Manual switch off on discrete graphics for R2 users

Just disable the 335M in the Device Manager and youll be able to get maximum battery life. You can enable it later if you want it. No need to restart either.

 

Can you disable intel hd graphics to force the discrete on?

 

Wait.......so is there a way to run Aero using the discrete on R2? What if I dont want to use the intel at all?

 

Can't you just choose "High performance nvidia processor" from the nvidia control panel under the global settings --> preferred graphics processor? It seems like doing that would basically always give you the nvidia processor unless there is a program doesn't use the global setting and applies the Intel processor... but most of the generic things I've seen in the "Program Settings" tab all use "Use Global Setting"


John

 

Nah that doesn't seem to actually dmake it 100% 335m. I think that tells it if it's going to guess for a new programs, or doesnt know, use the 335m. It will still try to use intel all the time for windows itself and simple tasks (which isnt a big deal really)

 

No. See my post in the Optimus thread. The 335M actively hijacks the Intel IGP to do all the drawing and rendering. The NVidia card just does all the calculations. To sum it up. It's a lot more complex that that, but I'm not qualified to explain beyond that.

 

I've sent a note to video engineering to ask if it is even possible to have a hard switch (like the R1) with the Optimus solution. I don't know if this is even possible with the way the hardware is implemented.

 

AFAIK it just hijacks the framebuffer.

So it is really just hijacking the output to the screen. At least that is what I get from it.

Apple with the new MacBook Pros utilize something very similar to Optimus. They have had the same issues and complaint until recently. Cody Krieger has developed a utility called gfxCardStatus for it which allows you to see which GPU is being utilized and force either Integrated or Discrete much like can be done in Hybrid Graphics. You can see more info here including descent screenshots and the program is actually found here. Cody has made his gfxCardStatus program open source so I don't see why someone with proper programming knowledge cant take a look at it and figure out what he is doing and do it for Optimus. (That is assuming it is all being done at the hardware level.)

 

Eh, okay I was right the first time then. I was thinking the Arrandale IGP was doing more than that. The amount of technical data on Optimus isn't very widely available (nor is it easy to read) from my position.

I saw gfxCardStatus (there's also a second utility which does the same, it came out maybe a week earlier). The MacBook Pro hardware is configured slightly differently though. Hypothetically speaking, Optimus is nonfunctional in Windows on a Macbook Pro, but if the 330 was really wired into the Arrandale IGP, then the 256.xx series drivers should hypothetically enable Optimus. I haven't looked into it recently, so I'm not qualified to make such statements. The above is just my theory.

 

No everyone is really just guessing at what Nvidia and Apple are doing with their Hybrid/Optimus technology as there isn't really too much concrete information. To top things off some of the information that is out there is false or made up.

For example from apple/ars technica:

Now with what we know about Optimus it works pretty much the same way that they describe in the first part. There has also been the erroneous statement out that Nvidia just uses a software list for Optimus switching which we know is also untrue if you read the whitepaper:

The only real difference between Apples approach and Optimus seems to be the claim that it can power off the iGPU while utilizing the dGPU. We know this can not be achieved with Optimus because the iGPU must be on in order to utilize the framebuffer. I question this until I see some real proof however.

This is true. It seems however that Apple may be intentionally disabling one of the GPUs when utilizing Windows though. My assumption is that they are doing something similar to switching the bios setting to static (single GPU).

 

From what I understand also, because what Apple did wasn't Optimus Technology, if you did get the new Mac and run boot camp, Windows wouldn't be able to switch GPU. And probably won't, as this seem like apple own this new tweak of the technology, nVidia probably can't write Windows driver with that technology. Another attempt from Apple to keep you in iJail.

 

I strongly doubt that. I've got a Macbook with the X3100 and I have full system control in software. Take GMABooster for example. Software overclocking on the old GMA 950 chipset. There isn't one for the 965 (what I have) yet, but if anything, Apple hasn't bothered to lock anything down.

My brother has a Macbook Pro with the 9400M/9600GT from 2009. In Windows, it always defaults to the 9600GT. My theory is that Windows looks for a discrete display option first and uses it as the primary if it finds one, otherwise it defaults to the IGP. In an Optimus situation, the only GPU that would appear to Windows IS the IGP, which at a hardware level is powered by the GT335 (or 330). Thus, it, along with computer model identifier, would allow Optimus to be utilized by the drivers.

Now, consider the new MBP, assuming for the moment, that what I put forth about Windows is true. If hypothetically, the GT330 was split from the IGP, then Windows would autodetect the GT330 or in my brother's case, the 9600GT and just use it. It doesn't bother to look for an IGP because it found a better primary display source. If this were the case, then Optimus on the MBP would be impossible, unless a software implementation could be created to bridge the gap between IGP and GT330.

Final comment about Apple. It would be a poor marketing (and business) decision to disable the IGP in Windows. You'd have to lock it out in the EFI, and there's no way of getting EFI to disable or enable hardware based on what OS is detected as booting. Besides, somebody would complain to Apple that their battery life sucks in Windows... wait, maybe they want that, so people buy more Macs... Hmmm......

This isn't Mac Land though, so I'll finish up with a final statement regarding Apple. I don't think they locked it out. If anything, I don't think they have control over the process at all, since the same IGP needs to run OSX too. What the gfxStatus application most likely does to force one or the other is to intercept OGL calls (or any graphically intensive calls) and redirect them to the IGP. To force the 330 on in the MBP, it probably does the opposite by sending a very non-intensive OGL call.


/ramble

 

You guys keep saying there's no concrete info on Optimus, but the whitepaper which you mention in the quote from apple/arstechnica explains the whole thing. (Did you read the whitepaper?)

...but I'll save you some time. Page 11 answers your question as to whether the iGPU must be on: "Using NVIDIA'S Optimus technology, when the discrete GPU is handling the rendering duties, the final image output to the display is still handled by the Intel integrated graphics processor (IGP). In effect, the IGP is only being used as a simple display controller.... When less critical or less demanding applications are run, the discrete GPU is powered off and the Intel IGP handles both rendering and display calls to conserve power". As you can see, the IGP is always on, and that's how it provides the flicker-free transition.

edit: whoops, nevermind, I thought you meant you questioned Optimus needing to use the IGP and not Apple powering off one or the other GPUs, but I'll keep the post in tact to point out that the whitepaper provides good info on Optimus.

edit2: *dance dance dance*

 

We've established that already. We're discussing how gfxStatus manages to "disable" one of the GPUs in OSX.

Edit: The main topic prior to derailment was the discussion of how to force the R2 to stay on one GPU.

Edit 3: The above poster and I seem to be having some kind of edit dance. :O

 

AH ha.... from anantech

I believe I saw this method in Nvidias patent. I think Nvidia may be letting Apple claim they came up with this method due to their dispute with Intel.

EDIT: So yeah. As it is Apples implementation may actually be closer to Hybrid Graphics than Optimus. So I guess a no go looking there.