Running top with a 10 second sampling window my iTunes takes about 4% CPU time while playing (currently MP3 encoded songs) out a pair of USB connected speakers. If I stop it playing iTunes drops down to about 0.5%, the window is still scrolling song text. If I hide iTunes it drops down to basically 0%.
MSWord chews far more CPU doing nothing then iTunes does playing music on this system.
...but its not a progress bar but a meter of some type (bandwidth or audio?)...
Don't forget that even if the screen shots are real, they are of an OS that won't be released for many more months. Apple has been known to change things, sometimes a lot, (add/drop/polish features) in the final couple of months before release.
This joke, if is one, will get this guy in a WORLD of hurt and I am not just talking about legally... you know Steve might just jump in his jet and pay him a visit.
First off... the USB 2.0... is that "Full Speed", or "High Speed" USB?
Most defiantly "High Speed" as in 480 Mbits second.
Secondly, how would a 64Bit 2Ghz (with a 1Ghz System Bus) compare to an AMD, or an Intel chip running at 2Ghz, or what would it be equal to (Can you call the 2Ghz G5 a G5 3000+?
Hard to answer, it depends on what you are doing.
For integer performance the PPC 970 should be around 1.3-1.5x clock for clock against a P4. For floating point it should be around 1.4-1.6x clock for clock against a P4. For SIMD (vector calculations) it could well into 2+x clock for clock (don't have solid numbers for this). This is based on SPECint2000 & SPECfp2000 published/estimated numbers and assume linear scaling with clock speed, neither of which is true in general.
In other words a PPC 970 @ 2GHz aligns approx. with a P4 @ 3GHz for int and floating point and a P4 @ 4+GHz for SIMD.
Also the PPC 970 can move a little more data over its FSB then a P4 can at the moment (assuming the PPC 970 @ 2GHz is true). P4s @ 3GHz can move around 6.4GB/s and PPC 970 @ 2GHz can move around 7.1GB/s (for the later I am accounting for overhead, not sure if the P4 numbers are doing the same).
Anyway we will have to see how things do once they get out into the real world. At the least the new PowerMacs should be on par with P4 based systems.
And the PCI-X... again, I'm going to assume that this is PCI Express?
Nope not PCI Express but most likely the existing PCI-X standard (133MHz/64bit).
Could this also mean the announcement of a x86 version of Mac OS X?
Very unlikely. Most likely he will show off 64 bit support, PPC 970 optimizations, HyperTransport support, evolutionary extensions to the audio sub-system, evolutionary (possibly revolutionary) extensions to Quartz 2D / QuartzExtreme, and a few new things that we havn't really thought of, etc.
Apple's high-end systems are generally targeted towards the EXACT folks who would be doing stuff like this. Why shouldn't they provide systems like this.
I guess Dell/HP/others shouldn't bother making multi-processor system at all then...
Anyway, my point is still 100% true (go reread what I wrote).
But hey, its apple, so its inherently good, right?
I guess labeling me (without knowing who I am and what I do/think) makes me wrong, I guess.
The PPC 970 bus is not HyperTransport (based on released information from IBM). It can however be easily interfaced to HT via a bridge chip. I do believe that the memory bus will not be HT but more direct, the rest will utilize HT as needed... we shall see.
Serial ATA is still faster then a single drive can pump out, it is a none issue at the moment. If you look at Apple's hardware they have generally favored slightly slower but dedicated buses, ones for each drive.
Dual prcessors arent cumulative, but thanks for playing!
Not true... Multiple processors can be cumulative for many operations and sometime even better then cumulative.
For example many image, video and audio tasks can be split up and worked on in pieces by multiple CPUs (or cluster nodes) in such a way that little to no dependencies happen between the threads. Each CPU could for example get a part of frame or every other frame. This can result in a cumulative scaling of performance.
In the best case you get better then cumulative because you can avoiding some thread swapping.
Actually the software in question is free since his website is not for commercial purposes. The apparent violation is that of license agreement; which requires registration and an in source link.
Free software is most often not "free" in the sense it often comes with purposely restrictive licensing agreements and this is an example of that.
(not saying free / open / etc. is good or bad, just pointing out the facts)
The FSB is only limiting for certain usage scenarios and in many cases the G4's backside cache can cover FSB issues.
Anyway FSB would only affect SMP if the processes you are running are stressing the FSB, which compilation wouldn't really do (it is disk bound mostly).
I would love to see a 970 with a large backside cache... but it would be overkill given its FSB support.
I love this... already pulled down 4 albums I have been meaning to buy for a while now, it only took me about 3 mins to sign on, find the albums, and buy them.
Downloading them as fast as my DSL line can go, I may spend myself into the red;-)
I get what you are trying to say now... your wording had me confused. You didn't intend to mix 1 mph and aerodynamic flight per say... you could use just enough power from say a rocket to keep you moving at 1 mph vertically as you gained altitude.
Unless you injected a rotational force (torque) you could not reach and maintain orbit (ignoring other gravitational sources).
It also depends on how far out you want to orbit... anyway back to work.
Escape velocity is only high if you are using ballistic flight. Flying under continuous power, you can go 1 mph and still make it to orbit, provided you can sustain that speed for long enough.
I think you mean you can accelerate slowly and still reach orbit if you have other means of avoiding chrashing back down to earth, which you can do in the case of aerodynamic flight. You do need enough propulsive force to over come aerodynamic drag and energy used in aerodynamic lift, this force would vary during your flight.
If you fly at a fixed velocity (1 mph) you could in theory climb (assuming you could maintain lift as you climbed) using aerodynamic lift to the edge of space, you could not however simply climb a latter out into space and be in _orbit_. You would not have enough angular velocity to over come gravity (you lose your aerodynamic lift out in space). You could not remain in orbit as a result, especially at 1 mph.
In other words the object you wanted to orbit would need a massive boost in angular velocity to remain in orbit.
If they have no over ground velocity then they are in geo-synchronous _orbit_, orbiting the earth in sync with the spin of the earths surface. In other words the have an angular velocity around the center of the earth at their orbital radius that matches the angular velocity of the surface of the earth.
Think of a ball (with a decent amount of mass, so you don't get dizzy) on the end of a rubber band with you holding the other end. If you spin around in a fixed spot the ball will sweep a circle around you. The faster you spin the farther the ball moves away (assuming the rubber band stretches) from you because you are giving the ball more energy, its moves faster rotationally because you are. If it wasn't for the rubber band the ball would fly in a line away from you, let go of it as you spin to see that.
This is similar to what happens with satellites (things that orbit a body because of gravity and velocity), at least in the case of circular orbits. The rubber band is gravity, the ball is the satellite and you are the earth with a rotational velocity, in the case of the earth a fixed velocity (at least fixed enough for this discussion). The main difference is that rubber bands pull harder the more you stretch them but the pull of gravity lesses as you get farther away (important detail to keep in mind that affects the radius of orbit). Also obviously the rubber band exerts a "rotational" force that gravity cannot but rockets can when something is launched into orbit.
Anyway... satellites do what they do because their angular velocity "generates a force" directed away from the earth (the ball wanting to fly away from you) that equals the force that the earths gravity puts on the satellite (the rubber band), a force that is directed at the earth. If these forces are balanced then the object will orbit the earth, it won't fly away and it won't crash down on your head.
So the closer you orbit the earth the more gravity you feel and hence the faster you have to orbit in order to offset gravity (low earth orbit objects circles the earth faster then the earths surfaces spins). As you move farther out gravity lessens a hence you can rotate slower. At some distance out form the earth the rotational speed of the satellite will match with the apparent rotational speed of the surface of the earth and those are called geo-synchronous satellites (this distance depends on the mass of the object).
Using the word "chose" or "choice" doesn't imply that some intent is in life over that that happens by random changes.
Generally it can be said about every "feature" of life that at some point in the past a choice was made by a series of random happen stances and confining conditions, which gave one solution or a small set of solution an advantage to win out over others.
No external hand need to be implied when using the word "choice".
---
Statically some would escape or resist the "toxic soup" as a result of interaction with others. They would accumulate over time allowing them to become potential building blocks.
This doesn't fully match with my experience and I work for HP in their storage software division... fortune 50, 100 and many 500 are using and or starting to setup storage networks based on Fibre Channel (not much iSCSI yet).
Most of the major data centers in the world are using FC interconnects and large storage arrays (5+ TB per array, the bigger (in capacity not physical size) the better generally). The trend is towards putting the data in storage arrays external from the servers, think about blade servers for example.
So in other words toward using storage networks regardless if they are based on FC interconnects, iSCSI, etc.
---
Also not what your point is about T3 and 5TB replicated storage... Many companies do that with far larger amounts of storage using metro fiber not T3s. This is not an Fibre Channel issue but a data desaster recover issue.
I got a almost new quality 16 port Brocade 2800 switch of off eBay, fully load with GBICs for around 4k, add on about 400-600 per host for adapters, and 5k-11k for something like Apple's xRAID you can get into a FC SAN for much less then 50k.
If you want 2gig fibre channel it would cost you of course more for the switch (2-3x currently).
Not that it is a cost effect thing to do for small SANs... iSCSI isn't that cheap either but in theory you do save on at least the switch costs (most companies are making iSCSI adapters instead of normal NICs for performance reasons).
I may start on a driver for QLogic's iSCSI cards
on
iSCSI for Mac OS X?
·
· Score: 1
I am about to finish work on a Fibre Channel driver for Mac OS X and QLogic's SANBlade family of adapters. Just waiting on my Apple xRAID to show up so I can finish testing (my current test array is toast, aka expensive brick).
Anyway I have been looking at writing or extending the driver to support QLogic's iSCSI capable adapters.
Slashdot Mirror
Not sure I see what you are talking about...
Running top with a 10 second sampling window my iTunes takes about 4% CPU time while playing (currently MP3 encoded songs) out a pair of USB connected speakers. If I stop it playing iTunes drops down to about 0.5%, the window is still scrolling song text. If I hide iTunes it drops down to basically 0%.
MSWord chews far more CPU doing nothing then iTunes does playing music on this system.
What system (CPU) do you have?
...but its not a progress bar but a meter of some type (bandwidth or audio?)...
Don't forget that even if the screen shots are real, they are of an OS that won't be released for many more months. Apple has been known to change things, sometimes a lot, (add/drop/polish features) in the final couple of months before release.
This joke, if is one, will get this guy in a WORLD of hurt and I am not just talking about legally... you know Steve might just jump in his jet and pay him a visit.
Hard to answer, it depends on what you are doing.
For integer performance the PPC 970 should be around 1.3-1.5x clock for clock against a P4. For floating point it should be around 1.4-1.6x clock for clock against a P4. For SIMD (vector calculations) it could well into 2+x clock for clock (don't have solid numbers for this). This is based on SPECint2000 & SPECfp2000 published/estimated numbers and assume linear scaling with clock speed, neither of which is true in general.
In other words a PPC 970 @ 2GHz aligns approx. with a P4 @ 3GHz for int and floating point and a P4 @ 4+GHz for SIMD.
Also the PPC 970 can move a little more data over its FSB then a P4 can at the moment (assuming the PPC 970 @ 2GHz is true). P4s @ 3GHz can move around 6.4GB/s and PPC 970 @ 2GHz can move around 7.1GB/s (for the later I am accounting for overhead, not sure if the P4 numbers are doing the same).
Anyway we will have to see how things do once they get out into the real world. At the least the new PowerMacs should be on par with P4 based systems.
Nope not PCI Express but most likely the existing PCI-X standard (133MHz/64bit).
Very unlikely. Most likely he will show off 64 bit support, PPC 970 optimizations, HyperTransport support, evolutionary extensions to the audio sub-system, evolutionary (possibly revolutionary) extensions to Quartz 2D / QuartzExtreme, and a few new things that we havn't really thought of, etc.
That's you and your desire... it doesn't change the reality of this instance.
I guess Dell/HP/others shouldn't bother making multi-processor system at all then...
Anyway, my point is still 100% true (go reread what I wrote).
I guess labeling me (without knowing who I am and what I do/think) makes me wrong, I guess.
The PPC 970 bus is not HyperTransport (based on released information from IBM). It can however be easily interfaced to HT via a bridge chip. I do believe that the memory bus will not be HT but more direct, the rest will utilize HT as needed... we shall see.
Serial ATA is still faster then a single drive can pump out, it is a none issue at the moment. If you look at Apple's hardware they have generally favored slightly slower but dedicated buses, ones for each drive.
For example many image, video and audio tasks can be split up and worked on in pieces by multiple CPUs (or cluster nodes) in such a way that little to no dependencies happen between the threads. Each CPU could for example get a part of frame or every other frame. This can result in a cumulative scaling of performance.
In the best case you get better then cumulative because you can avoiding some thread swapping.
Free software is most often not "free" in the sense it often comes with purposely restrictive licensing agreements and this is an example of that.
(not saying free / open / etc. is good or bad, just pointing out the facts)
The FSB is only limiting for certain usage scenarios and in many cases the G4's backside cache can cover FSB issues.
Anyway FSB would only affect SMP if the processes you are running are stressing the FSB, which compilation wouldn't really do (it is disk bound mostly).
I would love to see a 970 with a large backside cache... but it would be overkill given its FSB support.
100% agree... it is a great way to browse for music, especially if you are used to iTunes already.
I love this... already pulled down 4 albums I have been meaning to buy for a while now, it only took me about 3 mins to sign on, find the albums, and buy them.
;-)
Downloading them as fast as my DSL line can go, I may spend myself into the red
I get what you are trying to say now... your wording had me confused. You didn't intend to mix 1 mph and aerodynamic flight per say... you could use just enough power from say a rocket to keep you moving at 1 mph vertically as you gained altitude.
Unless you injected a rotational force (torque) you could not reach and maintain orbit (ignoring other gravitational sources).
It also depends on how far out you want to orbit... anyway back to work.
I think you mean you can accelerate slowly and still reach orbit if you have other means of avoiding chrashing back down to earth, which you can do in the case of aerodynamic flight. You do need enough propulsive force to over come aerodynamic drag and energy used in aerodynamic lift, this force would vary during your flight.
If you fly at a fixed velocity (1 mph) you could in theory climb (assuming you could maintain lift as you climbed) using aerodynamic lift to the edge of space, you could not however simply climb a latter out into space and be in _orbit_. You would not have enough angular velocity to over come gravity (you lose your aerodynamic lift out in space). You could not remain in orbit as a result, especially at 1 mph.
In other words the object you wanted to orbit would need a massive boost in angular velocity to remain in orbit.
If they have no over ground velocity then they are in geo-synchronous _orbit_, orbiting the earth in sync with the spin of the earths surface. In other words the have an angular velocity around the center of the earth at their orbital radius that matches the angular velocity of the surface of the earth.
Think of a ball (with a decent amount of mass, so you don't get dizzy) on the end of a rubber band with you holding the other end. If you spin around in a fixed spot the ball will sweep a circle around you. The faster you spin the farther the ball moves away (assuming the rubber band stretches) from you because you are giving the ball more energy, its moves faster rotationally because you are. If it wasn't for the rubber band the ball would fly in a line away from you, let go of it as you spin to see that.
This is similar to what happens with satellites (things that orbit a body because of gravity and velocity), at least in the case of circular orbits. The rubber band is gravity, the ball is the satellite and you are the earth with a rotational velocity, in the case of the earth a fixed velocity (at least fixed enough for this discussion). The main difference is that rubber bands pull harder the more you stretch them but the pull of gravity lesses as you get farther away (important detail to keep in mind that affects the radius of orbit). Also obviously the rubber band exerts a "rotational" force that gravity cannot but rockets can when something is launched into orbit.
Anyway... satellites do what they do because their angular velocity "generates a force" directed away from the earth (the ball wanting to fly away from you) that equals the force that the earths gravity puts on the satellite (the rubber band), a force that is directed at the earth. If these forces are balanced then the object will orbit the earth, it won't fly away and it won't crash down on your head.
So the closer you orbit the earth the more gravity you feel and hence the faster you have to orbit in order to offset gravity (low earth orbit objects circles the earth faster then the earths surfaces spins). As you move farther out gravity lessens a hence you can rotate slower. At some distance out form the earth the rotational speed of the satellite will match with the apparent rotational speed of the surface of the earth and those are called geo-synchronous satellites (this distance depends on the mass of the object).
No I don't.
So you can only charge for things that have a physical form?
Using the word "chose" or "choice" doesn't imply that some intent is in life over that that happens by random changes.
Generally it can be said about every "feature" of life that at some point in the past a choice was made by a series of random happen stances and confining conditions, which gave one solution or a small set of solution an advantage to win out over others.
No external hand need to be implied when using the word "choice".
---
Statically some would escape or resist the "toxic soup" as a result of interaction with others. They would accumulate over time allowing them to become potential building blocks.
This doesn't fully match with my experience and I work for HP in their storage software division... fortune 50, 100 and many 500 are using and or starting to setup storage networks based on Fibre Channel (not much iSCSI yet).
Most of the major data centers in the world are using FC interconnects and large storage arrays (5+ TB per array, the bigger (in capacity not physical size) the better generally). The trend is towards putting the data in storage arrays external from the servers, think about blade servers for example.
So in other words toward using storage networks regardless if they are based on FC interconnects, iSCSI, etc.
---
Also not what your point is about T3 and 5TB replicated storage... Many companies do that with far larger amounts of storage using metro fiber not T3s. This is not an Fibre Channel issue but a data desaster recover issue.
The disk drives, block serving NAS boxes, protocol bridges, disk arrays, etc. are your iSCSI servers.
You normally wouldn't run a iSCSI server on a server box, better to run some type of file server on those.
I got a almost new quality 16 port Brocade 2800 switch of off eBay, fully load with GBICs for around 4k, add on about 400-600 per host for adapters, and 5k-11k for something like Apple's xRAID you can get into a FC SAN for much less then 50k.
If you want 2gig fibre channel it would cost you of course more for the switch (2-3x currently).
Not that it is a cost effect thing to do for small SANs... iSCSI isn't that cheap either but in theory you do save on at least the switch costs (most companies are making iSCSI adapters instead of normal NICs for performance reasons).
I am about to finish work on a Fibre Channel driver for Mac OS X and QLogic's SANBlade family of adapters. Just waiting on my Apple xRAID to show up so I can finish testing (my current test array is toast, aka expensive brick).
Anyway I have been looking at writing or extending the driver to support QLogic's iSCSI capable adapters.
Hexley
Same with T1s, they still get rolled up some place.
The keys are the same shape and are made of the same material with little variation in the material itself from key to key.
That alone will make it likely that the keys will warp in very similar fashions when subjected to roughly the same amount heat.
The other damage doesn't look faked to me either...