best not to watch it with your wife unless she's into that kind of thing.
Should all TV shows be short, mindless entertainment and all books should be long and require deep introspection? Is there no place for long stories on TV or short stories in books?
Do you usually get horny watching an action movie?
Where've you been? It's on about 15 times a week, at least in DC. Fox, WB, and FX are all showing several episodes a week.
Hamlet made his share of bad puns and quips while contemplating suicide and murder. What's your point? The show has its own language and style. The cookie analogy was appropriate for the characters and situation.
Um, he's already got Angel.
Are you kidding? What hair he didn't lose turned white!
Oh. you mean artistically.
file neo
neo: MATRIX V7.0 core dump of kernel
To clarify: Both "the Constitution" and "or laws" apply to "of any State". So neither state laws nor state constitutions have precedence over federal treaties, but the US Constitution does.
That angle was added to the advertising after the DRM flap.
Taxcut is NOT just as easy to use. I switched to it because of the DRM, and it was torture. The interface is awful if you are doing anything even slightly outside the expected path, and there is extremely poor communication between the Federal and State products. Changes made to a box are not always automatically reflected in boxes that rely on that box.
I am quite happy that Intuit has caved. The free market occasionally does work.
As a programmer, I develop a test plan before I even start writing code. This is similar to someone giving me a requirement, and then changing the requirement after I've built a test plan and developed code toward that test. . . it's not really fair to the driver developers.
When you develop your test plan, do you include particular inputs to test? And when you develop code, do you check if the input is one of those values? If so, do you bypass the calculation of the output and instead check a pregenerated table?
Of course, putting common inputs into a hash table to improve performance is a perfectly valid practice. However, benchmarks are NOT common inputs for Nvidia's customers. The vast majority of customers will never run a benchmark. Therefore, this "performance enhancement" is clearly aimed only at reviewers, in order to trick customers into thinking they are getting more for their money, if ExtremeTech's suspicions are correct. Note that the article is not saying that this is what is happening. They are just showing the results, which are consistent with this possibility. Nvidia is "currently investigating this issue, and will be letting us know what they discover", according to the article.
Actually, in order to fully test the new card features, we need to get some BOOM ]I[ benchmarks.