May be ignorance on my part, but I disagree. I don't think SPEC is good for anything in the real world.
Now that I've primed the flames, I'll explain why I say that and maybe someone can point out my folly.
SPEC_int tests how fast the integer units can process data and how well the CPU and the rest of the computer architecture can keep these pipelines "stuffed" with operations.
SPEC_fp does the same for floating point operations.
In the real world, both types of operations are used. What happens to integer performance if tossing in a few floats causes bubbles to appear in the integer unit pipelines?
In the real world, what often determines processor speed isn't how fast a given processor unit is, but how well all the units can be kept busy.
So where are the SPEC tests that mix and match data and instructions?
If its actually that important, you shouldn't be depending on any third party benchmarks. The only way to know is to run your tasks on both platforms and compare.
Just a question. Is all your number crunching integer, floating point, or both? If both, independent int and float benchmarks may be completely useless.
Therein lies my problem with the venerable SPEC benchmarks. For most real-world applications its pretty much useless to test one functional unit at a time and draw any kind of conclusion about how this translates into the performance of the CPU as a whole.
I wouldn't, primarily because that would make the performance even less a good reflection of "eveeryday" use.
Of course the only benchmarks that matter to me is how fast it runs the apps I want to run, the way I want to run them (none of this lets hit integer performance for a while, now lets test fp).
The PPC is essentially a collaborative work betweein the AIM (Apple-IBM-Motorola) alliance. So, while Motorola calls its vector engine Altivec, IBM calls their implementation "instruction set compatible" (or something like that) and Apple just calls them both Velocity Engine.
You're right, every community has its crackpots. There aren't any more Mac Zeolots than any other kind, its just that its "in" to reinforce the idea. I'm sure there were many well-reasoned rebuttals sent in (I've seen a couple that were also posted elsewhere). Were they posted for all to see, or just the small percentage of flames?
As long as the speed isn't outrageous, I agree its the speed differential that kills. In fact I remember reading a newspaper article a few years ago (no idea where) that some cities were causing accidents by posting speed limits that were too low.
The crux of the argument was that most people will naturally drive at a relatively safe speed for any given set of road conditions. If the posted speed limit is significantly lower than this, some people will tend to drive at what feels like a safe speed, while others will be driving the speed limit. This leads to the speed differential you mentioned with the end result being more accidents.
OK, I'm missing the point. He said, "convicted monopolist."
Microsoft is a monopoly, they were conviceted of using their monopoly power illegally. Thus they are a convicted monopolist.
Being a mechanic is also not illegal, but if you're referring to a mechanic who was convicted of a crime, isn't it OK to refer to him/her as a convicted mechanic?
As bobsledbob said, I was primarily thinking of mass produced art, but also of unique artwork that hasn't reached the level of being a national/historical treasure.
For example, at one of the fairs around here my daughter talked me into getting one of those cheap, hand-drawn characatures (sp?) of her done. Although nobody told the artist, it wasn't really any good and nothing worth keeping or hanging up. Last I saw it, she gave it to her little brother to color.
I can buy a painting from any artist around ( if I have the $$$), take it home and alter it in any way I want (say paint clothes on the naked monkey). Then I can even resell it.
I can buy a book, take it home and do whatever I want with it (including writing in it, marking out parts or ripping out pages) and sell it to someone else if I want.
Why should motion pictures and music be any different?
Which, IMO, is why copyrights should be shorter. I find it simply arrogant that some "creative" individuals think that its impossible for someone to take part of something they've created and make something better -or just plain different.
For scientific works, we understand the necessity of a short patent life so that others can build on ideas for the benefit of society. We understand the necessity to refine and reuse IP in ways the originator may have never considered.
For some reason we think it less necessary to do the same for artistic works, which IMO is just plain wrong.
Seriously, I don't know of any home DVD players that can do what you're talking about without major modifications. There's no way to program any of my DVD players other than by hand, and you lose that as soon as you pop in another DVD.
No, they don't contradict. The equation is essentially (leaving out the geometry):
Frictional Force = (coefficient of friction) x weight
Thus if you decrease the coefficient of friction, you decrease frictional force.
WRT moving heavy objects, its not due to a "bodies in motion" effect, its the change in the coefficient of friction. You left out the part about "unless acted upon by an outside force" (i.e. friction). We're all familiar with the formula
Force = Mass x Acceleration
The force due to friction must be equal and opposite to the force you are pushing with, or the object will "accelerate" -either speed up or slow down (assuming mass is constant).
Stated another way, in order to keep the object moving at roughly a constant speed, you need to continue to apply a force equal to that of the frictional force. That force is less than what you need to initially get the object moving. The technical term for this is "stiction" and its the reason why tapping/thumping/hammering mechanical hardware that's stuck can sometimes get things moving.
Slashdot Mirror
May be ignorance on my part, but I disagree. I don't think SPEC is good for anything in the real world.
Now that I've primed the flames, I'll explain why I say that and maybe someone can point out my folly.
SPEC_int tests how fast the integer units can process data and how well the CPU and the rest of the computer architecture can keep these pipelines "stuffed" with operations.
SPEC_fp does the same for floating point operations.
In the real world, both types of operations are used. What happens to integer performance if tossing in a few floats causes bubbles to appear in the integer unit pipelines?
In the real world, what often determines processor speed isn't how fast a given processor unit is, but how well all the units can be kept busy.
So where are the SPEC tests that mix and match data and instructions?
If its actually that important, you shouldn't be depending on any third party benchmarks. The only way to know is to run your tasks on both platforms and compare.
Just a question. Is all your number crunching integer, floating point, or both? If both, independent int and float benchmarks may be completely useless.
Therein lies my problem with the venerable SPEC benchmarks. For most real-world applications its pretty much useless to test one functional unit at a time and draw any kind of conclusion about how this translates into the performance of the CPU as a whole.
And if they hadn't disabled those "features", they would have been slammed for not having done so.
We're talking Apple. They're damned if they do, damned if they don't.
I wouldn't, primarily because that would make the performance even less a good reflection of "eveeryday" use.
Of course the only benchmarks that matter to me is how fast it runs the apps I want to run, the way I want to run them (none of this lets hit integer performance for a while, now lets test fp).
Why not just compare raw benchmarks then? IIRC IBM and Intel have both published their benchmarks for their respective processors.
The PPC is essentially a collaborative work betweein the AIM (Apple-IBM-Motorola) alliance. So, while Motorola calls its vector engine Altivec, IBM calls their implementation "instruction set compatible" (or something like that) and Apple just calls them both Velocity Engine.
You're right, every community has its crackpots. There aren't any more Mac Zeolots than any other kind, its just that its "in" to reinforce the idea. I'm sure there were many well-reasoned rebuttals sent in (I've seen a couple that were also posted elsewhere). Were they posted for all to see, or just the small percentage of flames?
WRT SSE2, they also didn't use Altivec on the G5 did they?
Also, (don't know for sure) isnt SSE2 faster partly because it reduces the precision? That really wouldn't be a "fair" test (whatever "fair" is).
Or maybe its the not stretching that distorts his color so much?
As long as the speed isn't outrageous, I agree its the speed differential that kills. In fact I remember reading a newspaper article a few years ago (no idea where) that some cities were causing accidents by posting speed limits that were too low.
The crux of the argument was that most people will naturally drive at a relatively safe speed for any given set of road conditions. If the posted speed limit is significantly lower than this, some people will tend to drive at what feels like a safe speed, while others will be driving the speed limit. This leads to the speed differential you mentioned with the end result being more accidents.
OK, you call someone who commits a crime a criminal.
What do you call someone who is "civilly liable"? (Other than broke)
But what if instead of $100, it was $1,000? How about $100,000?
Would you still be speeding?
We're just lucky there's not an anti-speeding lobby equivalent to the entertainment industry lobby.
The ironic part is when you think about possible consequences. Potentially losing a few$ on sales vs. petentially losing a life.
I thought they were going for life+70.
OK, IANAL, so help me out here.
There is criminal law and there is civil law.
Both contain the word "law".
So in either case, its breaking the law.
Where am I going wrong?
Better do it fast!
Seems Apple's stock always drops after a big event. The bigger the announcement, the bigger the drop.
Let's see. Powerbook.
...
Want 12, 15 or 17" screen?
Combo drive or SuperDrive?
256, 512 or 1000 MB of RAM?
WiFi or no?
Bluetooth or no?
What were those non-existant options again?
OK, I'm missing the point. He said, "convicted monopolist."
Microsoft is a monopoly, they were conviceted of using their monopoly power illegally. Thus they are a convicted monopolist.
Being a mechanic is also not illegal, but if you're referring to a mechanic who was convicted of a crime, isn't it OK to refer to him/her as a convicted mechanic?
Sitting here crying my heart out.
I just found out I'm not a real engineer.
Guess I'll have to find a new job.
Of course we all understand performance doesn't equal raw speed.
As bobsledbob said, I was primarily thinking of mass produced art, but also of unique artwork that hasn't reached the level of being a national/historical treasure.
For example, at one of the fairs around here my daughter talked me into getting one of those cheap, hand-drawn characatures (sp?) of her done. Although nobody told the artist, it wasn't really any good and nothing worth keeping or hanging up. Last I saw it, she gave it to her little brother to color.
Just some food for thought.
I can buy a painting from any artist around ( if I have the $$$), take it home and alter it in any way I want (say paint clothes on the naked monkey). Then I can even resell it.
I can buy a book, take it home and do whatever I want with it (including writing in it, marking out parts or ripping out pages) and sell it to someone else if I want.
Why should motion pictures and music be any different?
Which, IMO, is why copyrights should be shorter. I find it simply arrogant that some "creative" individuals think that its impossible for someone to take part of something they've created and make something better -or just plain different.
For scientific works, we understand the necessity of a short patent life so that others can build on ideas for the benefit of society. We understand the necessity to refine and reuse IP in ways the originator may have never considered.
For some reason we think it less necessary to do the same for artistic works, which IMO is just plain wrong.
... and a player that can read your notepad.
Seriously, I don't know of any home DVD players that can do what you're talking about without major modifications. There's no way to program any of my DVD players other than by hand, and you lose that as soon as you pop in another DVD.
No, they don't contradict. The equation is essentially (leaving out the geometry):
Frictional Force = (coefficient of friction) x weight
Thus if you decrease the coefficient of friction, you decrease frictional force.
WRT moving heavy objects, its not due to a "bodies in motion" effect, its the change in the coefficient of friction. You left out the part about "unless acted upon by an outside force" (i.e. friction). We're all familiar with the formula
Force = Mass x Acceleration
The force due to friction must be equal and opposite to the force you are pushing with, or the object will "accelerate" -either speed up or slow down (assuming mass is constant).
Stated another way, in order to keep the object moving at roughly a constant speed, you need to continue to apply a force equal to that of the frictional force. That force is less than what you need to initially get the object moving. The technical term for this is "stiction" and its the reason why tapping/thumping/hammering mechanical hardware that's stuck can sometimes get things moving.
What seems to be forgotten is that we've been genetically modifying pets (and other animals and plants) for centuries.