Are we paying for inaccessible performance?

exigentsky · Oct 25, 2013

I had a maxed out MBP order about to ship but I became concerned after looking at the max TDP of the components. I hope I'm missing something because what I learned seems rather absurd if true.

The adapter outputs 85W so this is the maximum power draw of the 15-in rMBP. The TDP of the 4960HQ is 47W and the TDP for the 750M is ~45W. This itself adds up to 92W and that number would be greatly exceeded if we include the motherboard, screen and other components. Consider a program that requires maximum CPU and GPU performance while you're blasting music with maximum screen brightness (probably another 10W or more) plus Wi-Fi and BlueTooth on. Let's say you're connected to an external display via HDMI, have a USB 3 drive and a Thunderbolt 2 device too. The amount of power required to do all this would probably be 2x as much as that 85W adapter is capable of outputting.

It's not the most realistic use scenario but the point stands. The only way I see the MBP doing this is by throttling down the CPU and GPU a ton. But then, why do I pay for the highest end CPU and a discrete GPU if they're running at 60% when I actually need them? How can they get away with an 85W PSU if only the graphics card and CPU TDP exceeds that?

Quu · Oct 25, 2013

4960HQ's TDP includes the GPU part. It's the entire package. If you're using the 750m then the Iris Pro graphics on the 4960HQ will be idle, drastically lowering the CPU's TDP.

GSPice · Oct 25, 2013

exigentsky said:
But then, why do I pay for the highest end CPU and a discrete GPU if they're running at 60% when I actually need them?

You'd also have to ask why Apple would pay for a part that they'd never implement at anywhere near full potential in their notebook. Trickery? Gimmicks? Conspiracy? To compete with other notebooks on paper only then to be destroyed in benchmarks?

See above post.

sephirot · Oct 25, 2013

You forgot the battery in your equation. Under full load your Macbook will draw power form the MagSafe and from the battery simultaneously.
I know this because on my old 2008 MacBook Dual-Core 2.4 Ghz the battery died and without battey it's locked to 1 Ghz even though it's plugged in via powercord. It's a limitation by the MagSafe.

magbarn · Oct 25, 2013

exigentsky said:
I had a maxed out MBP order about to ship but I became concerned after looking at the max TDP of the components. I hope I'm missing something because what I learned seems rather absurd if true.

The adapter outputs 85W so this is the maximum power draw of the 15-in rMBP. The TDP of the 4960HQ is 47W and the TDP for the 750M is ~45W. This itself adds up to 92W and that number would be greatly exceeded if we include the motherboard, screen and other components. Consider a program that requires maximum CPU and GPU performance while you're blasting music with maximum screen brightness (probably another 10W or more) plus Wi-Fi and BlueTooth on. Let's say you're connected to an external display via HDMI, have a USB 3 drive and a Thunderbolt 2 device too. The amount of power required to do all this would probably be 2x as much as that 85W adapter is capable of outputting.

It's not the most realistic use scenario but the point stands. The only way I see the MBP doing this is by throttling down the CPU and GPU a ton. But then, why do I pay for the highest end CPU and a discrete GPU if they're running at 60% when I actually need them? How can they get away with an 85W PSU if only the graphics card and CPU TDP exceeds that?

Apple has chosen the 85watt for size and weight. MBP 15 power draw in the last several years can exceed the AC adapters output. When this happens, it will draw the remaining wattage >85 watts from the battery. I've noticed if I'm playing BF3 on my rMBP, the battery will slowly drain. (maybe about 5% or so per hour so it's not fast) I suppose if you drain the battery all the way, then it will throttle at that point, but I've never gamed long enough to get there. OTOH, my previous 2011 MBP would overheat and it would throttle not because of power, but because of heat.

Starfyre · Oct 25, 2013

magbarn said:
Apple has chosen the 85watt for size and weight. MBP 15 power draw in the last several years can exceed the AC adapters output. When this happens, it will draw the remaining wattage >85 watts from the battery. I've noticed if I'm playing BF3 on my rMBP, the battery will slowly drain. (maybe about 5% or so per hour so it's not fast) I suppose if you drain the battery all the way, then it will throttle at that point, but I've never gamed long enough to get there. OTOH, my previous 2011 MBP would overheat and it would throttle not because of power, but because of heat.

Is this with your 2.6? This is the case where I wonder if going 2.3 for the lower heat and power draw would be a good idea... especially for high loads... I expect to actually increase power, albeit slowly when plugged into the outlet.

The iPad 3 is like what you described and it sucks. All classic macbook pros, when playing graphics intensive games under load don't do what you described.

john123 · Oct 25, 2013

GSPice said:
You'd also have to ask why Apple would pay for a part that they'd never implement at anywhere near full potential in their notebook. Trickery? Gimmicks? Conspiracy? To compete with other notebooks on paper only then to be destroyed in benchmarks?

See above post.

I actually think "conspiracy" is correct. Remember that it's really Apple that's nudged Intel on this whole iGPU thing in the first place. It's something of a symbiotic relationship at this point. I was shocked that we didn't see the HD4600 in the high-end configuration, but I'd be willing to bet Intel gave Apple a sweetheart deal to stick with the Iris Pro there.

exigentsky · Oct 25, 2013

GSPice said:
You'd also have to ask why Apple would pay for a part that they'd never implement at anywhere near full potential in their notebook. Trickery? Gimmicks? Conspiracy? To compete with other notebooks on paper only then to be destroyed in benchmarks?

See above post.

Only stress test benchmarks monitoring the throttling on the GPU and CPU can answer that. You'd still get extra performance when you need the max CPU performance but not for all 4 cores and that is usually not relevant since few programs are multi-threaded well.

It's a really good point that Iris Pro would be idle while using the 750M. That probably means the TDP would be more in line with the 4702HQ and probably lower since that has integrated graphics too. I'd guess ~30W.

On the other hand, a Dell m3800 which has a slower CPU with a lower TDP and a screen that consumes ~35% less was pulling 110W from the wall according to a guy that has a pre-release version. That one has a 130W adapter so all signs are pointing to heavy down-throttling for extended high load periods.

BTW: The TDP of the 2.3 is the same as the 2.6.

vanc · Oct 25, 2013

magbarn said:
MBP 15 power draw in the last several years can exceed the AC adapters output. When this happens, it will draw the remaining wattage >85 watts from the battery.

Exactly. That's what I learned with my first 15' MBP purchase in 2009.

exigentsky · Oct 25, 2013

sephirot said:
You forgot the battery in your equation. Under full load your Macbook will draw power form the MagSafe and from the battery simultaneously.
I know this because on my old 2008 MacBook Dual-Core 2.4 Ghz the battery died and without battey it's locked to 1 Ghz even though it's plugged in via powercord. It's a limitation by the MagSafe.

Ah, interesting. That definitely could cover some of it. Although it would be awkward for me to run a chess program like the Mac HIARCS and then find that my battery is empty even when plugged in after a few hours.

Still, I question why they'd simply not use a higher wattage rated adapter instead of dipping into the battery. What would be the disadvantage?

BTW: I'm reading that the TDP might be similar even without Iris Pro if at max load. It might just stay at a higher turbo.

"With the CPU graphics cores at idle, it just allows the processing cores to stay in turbo mode all the time. The chip will want to pull 55 watts in the short term PL1 turbo mode. The chip will then fall out of PL1 mode and stay at 47 watts of draw unless there is a heat or electrical supply issue. Loading the IRIS side will just take away from the clocks on the CPU side."

sephirot · Oct 25, 2013

exigentsky said:
Ah, interesting. That definitely could cover some of it. Although it would be awkward for me to run a chess program like the Mac HIARCS and then find that my battery is empty even when plugged in after a few hours.

Still, I question why they'd simply not use a higher wattage rated adapter instead of dipping into the battery. What would be the disadvantage?

IDK, it could be a safety measurement, power limitation of the MagSafe, prevent sparks from flying when the MagSafe comes loose or some technical limitation, no idea.
I only know before the MagSafe this wasn't the case, you could take out the battery and run your MBP with full speed when plugged into the wall.

Polymorphic · Oct 25, 2013

In a different thread on this forum yesterday, someone posted a screenshot suggesting that the GT 750M in their MacBook Pro Retina was underclocked to 925 MHz. If that's the case (I'm waiting for the AnandTech review to be sure), it's possible the 750M isn't drawing as much power as its maximum TDP suggests.

GSPice · Oct 25, 2013

john123 said:
I actually think "conspiracy" is correct. Remember that it's really Apple that's nudged Intel on this whole iGPU thing in the first place. It's something of a symbiotic relationship at this point. I was shocked that we didn't see the HD4600 in the high-end configuration, but I'd be willing to bet Intel gave Apple a sweetheart deal to stick with the Iris Pro there.

I too was shocked that Apple managed to select 5200+750. I remember reading much about Apple having a bit of leverage with Intel.

There are folks that have much more empirical evidence than I regarding the OP's question, but I'm quite willing to bet that Apple doesn't waste their money with components that are not/cannot be reasonably optimized to their liking. In fact, Apple seems a bit stubborn to ensure that components are tailored to their liking. Not saying it's a bad thing, but it goes back to the leverage bit - not many companies can play the pipe like Apple can.

john123 · Oct 25, 2013

GSPice said:
Not saying it's a bad thing, but it goes back to the leverage bit - not many companies can play the pipe like Apple can.

Totally agreed, and that's why I'd kill to know what they're paying Intel for those 2.3Ghz+ Haswell chips. All summer long, every time someone on here suggested we might see a 5200+750/760/765, I scoffed at it for cost reasons and insisted it wouldn't make any sense. After all, this is the company that's had its share price hammered not because of declining revenues (they've continued to increase), but rather erosion in margins.

So there's no way they did this without the economics working in their favor, and that's why I'm curious what pricing Intel gave them. Unfortunately we won't be able to figure it out very well until the Q1 earnings are released, if then, but it's a thought provoking question nonetheless.

GSPice · Oct 25, 2013

john123 said:
Totally agreed, and that's why I'd kill to know what they're paying Intel for those 2.3Ghz+ Haswell chips. All summer long, every time someone on here suggested we might see a 5200+750/760/765, I scoffed at it for cost reasons and insisted it wouldn't make any sense. After all, this is the company that's had its share price hammered not because of declining revenues (they've continued to increase), but rather erosion in margins.

So there's no way they did this without the economics working in their favor, and that's why I'm curious what pricing Intel gave them. Unfortunately we won't be able to figure it out very well until the Q1 earnings are released, if then, but it's a thought provoking question nonetheless.

I mirror your sentiments exactly. And on top of all this, in context of the slumping PC and notebook market, folks had even more reason to believe that Apple would trade performance for profit margins.

The Haswell rMBP is no first-gen MBA, but it seems that Apple continues to push, not be pulled.

magbarn · Oct 25, 2013

Starfyre said:
Is this with your 2.6? This is the case where I wonder if going 2.3 for the lower heat and power draw would be a good idea... especially for high loads... I expect to actually increase power, albeit slowly when plugged into the outlet.

The iPad 3 is like what you described and it sucks. All classic macbook pros, when playing graphics intensive games under load don't do what you described.

Well assuming the Haswell Iris Pro really sticks to it's TDP (Intel's chips have been known to exceed their TDP's in the past) yes, you'll be using much less power than 85 watts. It's funny but PC laptop manufacturers like Dell will bundle a 65 watt power adapter on their IGP only laptops while bundling a 90 watt adapter on the same model with a dGPU.

Count Blah · Oct 25, 2013

sephirot said:
You forgot the battery in your equation. Under full load your Macbook will draw power form the MagSafe and from the battery simultaneously.
I know this because on my old 2008 MacBook Dual-Core 2.4 Ghz the battery died and without battey it's locked to 1 Ghz even though it's plugged in via powercord. It's a limitation by the MagSafe.

What are you using to see that it's stuck at 1Ghz without the battery?

exigentsky · Oct 25, 2013

magbarn said:
Well assuming the Haswell Iris Pro really sticks to it's TDP (Intel's chips have been known to exceed their TDP's in the past) yes, you'll be using much less power than 85 watts. It's funny but PC laptop manufacturers like Dell will bundle a 65 watt power adapter on their IGP only laptops while bundling a 90 watt adapter on the same model with a dGPU.

I don't get why that's funny or surprising. A dGPU consumes more and thus requires an adapter with a higher wattage rating.

This is actually my entire problem with the MBP. I want to be able to use the performance I pay for and it's not designed to be used intensely for longer periods. The adapter is incapable of supporting a MBP on full-load. Thus, the MBP will consume battery to compensate and probably throttle down too. I'd love to see some stress tests and benchmarks to clarify the details for this.

My guess is that the MBP thermals wouldn't have been able to support full-load for long anyway so they decided to not use an adapter that can support full-load like 130W but I really don't know. It seems very odd for a pro-machine.

exigentsky · Oct 26, 2013

Apparently others ran into the exact issue I was anticipating.

After further analysis, i7-3720QM and GT 650M are pushing the system beyond its 80W power limit. When that occurs, the GT 650M will throttle. Using ThrottleStop or a similar utility, you can clamp CPU frequency to below the power limit and throttling should not occur. In CPU heavy games, this can be a huge problem. The stock 85 Watt Power Supply is simply not enough.

GPU-Z lets me see voltage, but that is not important.

AIDA64 Extreme Edition (Trial) lets me see how many watts are running though the power supply. Earlier today, I monitored that with GTA IV with my CPU at 3.0 GHz with my GPU overclocked.
At peak before throttling the GPU, it was at 77-78W. Underclocking the CPU or the GPU brought it down to 55-60W.
Both the CPU and GPU are rated at 45W and are extremely power hungry at high clocks. I am pretty convinced that I cannot access beyond 85W before my GPU starts throttling.

When I first got this computer, I did not have this problem. Instead, I noticed my battery draining after several hours of gaming.
There was recently an update to my Macbook Pro (I am also running 10.9 Preview) that fixed the battery draining; but I think the GPU throttling is the fix...

Can someone else confirm that the battery is no longer drained on 2013 models? This would mean even more throttling and worse performance when under stress.

MaclandValhalla · Oct 26, 2013

What about the 15 inch (38.1 cm) Haswells that don't have a dGPU? Would they ever have to draw from their batteries while plugged in?

I don't like the idea of using battery cycles while tethered to a power outlet.

Wuiffi · Oct 26, 2013

john123 said:
Totally agreed, and that's why I'd kill to know what they're paying Intel for those 2.3Ghz+ Haswell chips. All summer long, every time someone on here suggested we might see a 5200+750/760/765, I scoffed at it for cost reasons and insisted it wouldn't make any sense.

So did I ^^
Guess it went like this:
Apple: Intel! Your graphics don't deliver, how are we supposed to go iGPU only?
Intel: that's your problem. You asked for this chips, now use them!
Apple: *****! They are not powerfull enough. We have to put in a dGPU too. No way we gonna pay you for this?
Intel: how about we meet in the middle - we give you a great deal and you take all of them because noone else wants em
Apple: ok. And let's try again next year

:>

juliant · Oct 26, 2013

In the market to jump to MBP and this is not something encouraging... especially not at this premium price point.

Orr · Oct 26, 2013

GSPice said:
I too was shocked that Apple managed to select 5200+750. I remember reading much about Apple having a bit of leverage with Intel.

There are folks that have much more empirical evidence than I regarding the OP's question, but I'm quite willing to bet that Apple doesn't waste their money with components that are not/cannot be reasonably optimized to their liking. In fact, Apple seems a bit stubborn to ensure that components are tailored to their liking. Not saying it's a bad thing, but it goes back to the leverage bit - not many companies can play the pipe like Apple can.

Definitely agree. The consensus around here thus far seems to be that the overlap between Iris Pro and the 750M is overwhelming and the performance gains are non-existent to minimal at best. I highly doubt this and think that the popular opinion will begin to shift as more performance data trickles out.

john123 said:
Totally agreed, and that's why I'd kill to know what they're paying Intel for those 2.3Ghz+ Haswell chips. All summer long, every time someone on here suggested we might see a 5200+750/760/765, I scoffed at it for cost reasons and insisted it wouldn't make any sense. After all, this is the company that's had its share price hammered not because of declining revenues (they've continued to increase), but rather erosion in margins.

So there's no way they did this without the economics working in their favor, and that's why I'm curious what pricing Intel gave them. Unfortunately we won't be able to figure it out very well until the Q1 earnings are released, if then, but it's a thought provoking question nonetheless.

Very, very interested as well. It was morbidly amusing to see the few souls that dared to suggest the above in advance mercilessly put down by the masses all summer long.

Wuiffi said:
So did I ^^
Guess it went like this:
Apple: Intel! Your graphics don't deliver, how are we supposed to go iGPU only?
Intel: that's your problem. You asked for this chips, now use them!
Apple: *****! They are not powerfull enough. We have to put in a dGPU too. No way we gonna pay you for this?
Intel: how about we meet in the middle - we give you a great deal and you take all of them because noone else wants em
Apple: ok. And let's try again next year

:>

I actually have a strong suspicion that this isn't far off from the truth.

magbarn · Oct 26, 2013

juliant said:
In the market to jump to MBP and this is not something encouraging... especially not at this premium price point.

As others have come to the conclusion before, if you're going to be doing heavy gaming, just save yourself the headache and get a >6 lbs. Alienware 14 that can game for hours on end. Apple lately has been all about thinness and weight. And to think that people were complaining about GT750M and wishing for a GTX765M which consumes even more power!

juliant · Oct 26, 2013

magbarn said:
As others have come to the conclusion before, if you're going to be doing heavy gaming, just save yourself the headache and get a >6 lbs. Alienware 14 that can game for hours on end. Apple lately has been all about thinness and weight. And to think that people were complaining about GT750M and wishing for a GTX765M which consumes even more power!

Who said anything about gaming for hours on ends.

Before you start to assume that I am getting a MBP for gaming due to your assumption that I own an ASUS G53-JW. I am not into gaming as much as 3 years ago. I am having interest in Macs as well as nix based system, which also aligns to my work.

Just saying that to be shelling out this amount of money to find that we are at the mercy of cooperate ********* is just not very encouraging for Mac adopters at this point.

Granted nothing is perfect, still, at this price premium, its not encouraging like I said.

Are we paying for inaccessible performance?

macrumors regular

macrumors 68040

macrumors 68000

macrumors newbie

macrumors 68030

macrumors 68030

macrumors 68030

macrumors regular

macrumors 6502

macrumors regular

macrumors newbie

macrumors regular

macrumors 68000

macrumors 68030

macrumors 68000

macrumors 68030

macrumors 68040

macrumors regular

macrumors regular

macrumors member

macrumors 6502a

macrumors newbie

macrumors 6502

macrumors 68030

macrumors newbie

Our Staff