Sidebar Sidebar
Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.

Will there be a new Mac Pro in 2015?

Status
Not open for further replies.
JesterJJZ said:
First I've heard of this. Dual CPUs is a must before I even give the nMP a second look.
Click to expand...

MacVidCards said:
Really, wow, did they hide the second socket under a sticker?
Click to expand...

I assume that the post meant that the chipset on the MP6,1 motherboard (that circular thing in the bottom) has pinouts to support a second E5-v2xxx.v2 processor.

Apple chose to castrate the capabilities of the chipset by putting it in a "pretty" tube without the power, cooling or space to support the chipset. (For example, 80 PCIe 3.0 lanes and 24 DIMM slots....)
 
mugwump said:
Oops, wrote that too quickly. What I meant is that there is an identical open space where a second socket should be, but it's empty. I assume it will get used when they can get more wattage power in there.
Click to expand...

You're thinking of the "open slot" where a 2nd SSD could go. This is a view of the nMP CPU:

pro.png
 
handsome pete said:
You're thinking of the "open slot" where a 2nd SSD could go. This is a view of the nMP CPU:
Click to expand...

Yeah, yor're right about that open slot. Unless they engineered it too small, a second socket should fit there. It's the 450watt power source that is the limitation. I would assume the next revision will remedy that since sky lake runs cooler on the smaller die.
 
mugwump said:
.... It's the 450watt power source that is the limitation. I would assume the next revision will remedy that since sky lake runs cooler on the smaller die.
Click to expand...

That isn't the only limitation. If they want the 2nd slot to be a 'mirror' in functionality to the 1st PCIe SSD slot the other , probably more major, limitation is that they are out of PCIe bandwidth. There are no lanes to provision a second slot with. They might be able to throw in a PCIe switch but then in a "rob Peter to pay Paul" situation of taking bandwidth away from the 1st SSD slot to 'give' to the second. Used in tandem you wouldn't get max speed out of either.


Power isn't so much a major issue. The OWC "replacement" drive bleeds some power off the GPU and pragmatically is a multi SSD implementation that merely presents as a single drive. ( multiple SSD controllers bound up behind a SATA RAID controller. ). if chop that in "half" ( toss the SATA raid controller for a PCIe SSD controller) and split the Flash packages over two PCIe SSD I don't think the power is going to spike much.


If not a slave to design symmetry the 2nd SSD could be hooked to the under leveraged SATA lanes inside the Mac Pro. That SSD wouldn't have the same "max bandwidth". It is also a bit of a bandwidth ballon squeeze in that the SATA subsystem and the chipset PCIe lanes all feed back to the memory+CPU along the same DMI link. ( DMI 2.0 20Gb/s ... x4 PCIe v2.0 20Gb/s ). As the SSD drives push right up to the max capacity of the x4 PCIe v2 things get completely saturated with just one drive. The 2nd drive would primarily just be for additional capacity, not speed ( hence it is asymmetrical. )

I don't see Apple backsliding into SATA SSD drives though. All the new Mac designs point to SSD requiring PCIe going forward. Standard Apple simplified logistics and design practice. Unless Apple wants to chuck one of the Thunderbolt controllers, there isn't bandwidth for a equally capable PCIe SSD in the next set of chipsets+CPUs coming from Intel in this Xeon E5 space.
 
Last edited:
mugwump said:
Yeah, yor're right about that open slot. Unless they engineered it too small, a second socket should fit there. It's the 450watt power source that is the limitation. I would assume the next revision will remedy that since sky lake runs cooler on the smaller die.
Click to expand...

It's not the power supply, but perhaps the thermals.

My new HP ProLiant DL380 Gen 9 servers have dual (N+1) 1400 watt power supplies, and they look like they could fit in an MP6,1. (see attachment - a photo of the 1400w 94% efficiency PS with a 12 oz Coke can sitting on it)

Of course the MP6,1 doesn't need a 240v only 1400 watt power supply - but if Apple borrowed a little innovation from HP they could get 600w to 700w in that space.
 

Attachments

  • 1400.jpg
    1400.jpg
    571.5 KB · Views: 91
AidenShaw said:
I
Of course the MP6,1 doesn't need a 240v only 1400 watt power supply - but if Apple borrowed a little innovation from HP they could get 600w to 700w in that space.
Click to expand...

At what decibel level? Wattage and 2D area isn't the only design constraint.

That powers supply is meant to be used in a room where generally band humans from staying in long term. The Mac Pro power supply is suppose to sit on top of someone's work desk.
 
deconstruct60 said:
At what decibel level? Wattage and 2D area isn't the only design constraint.

That powers supply is meant to be used in a room where generally band humans from staying in long term. The Mac Pro power supply is suppose to sit on top of someone's work desk.
Click to expand...

My point is that 450 watts isn't a hard ceiling. If the nnMP needed 525 watts, I'm sure that Apple could figure out how to do that without adding a turbo fan.

For starters, increasing the PS efficiency might be all that it takes - the MP6,1 supply is only 90% efficient, whereas the HP one above is 94%. You can get 96% HP supplies as well.
 
Last edited:
If Apple can put a song & dance light show into a wristwatch they could certainly improve on nMP. Obviously I still think nMP should be like iPhone 5 is now, their 2nd tiered computer.

The 7,1 Mac Pro could be an amazing machine rivaling the best of the best. Apple has the ability to make it be that. They have obviously decided that competing at the top isn't worth the trouble and have steered the line into the slow lane.

Very similar to how AMD has had to aim the R9 290X at the GTX970 as they have absolutely nothing to compete with 980/Titan-X.

"Pretty but hamstrung by compromises" is their direction, so expect more 2nd rate AMD chips and other choices dictated by fashion rather than function.
 
AidenShaw said:
For starters, increasing the PS efficiency might be all that it takes - the MP6,1 supply is only 90% efficient, whereas the HP one above is 94%. You can get 96% HP supplies as well.
Click to expand...

Just to elaborate, the math is quite interesting.

The 450 watt 90% power supply in the MP6,1 has the following characteristics:
  • it supplies a continuous 450w DC to the computer
  • it draws 500w of power from the mains
  • it produces 50w of heat to be dissipated

A 94% efficiency power supply like the HP in my picture that produced 50 watts of heat would:
  • supply 785w to the computer
  • draw 835.1 watts from the mains
  • produce 50.1 watts of heat

A 96% efficiency power supply that produced 50 watts of heat would:
  • supply 1200w to the computer
  • draw 1250 watts from the mains
  • produce 50 watts of heat

(Think about it - a 100% efficient power supply would supply infinite power and produce no heat.)

So, the issue is *not* cooling the power supply. Just increasing the efficiency of the power supply to what is easily available on the open market would nearly triple the available power.

So the question becomes, assuming that the vaunted "thermal core" could dissipate additional heat with more airflow, do you want:
  • an underpowered nnMP that is always quiet
  • a much faster nnMP that has audible airflow when pushed hard

And it doesn't even have to be an "either-or" question. The power setting control panel on my laptop let's me choose "reduce performance to stay quiet" or "spin the fans up for max performance".
 
Last edited:
AidenShaw said:
Just to elaborate, the math is quite interesting.

The 450 watt 90% power supply in the MP6,1 has the following characteristics:
  • it supplies a continuous 450w DC to the computer
  • it draws 500w of power from the mains
  • it produces 50w of heat to be dissipated

A 94% efficiency power supply like the HP in my picture that produced 50 watts of heat would:
  • supply 785w to the computer
  • draw 835.1 watts from the mains
  • produce 50.1 watts of heat

A 96% efficiency power supply that produced 50 watts of heat would:
  • supply 1200w to the computer
  • draw 1250 watts from the mains
  • produce 50 watts of heat

(Think about it - a 100% efficient power supply would supply infinite power and produce no heat.)

So, the issue is *not* cooling the power supply. Just increasing the efficiency of the power supply to what is easily available on the open market would nearly triple the available power.

So the question becomes, assuming that the vaunted "thermal core" could dissipate additional heat with more airflow, do you want:
  • an underpowered nnMP that is always quiet
  • a much faster nnMP that has audible airflow when pushed hard

And it doesn't even have to be an "either-or" question. The power setting control panel on my laptop let's me choose "reduce performance to stay quiet" or "spin the fans up for max performance".
Click to expand...

The only reason they are getting that small package to 1400watts with a high efficiency is because it is running at 240V. For power supplies, running at 240V vs 120V means the computer, or sever in the case of the PS you listed, will draw much less current. The much less current means much much less heat...
 
gman36 said:
The only reason they are getting that small package to 1400watts with a high efficiency is because it is running at 240V. For power supplies, running at 240V vs 120V means the computer, or sever in the case of the PS you listed, will draw much less current. The much less current means much much less heat...
Click to expand...

Sorry, but that's nonsense. Complete nonsense.

Heat (watts) is amps*voltage. Double the voltage - you halve the amps but have the same heat. *the same heat*

The big thing about higher efficiency power supplies is that the heat (for a given draw from the mains) comes more from the electronics and less from the PS itself. As I pointed out earlier, a 450w 90% power supply and a 785w 94% power supply generate about the same amount of heat from the power supply itself.

There is an advantage to 240v-only in that the amperage drops by half, so you have thinner copper wires and it can be a bit more compact due to that.

The big advantage to using higher voltage, though, is that 1400w is too much to run on a standard American 15 amp 110v circuit - so there's an electrical code issue to trying to run it on a standard outlet, it's against code.

My racks have 84 outlet PDUs (power strips) that have dual 208 volt 30 amp 3-phase plugs (L21-30). That's about 18.7 kW per PDU. Life is simpler when you get rid of that silly American 110v connection.
 
Last edited:
AidenShaw said:
Sorry, but that's nonsense. Complete nonsense.

Heat (watts) is amps*voltage. Double the voltage - you halve the amps but have the same heat. *the same heat*

The big thing about higher efficiency power supplies is that the heat (for a given draw from the mains) comes more from the electronics and less from the PS itself. As I pointed out earlier, a 450w 90% power supply and a 785w 94% power supply generate about the same amount of heat from the power supply itself.

There is an advantage to 240v-only in that the amperage drops by half, so you have thinner copper wires and it can be a bit more compact due to that.

The big advantage to using higher voltage, though, is that 1400w is too much to run on a standard American 15 amp 110v circuit - so there's an electrical code issue to trying to run it on a standard outlet, it's against code.

My racks have 84 outlet PDUs (power strips) that have dual 208 volt 30 amp 3-phase plugs (L21-30). That's about 18.7 kW per PDU. Life is simpler when you get rid of that silly American 110v connection.
Click to expand...

Do you have a masters in electrical engineering? More specifically power systems? Watts is power, which is not heat. Ill let you research this more, look into power supply designs and analyze the circuits. You can learn on your own which part of the power supply contributes significantly to heat, this component, will be greatly influenced on current. Let me know what you find!! Good luck.
 
gman36 said:
Do you have a masters in electrical engineering? More specifically power systems? Watts is power, which is not heat. Ill let you research this more, look into power supply designs and analyze the circuits. You can learn on your own which part of the power supply contributes significantly to heat, this component, will be greatly influenced on current. Let me know what you find!! Good luck.
Click to expand...

so the EU mac pros run twice as cool since they're drawing half the amps?

(* note - i do not have a masters in electrical engineering.. just that if i read you correctly, you're saying this)

my understanding is that volts are the 'speed' of the current.. higher voltage with equal wattage means yes, lower current is being used.. but that lower current is being delivered much faster.
 
Last edited:
Apple probably designs the Eu MacPro proper supply by stepping down the AC to 120V. That is cheap and easy to do, rather than doing product development for two entirely different power supplies.
 
gman36 said:
Apple probably designs the Eu MacPro proper supply by stepping down the AC to 120V. That is cheap and easy to do, rather than doing product development for two entirely different power supplies.
Click to expand...

maybe someone will goggle it by the morning.. going to sleep
 
flat five said:
so the EU mac pros run twice as cool since they're drawing half the amps?

(* note - i do not have a masters in electrical engineering.. just that if i read you correctly, you're saying this)

my understanding is that volts are the 'speed' of the current.. higher voltage with equal wattage means yes, lower current is being used.. but that lower current is being delivered much faster.
Click to expand...

It isn't a linear relationship, since these are very non linear systems. Inductors and capacitors are very dynamic. But essentially if apple designed power supplies specifically for the EU market, I am claiming they will run cooler.

The voltage isn't a speed. Current all travels at the same rate.
 
Status
Not open for further replies.
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.
Top Bottom
Back