Its not there. You must add it to the cfg

Thanks for the information.  Is there a specific place in the file that it should be placed?

No, just dont put in the middle of another string.

 

Ex:

 

[Display]
ChangeTime=4.000000
TransitionTime=4.000000
TEXTURE_FILTERING=4
MSAA=8
SSAA=0
VSYNC=0
FXAA=0
TRIPLE_BUFFER=0
ActiveWindowTitleTextColor=255,255,255
ActiveWindowTitleBackGroundColor=0,28,140,64
NonActiveWindowTitleTextColor=255,255,255
NonActiveWindowTitleBackGroundColor=24,33,87,64
InfoUpperRightTextColor=255,0,0
InfoUpperRightBackGroundColor=0,0,0,0
InfoLowerLeftTextColor=255,255,255
InfoLowerLeftBackGroundColor=255,0,0,128
InfoLowerRightTextColor=255,255,255
InfoLowerRightBackGroundColor=255,0,0,128
InfoBrakesEnable=True
InfoParkingBrakesEnable=True
InfoPauseEnable=True
InfoSlewEnable=True
InfoStallEnable=True
InfoOverspeedEnable=True
InfoWeaponSystemEnable=True
InfoCountermeasureEnable=True
SKINNED_ANIMATIONS=1
TEXTURE_BANDWIDTH_MULT=30
UPPER_FRAMERATE_LIMIT=0
WideViewAspect=True
FullScreenExclusive=True
FullScreenAutoFill=True
FullScreen=False

[JOBSCHEDULER]
AffinityMask=14

[sCENERY]
POP_FREE_AUTOGEN=1
AUTOGEN_BATCH_LOD=2
AUTOGEN_TREE_LOD=2
IMAGE_COMPLEXITY=3

Much appreciated.  By the way, is the 14 value for a cpu without hyperthreading and the 15 value is for a cpu with hyperthreading?

No prob. Ive experimented with this number and not seen any differences. Specially on Quad cores. Hyperthreading disabled.

Much appreciated.  By the way, is the 14 value for a cpu without hyperthreading and the 15 value is for a cpu with hyperthreading?

 

 

BTW, here's a simple way to figure out the AM (even though for P3d2, changing the AM shouldn't have much of an effect one way or the other). First, run the Windows Calculator app. Put the calculator in Programmer mode from the View menu item. Next put the app in binary mode by clicking that radio button. Now, for a quad core CPU, with no hyperthreading, all cores on would be:

 

1111

 

Enter that and then hit the decimal button. The display should now read 15.

 

For a quad core CPU with HT on (8 logical cores), the binary is:

 

11111111

 

and the decimal is 255.

 

To turn a core off, the lowest core (Core 0) is the right hand most digit. So for a quad core CPU and forcing Core 0 to not be used by P3d2, the binary is

 

1110

 

and the decimal is 14.

 

For a quad core CPU with HT on (8 logical cores), the binary is:

 

11111100

 

and the decimal is 252.

This worked well for me-

edited the cfg file to give me affinity mask of 252, as suggested by the post above, gave me ~65% load whereas I was only getting ~37% before.

 

Thanks guys

 

System is 4770K at 4.4ghz, 16gb ram, SSD, 2x GTX760.  still waiting for the SLI patch!!!  

 

Awesome, thanks guys.  Gotta love the forums

 

Runs at about 25fps average, running everything pretty much maxed out, full shadows, etc.  Seems to run smoother and loads a bit faster, in my opinion.  The CPU is definitely working harder, for sure.  Also, my GPU monitor is pegged at 100, which makes sense.  SLI PATCH!!!!

 

 

Screenshot of the FTX Juneau airport with Global, south Alaska, etc.  Looks awesome, really impressed with P3dv2. Also Running the REX overdrive stuff. wow. Did I mention SLI Patch needed?  lol

AffinityMask=5 if no one else has suggested it.

 

SLI does not work as the code is too old. Also its a 32bit sim so don't expect it to use more than 4gigs of ram.

AffinityMask=5 if no one else has suggested it.

 

SLI does not work as the code is too old. Also its a 32bit sim so don't expect it to use more than 4gigs of ram.

Joined the forums just to put out bogus information???

 

1 - That mask is the strangest suggested value I've ever seen.

2 - SLI functions with Prepar3D v2 and given time will work even better.

3 - So it's 32bit... if you need more than 4gigs... maybe your software suffers from bloat? You never know. :wink:

LOL ... I am using SLI now and have my AM set to 1364 which lifted my FPS by about 25%. Between the two, SLI and AM, I have an improvement of at least 50% in overall performance.

 

Mac

Anyone tried AffinityMask=0?

 

Here's an interesting thought; just because an address bus only has 32 bits does not mean it can only access 2^32=4294967296 bytes.

 

When I first worked on computers they could access way more RAM than their address busses could read, through switched pages of RAM in hardware.

 

Later, certain versions of MS SQL server could access 32Gb through a specially supplied method of Virtual Paging.

 

Back then a few boards had 2 or 4 CPUs some had uprated cooling for some CPUs and so the Affinity Mask kept the high throughput processes off of the CPUs with regular cooling.

 

The Affinity Mask can only be used to block a core.

 

Steve

Anyone tried AffinityMask=0?

 

Here's an interesting thought; just because an address bus only has 32 bits does not mean it can only access 2^32=4294967296 bytes.

 

When I first worked on computers they could access way more RAM than their address busses could read, through switched pages of RAM in hardware.

 

Later, certain versions of MS SQL server could access 32Gb through a specially supplied method of Virtual Paging.

 

Back then a few boards had 2 or 4 CPUs some had uprated cooling for some CPUs and so the Affinity Mask kept the high throughput processes off of the CPUs with regular cooling.

 

The Affinity Mask can only be used to block a core.

 

Steve

 

Hasn't worked for me, netted either worse stutters or no effect in the instances I've tried it over AM = 14 on my system.

 

 

The Affinity Mask can only be used to block a core.

 

That is true but an AM of all zeros would literally block all cores, so I sure that a value like that probably defaults to something else, like maybe one core assigned to FSX.

AffinityMask=0? Hehe, I'm not suggesting it's going to be good, but what I'm thinking is it just blocks the first core, and so cores not included in the mask are all on.

SLI does not work as the code is too old. Also its a 32bit sim so don't expect it to use more than 4gigs of ram.

 

Can you cite a source for your claim about SLI? This is what Beau Hollis posted on April 8 in the LM P3D forums:

 

 

 

A quick note on SLI:

- If you have multiple monitors and multiple views, then you'll generally be better off driving a display from each video card than using SLI. If you have 2 cards and 1 monitor than it's worth giving SLI a shot.

 

- We're still waiting on NVidia to create an SLI driver profile for Prepar3D v2 so the drivers may not be picking the best SLI mode for our app be default.

 

- We still have more work to do to improve SLI performance, but we chose to focus on autogen vegetation performance for this release as it was a big contributor to OOMs and CPU boundedness.

 

- If you want to check out what the best-case improvement of SLI might be in the future, you can try this:

 

1. Make a copy of your Prepar3D.exe and rename it AFR-FriendlyD3D.exe. Now make a new shortcut to the AFR-FriendlyD3D.exe and launch that. This is a developer trick for telling the SLI drivers not to copy any resources between GPUs each frame. (See NVidia's SLI Best Practices Doc for more info: http://developer.download.nvidia.com/whitepapers/2011/SLI_Best_Practices_2011_Feb.pdf) In a perfect world, we wouldn't need any resources to be copied and this would result in a nice perf boost with no side effects.

 

In the real world, we do have some features that rely on resources that persist frame-to-frame. HDR is one of them so you'll probably want to turn that off to avoid getting an odd pulsing effect. You may also see some garbage geometry streak across the screen here and there. (at least that's what I was seeing on a dual titan rig with the drivers I was using at the time.)

 

Also, any time you enable/disable SLI in the driver settings you should clear out your shader cache or you may get driver crashes. This really shouldn't happen since cached shaders are compiled in a non-device-specific way, but it does seem to happen nonetheless. This is also sometimes true of doing driver updates.

Beau Hollis

Rendering System Lead - Prepar3D® Team

 

Also, don't waste your time trying to calculate AM settings. Download and install SimStarter and go to the Tweaks tab. You can select which cores you want to use and it will insert the appropriate AM value. This is what is looks like:

 

 

Sorry to start an argument. I was going off a quad core cpu for affinity. Tried AffinityMask=0, textures blurred but performance picked up. As I have Prepar3d 2.2 im just amazed how they've tweaked it (shaders,tree lod etc) It kicks FSX to the curb framewise.