rMBP: battery lower than 30% performance drops?

infernohellion

macrumors 6502
Original poster
May 14, 2006
395
14
København
Hi guys,

I've been looking for a few pages and couldn't find a threat about this issue yet though I recall there were discussions about this issues inside one of the threads.

Anyway, I was just wondering how many of you out there notice this issue? For me it's become really annoying and the performance drop is often so drastic that it's starting to make scrolling unbearable and unnatural on many of the apps.

I am not the type that is extremely picky with the scroll lags in general and I find normal usage on the rMBP at 1920x1200 (HiDPI) mode alright.

If so, has there any solution?
Cheers,
 

GGJstudios

macrumors Westmere
May 16, 2008
44,419
759
Your battery charge being lower than 30% will not affect performance at all. If you're having performance issues, the source is elsewhere. When you're experiencing such lagging:
  1. Launch Activity Monitor
  2. Change "My Processes" at the top to "All Processes"
  3. Click on the CPU column heading once or twice, so the arrow points downward (highest values on top).
  4. Click on the System Memory tab at the bottom.
  5. Take a screen shot of the entire Activity Monitor window, then scroll down to see the rest of the list, take another screen shot
  6. Post your screenshots.

If you're having performance issues, this may help:
 

infernohellion

macrumors 6502
Original poster
May 14, 2006
395
14
København
I'm quite sure the source isn't from else where. I've been continuously running the same apps: surfing the web, and using Twitter. Hmmm.

And no, I'm not new to Macs.
 

jav6454

macrumors P6
Nov 14, 2007
16,871
1,492
1 Geostationary Tower Plaza
Hi guys,

I've been looking for a few pages and couldn't find a threat about this issue yet though I recall there were discussions about this issues inside one of the threads.

Anyway, I was just wondering how many of you out there notice this issue? For me it's become really annoying and the performance drop is often so drastic that it's starting to make scrolling unbearable and unnatural on many of the apps.

I am not the type that is extremely picky with the scroll lags in general and I find normal usage on the rMBP at 1920x1200 (HiDPI) mode alright.

If so, has there any solution?
Cheers,
It is the GPU struggling to keep up. No science to it. Apple put a mediocre GPU into a "Premium" product. Bad decision and reason why I don't buy the Retina Mac just yet.
 

GGJstudios

macrumors Westmere
May 16, 2008
44,419
759
I'm quite sure the source isn't from else where. I've been continuously running the same apps: surfing the web, and using Twitter. Hmmm.

And no, I'm not new to Macs.
I didn't suggest that you're new to Macs, but the battery charge being below 30% will not affect performance. If you care to diagnose the issue, you can post the screen shots. It's up to you.
 

DTKblaster

macrumors member
Aug 3, 2012
74
4
If you're plugged in, and pulling more power than the 85w provided by the adapter, then your battery will supplement the power adapter. I read somewhere that in order to stop power being cut once the battery is empty, the processor down clocks below 30% battery so that the power draw comes bellow 85w and is fully provided by the adapter.

You mention scrolling, which makes me doubt your doing anything which should put the processor and graphics card under full load. :/
 
Last edited:

GGJstudios

macrumors Westmere
May 16, 2008
44,419
759
In order to top power being cut once the battery is empty, the processor down clocks below 30% battery so that the power draw comes bellow 85w and is fully provided by the adapter.
Please state your source for this claim.
 

DTKblaster

macrumors member
Aug 3, 2012
74
4
Please state your source for this claim.
I can't find where I read it first.

After almost an hour of testing and a toasty laptop, all sorts of other throttling issues stemming from heat get in the way of maximizing power draw (even with wifi, bluetooth, full brightness screen and keyboard).

So i retract my statement.
 

TheRdungeon

macrumors 6502
Jul 21, 2011
382
0
I can definitely vouch for the downclocking, but it seems to happen at 10% for me. I play back logic files that are normally no sweat, then as soon as 10% comes up the exact songs are stressing the CPU a whole lot more, sometimes even stopping playback
 

Dangerous Theory

macrumors 68000
Jul 28, 2011
1,981
28
UK
It is the GPU struggling to keep up. No science to it. Apple put a mediocre GPU into a "Premium" product. Bad decision and reason why I don't buy the Retina Mac just yet.
Mediocre? The GPU is incredible. I've been playing Just Cause 2 at almost max settings bar anti aliasing at 1680x1050 and GTA 4 at 1920x1200. Absolutely brilliant for a laptop, let alone one so thin and light. Maybe it's cause I'm used to the 13", but this does all I wanted and more.
 

Textureboy

macrumors 6502
Jul 25, 2012
322
40
It is the GPU struggling to keep up. No science to it. Apple put a mediocre GPU into a "Premium" product. Bad decision and reason why I don't buy the Retina Mac just yet.
You should buy before you make absurd claims
 

katmeef

macrumors 6502
Jul 20, 2010
368
0
is it documented by apple that the CPU goes into reduced power mode when the batteries hit a certain level, or that the batteries supplement the charger (both mentioned above)

Pretty sure that is wild conjecture. I use my Rmbp working (using parallels) from the garage at night (so i can smoke my ciggies while I work), I run the battery almost right down nightly (too lazy to bring an extension cord) and have noticed no performance differences...
 

jav6454

macrumors P6
Nov 14, 2007
16,871
1,492
1 Geostationary Tower Plaza
Mediocre? The GPU is incredible. I've been playing Just Cause 2 at almost max settings bar anti aliasing at 1680x1050 and GTA 4 at 1920x1200. Absolutely brilliant for a laptop, let alone one so thin and light. Maybe it's cause I'm used to the 13", but this does all I wanted and more.
You should buy before you make absurd claims
When you two go on and learn what the 650 is marketed as on nVidia's website come back and talk. The 650 is a mediocre product for the retina challenge. There is no arguing that. Apple just makes it look like its the best thing ever when it is not. As it is the 650 is being pushed to limits and is struggling to keep up with the large pixel push.

Now a 660 onwards would be good. Hopefully next generation retina can use a better GPU.
 

terraphantm

macrumors 68040
Jun 27, 2009
3,740
552
Pennsylvania
jav6454 said:
When you two go on and learn what the 650 is marketed as on nVidia's website come back and talk. The 650 is a mediocre product for the retina challenge. There is no arguing that. Apple just makes it look like its the best thing ever when it is not. As it is the 650 is being pushed to limits and is struggling to keep up with the large pixel push.

Now a 660 onwards would be good. Hopefully next generation retina can use a better GPU.
GT 650M and GTX 660M have identical cores. In the past, the GT 650M had lower clocks and/or gimped memory compared to the GT 650M. That's not the case with our rMBPs - our core and memory clocks are actually higher than most GTX 660Ms, and we still have GDDR5. So our GPUs essentially are slightly overclocked GTX 660Ms. A 670 or 680 would require a lot more power and produce more heat. With that said, I think Apple should've stuck in 2GB video memory.

As far as the HD4000 - that's integrated into all IvyBridge CPUs. It doesn't make sense not to use it for the sake of battery life. But when you need to do anything that requires real performance, the discrete one should kick in.

Here's a thread on a non Apple forum that discusses Apple's "GT 650" vs others' "GTX 660s" http://forum.notebookreview.com/gaming-software-graphics-cards/672830-so-nvidia-gt-650m-gtx-660m-confused.html
 

AZREOSpecialist

macrumors 68020
Mar 15, 2009
2,117
925
It is the GPU struggling to keep up. No science to it. Apple put a mediocre GPU into a "Premium" product. Bad decision and reason why I don't buy the Retina Mac just yet.
Here we go again, more inaccurate information supplied without any corroborating evidence. Way to go!

Now, for those who appreciate facts and evidence… take a look at this video that shows tons of activity going on and NO LAG - NONE. This is at the 1920x1200 scaled resolution:

http://youtu.be/Dudx1GhHGiY

Until jav6454 or anyone else making the same claim can post a video, I call BS.
 

jav6454

macrumors P6
Nov 14, 2007
16,871
1,492
1 Geostationary Tower Plaza
Here we go again, more inaccurate information supplied without any corroborating evidence. Way to go!

Now, for those who appreciate facts and evidence… take a look at this video that shows tons of activity going on and NO LAG - NONE. This is at the 1920x1200 scaled resolution:

http://youtu.be/Dudx1GhHGiY

Until jav6454 or anyone else making the same claim can post a video, I call BS.
I have evidence. Can't post it right now since I am mobile. A mod can corroborate I'm posting mobile.

GT 650M and GTX 660M have identical cores. In the past, the GT 650M had lower clocks and/or gimped memory compared to the GT 650M. That's not the case with our rMBPs - our core and memory clocks are actually higher than most GTX 660Ms, and we still have GDDR5. So our GPUs essentially are slightly overclocked GTX 660Ms. A 670 or 680 would require a lot more power and produce more heat

As far as the HD4000 - that's integrated into all IvyBridge CPUs. It doesn't make sense not to use it for the sake of battery life. But when you need to do anything that requires real performance, the discrete one should kick in.
That is different. The 650 and the 660 have different architectures. NVidia's and ATI's strategy for mobile is not the same as desktop. In desktop the 4 top cards are the newest architecture starting with the highest and then slowly making it down. Example the GTX680, then the slightly binned 670 and the further binned 660. On mobile what manufacturers do is just give the two top mobile GPUs the new architecture. However the cores are usually binned to half their desktop counterparts. I can quickly provide the example of the desktop HD5870 and the mobile 5870 (1600 v 800 core count).

The rest of the series is just last generation architecture clocked higher. In the case of the 650 v 660 both are different architectures.
 

terraphantm

macrumors 68040
Jun 27, 2009
3,740
552
Pennsylvania
Here we go again, more inaccurate information supplied without any corroborating evidence. Way to go!

Now, for those who appreciate facts and evidence… take a look at this video that shows tons of activity going on and NO LAG - NONE. This is at the 1920x1200 scaled resolution:

http://youtu.be/Dudx1GhHGiY

Until jav6454 or anyone else making the same claim can post a video, I call BS.
Two things: When playing videos, the rMBP switches to the discrete GPU - so the UI will be smoother overall for things that would lag on the iGPU. Second, there is lag when you're scrolling through the verge.

I have an rMBP and absolutely love it - but especially at the 19x12 mode, there is lag switching through spaces and other similar tasks when on the iGPU. Most sites scroll okay though. The slight stutter that is present doesn't really bother me, since most macs I've used are about the same in that regard in my experience.

----------

That is different. The 650 and the 660 have different architectures. NVidia's and ATI's strategy for mobile is not the same as desktop. In desktop the 4 top cards are the newest architecture starting with the highest and then slowly making it down. Example the GTX680, then the slightly binned 670 and the further binned 660. On mobile what manufacturers do is just give the two top mobile GPUs the new architecture. However the cores are usually binned to half their desktop counterparts. I can quickly provide the example of the desktop HD5870 and the mobile 5870 (1600 v 800 core count).

The rest of the series is just last generation architecture clocked higher. In the case of the 650 v 660 both are different architectures.
They're not different architectures - at least not in the mobile versions. Do some research. They both use the GK107 GPU, and they both have 384 stream processors with 128-bit memory busses. So the only differentiators will be clocks and GDDR3 vs GDDR5.
 

GGJstudios

macrumors Westmere
May 16, 2008
44,419
759
is it documented by apple that the CPU goes into reduced power mode when the batteries hit a certain level
I think people are misreading this statement by Apple, which only applies to older Macs with removable batteries:

MacBook and MacBook Pro: Mac reduces processor speed when battery is removed while operating from an A/C adaptor
If the battery is removed from a MacBook or MacBook Pro, the computer will automatically reduce the processor speed. This prevents the computer from shutting down if it demands more power than the A/C adaptor alone can provide.
There is no Apple statement that I've ever seen that mentions downclocking or reducing processor speed based on the level of battery charge remaining.
 
Last edited:

mattkidd

macrumors regular
Hi guys,

I've been looking for a few pages and couldn't find a threat about this issue yet though I recall there were discussions about this issues inside one of the threads.

Anyway, I was just wondering how many of you out there notice this issue? For me it's become really annoying and the performance drop is often so drastic that it's starting to make scrolling unbearable and unnatural on many of the apps.

I am not the type that is extremely picky with the scroll lags in general and I find normal usage on the rMBP at 1920x1200 (HiDPI) mode alright.

If so, has there any solution?
Cheers,
I agree with you here on this one. When the battery is close to being drained/empty, I certainly see difference in the graphics performance with the Nvidia chip enabled. I have for all four of my rMBP's

From my analysis it appears as though the Nvidia is either chocking from not enough power (unlikely) or is being under-clocked by the software. This is very much noticeable when playing a game on OSX on battery power.

If you however launch the graphics switching app (GFX) and change to the HD4000 graphics the issue is resolved (until the battery is drained), for general use. Obviously you don't get the same gaming performance and should probably charge the battery ;)
 

AZREOSpecialist

macrumors 68020
Mar 15, 2009
2,117
925
Two things: When playing videos, the rMBP switches to the discrete GPU - so the UI will be smoother overall for things that would lag on the iGPU. Second, there is lag when you're scrolling through the verge.
You're seeing something I'm not.
 

terraphantm

macrumors 68040
Jun 27, 2009
3,740
552
Pennsylvania
You're seeing something I'm not.
How is that not visible to you? The framerate clearly drops to something sub 30 there. You're in serious denial if you can't see that. And from experience, it's significantly worse when the iGPU is being used instead. If you do all those same things *without* videos playing in the background, it will be on the integrated GPU and choppier.

----------

Please quote or link such documentation. I think people are misreading this statement by Apple, which only applies to older Macs with removable batteries:

MacBook and MacBook Pro: Mac reduces processor speed when battery is removed while operating from an A/C adaptor

There is no Apple statement that I've ever seen that mentions downclocking or reducing processor speed based on the level of battery charge remaining.
Pretty sure that quote was supposed to be a question.
 

alphaod

macrumors Core
Feb 9, 2008
22,164
1,209
NYC
The OP clearly states, the performance drops after the battery drops to less than 30%.

The GPU was a hunk of junk and isn't really good enough, why isn't the lack of performance consistent?

Simply drawing conclusions without fulling reading the OP isn't going to solve the issue at hand.


________


Speaking of this Retina business. The GPU on the iPad can run a retina display fine (it's definitely not even close to the 650M), and can draw a second display via the AV adapter. I would hazard a guess and say the 650M is definitely capable of powering this display or Apple would not have offered it.
 

Dangerous Theory

macrumors 68000
Jul 28, 2011
1,981
28
UK
When you two go on and learn what the 650 is marketed as on nVidia's website come back and talk. The 650 is a mediocre product for the retina challenge. There is no arguing that. Apple just makes it look like its the best thing ever when it is not. As it is the 650 is being pushed to limits and is struggling to keep up with the large pixel push.

Now a 660 onwards would be good. Hopefully next generation retina can use a better GPU.
I don't give a **** what the 650M is marketed as by nVidia, I give a **** how well it actually performs in my retina MBP - and that performance is outstanding. The 650M isn't even coming close to its limits during normal use, and in fact there is opportunity to over clock it by 30-40%.

How the hell you think the GPU is mediocre I don't know. Pretty obvious you don't own one of these MacBooks yourself. It performs like an absolute champ.