I've started to use this tweak again on a new build: i5 2500K overclocked to 4.4, 8 GB DDR3 RAM, and a 570 GTX. It definately improves smoothness with the only downside being that I occasionally see brief white flashes on the outside terrain when changing views. This happens only on terrain with high autogen and scenery complexity and is presumably due to texture loading delays. The flashes are so brief as to be only a momentary distraction.Might there be a tweak that I've overlooked to address this minor downside? Overall, it's not enought to drive me to stop using UsePools=0

I had the same thing - and this stabilized it..[bufferPools]//UsePools=0UsePools=1PoolSize=30000

With stuff using a lot larger textures (ie 4096x4096) you may run in graphic glitches on some situations using zero BP. I had to lower my REX clouds to 2048x2048 to prevent spikes on certain type heavy cloud weather with certain aircraft and ORBX scenery. In most occasions it works just fine, but spikes appear when your GPU-load goes over 100%. Using BP prevents this all together, but it also lowers performance by a large margin. I use BP=0 and I think with a decent computer it is a preferable option.

Edited February 12, 201214 yr by TheGrunt

I had the same thing - and this stabilized it..[bufferPools]//UsePools=0UsePools=1PoolSize=30000

More testing required, but this might be the sweet spot....thanks, Bert.

Here's a bit of a chart to demo some possible expectations with say, a 4.4-gig proc and a 480 combination.

Hi Paul,How this chart was derived?I use two 580 at 4.4 and the chart indicates zero. Does that mean no advantage to use BP and RT?Pierre

HiDoes everyone here use this tweak in the cfg with a GPU of 1GB or more ? Are there any downsides to using this and does it increase your virtual memory ?Thanks

If you add the following tweak:[bufferPools]Poolsize=0You should notice a jump in FPS. I've seen it double from 15-30 fps. No joke! This is an amazing discovery.If you have issues, try this one:[bufferPools]Poolsize=30000I've been using it successfully with the Nvidia 275.33 and latest Inspector. You might be as fortunate as I am. I do NOT have to run the water at High 2x. That water setting is known to eradicate flashes, but it comes at a serious performance hit.Have fun!

Edited February 13, 201214 yr by DJJose

The only other tweak that I know of is the HIGHMEMFIX=1, in the [GRAPHICS] section, Wayne - this gets rid of some of the flashing problems, but there's nothing wrong in doing aUsePools=1PoolSize=8388608RejectThreshold=131072 // 524288 // 262144 //PoolSize=0

Can you please explain the difference between the different RT sizes?

Hi Paul,How this chart was derived?I use two 580 at 4.4 and the chart indicates zero. Does that mean no advantage to use BP and RT?Pierre

All of the "0"'s mean a setting of BufferPools=0. It means that your GPU can paint more screens than your CPU can push. It means you can overclock some more, dependent upon the CPU.

Can you please explain the difference between the different RT sizes?

The RT is the point at which any data size larger than that figure will be pushed directly to the GPU. It effectively throttles the data stream.The bigger the RT - the greater the quantity of data that will be buffered, which is the case when the proc is faster than the GPU. If the GPU is faster than the proc - then that data doesn't need to be buffered, does it? - so it's pushed directly to the GPU. Make sense?

The RT is the point at which any data size larger than that figure will be pushed directly to the GPU. It effectively throttles the data stream.The bigger the RT - the greater the quantity of data that will be buffered, which is the case when the proc is faster than the GPU. If the GPU is faster than the proc - then that data doesn't need to be buffered, does it? - so it's pushed directly to the GPU. Make sense?

Yes, so the 131072 figure will be the number to use that will most closely resemble having a bp=0 setup while keeping the bp at 1 for stability?

Edited February 13, 201214 yr by B777ER

Yes! You have it! Absolutely - the faster the GPU, the smaller this can be - BP=0 giving the best performance. The default is an 8MB BP, with an unlimited RT - which means everything gets buffered.

Yes, so the 131072 figure will be the number to use that will most closely resemble having a bp=0 setup while keeping the bp at 1 for stability?

Here's a simple test:Set the RT to 131072. Now go to your favorite busy airport and check the fps. NYC KLGA is a good place. Note the fps.Now set the BP to 0. Now check your fps.BP=0 gives the best and smoothest performance if your card can handle it.

Chaps...

For the first time I thought I'd give BP a try.

I have a GTX580 1GB. And i7920 at 4GHZ.

I tried...

[bufferPools]Poolsize=0At KSEA on the runway, with the NGX and major thunderstorms I saw 26-38 FPS. Exactly the same as without bufferpools.Memory usage according to GPUz was 485 without BP and 480 with BPGPU load was exactly the same at 27%.Am I using the right settings there, any advice, should i be using reject threshold or something?

Edited February 13, 201214 yr by martin-w

Chaps...

For the first time I thought I'd give BP a try.

I have a GTX580 1GB. And i7920 at 4GHZ.

I tried...

[bufferPools]Poolsize=0At KSEA on the runway, with the NGX and major thunderstorms I saw 26-38 FPS. Exactly the same as without bufferpools.Memory usage according to GPUz was 485 without BP and 480 with BPGPU load was exactly the same at 27%.Am I using the right settings there, any advice, should i be using reject threshold or something?

That's the wrong setting, you need[bufferPools]UsePools=0

Oops. Okay I'll give that a bash, thanks.

Now...FPS = 34-39Mem usage = 480Load = 26%Without BP was...FPS = 26-38Mem usage = 485Load 27%