If you ar right a uppgrade of a GPU from 570 to Titan or 780 do a 40% fps increasd?

And the titan windforce single cards performs aprox 90% of 2x 680 in sli..

Fsmark 11 test its no gain in avg fps 670,680,690 , Titan run on same system only swap

Cards.

That is not correct.

As I wrote above adding a second GTX680 to my current system and using them in SLI gave me 40% more fps in the 2d cockpit and in spotplane view.

 

Also the extra performance gain with a 4770k compared to a 2600k is minimal.

Did you use the same type / speed of memory ?

Were the mobo's the same quality ?

Tweaks ? Etc.

Your situation may be unique since you're not using the VC, but probably 95% of people do use the VC, so not sure about that one.

 

But my performance when SLI is enabled on my 780's is actually worse than running in single card config.

 

I have tested same RAM/Faster RAM the works. Mobos were of comparable quality and settings in FSX/Tweaks you name it were all identical. 4770K came out on top time and time again.

 

I don't mind admitting when I'm wrong, but in this situation the results spoke for themselves.

 

Westman has done considerable testing across all his systems as he's mentioned and his results are in line with mine also.

2d cockpit running scenery only will always give extremely good frames. Those running external hardware and multimonitors benefit greatly from no VC and gauges having to be rendered. gets VERY expensive though. I*ve considered that setup for yrs but just cant justify the 1000.00's in hardware, GoFlight etc.IMO, It's more of what being displayed than an SLI setup that causes increased performance.

Here's a link to that PCIe 3.0/780 thread I was talking about. 

 

http://forum.avsim.net/topic/425500-this-is-when-pcie-30-matters/

http://www.ocaholic.ch/modules/smartsection/item.php?itemid=1123&page=14

 

Quote :

"Having a closer look at the results we gathered while testing eight different games and two different benchmarks with two different presets, we see that the Core i7-4770K, with our "low-preset" is on average 6.4 percent faster than the Core i7-2600K.

Switching to our "high-preset" makes the Core i7-4770K's become 1.1 percent quicker than the Core i7-2600K.

Overclocking the Core i7-4770K to 4.5 GHz makes the performance with our "low-preset" go up by 11 percent but when it comes to the high-preset the increase in performance is only 1 percent.

 

Regarding the Core i7-2600K the situation is similar: 8.5 percent gain with "low-preset" and 1.3 percent with "high-preset".

 

What's quite interesting to see is how the wattage of our test system increased while maintaining 4.5 GHz stably.

The 4770K needed 40 percent more power and in case of the 2600K the increase was "only" 17 percent.

 

If you bought a Core i7-2600K a while back, then we can tell you without the shadow of a doubt, that there is absolutely no need to get rid of it and replace it with a Core i7-4770K.

 

From a gaming performance point of view it really doesn't matter which of the two CPU's you have in your system. Should you actually want to upgrade your gaming PC, then don't go for a new CPU in this case, you should rather choose a new graphics card."

 

For non FSX titles that are less CPU dependent than FSX that is true, as the "High preset" more GPU related test showed, nothing new about that at all.

 

For FSX though a 2600K has to run at about 5.2+GHZ or more to =  4770K@4GHZ, as the 4770K is much stronger clock per clock.

 

And sadly no, SLI does not do even squat for FSX, only slight edge in super high res multiple displays and ultra high AA,  this has been tested to death too many times already by almost everyone who has tested it including Aces.

For non FSX titles that are less CPU dependent than FSX that is true, as the "High preset" more GPU related test showed, nothing new about that at all.

 

For FSX though a 2600K has to run at about 5.2+GHZ or more to = 4770K@4GHZ, as the 4770K is much stronger clock per clock.

 

And sadly no, SLI does not do even squat for FSX, only slight edge in super high res multiple displays and ultra high AA, this has been tested to death too many times already by almost everyone who has tested it including Aces.

You have taken your first remark right out of the sky,

 

Your second remark is totally incorrect. This has been tested and SLI proved to be working in certain circumstances ( spotplane view and 2d without instruments ).

That is seperate from high AA and multi monitor setups.

Word Not Allowed has published about it a year ago.

A 5.2gHZ 2600k would be quicker than a 4Ghz 4770K. It would be more like 4.5-4.6Ghz on the 4770K.

 

His numbers may have been off, but what he was saying is correct.

I have done some crosschecks with Ant, Idao and my FSmark11 runs.

As a dont have run a 2600K but 2700K , 4770k @4.6 ghz performs as a 2700K @ 5.4 ghz..

Thats fact Gerard if you dont agree show some proof, dont take your numbers from the air.

 

I spend a lot of time testing different systems and cooling setups from air up to extreme.

yesterday a did some test with a 4770k , base for Vantage with HT enabled.

Just to clear up that iam not a completly noob.

test vantage base HT on

Hasse, I have read the tests you did and it where a lot.

I remember you got better fps after upgrading your memory to you current memory.

 

Others , after upgrading from a oc 2500 or 2600k syrems gained not much fps with their 4770.

Surely you will see some improvement with the current fastest processor, fastest mobo , fastest memory and fastest graphics card.

 

However , as far as I have read the tests, you with your oc 2700k got the same results as some others with a 4770.

For me even a gain of 10% in fps isn't worth a € 1000 upgrade.

 

Now I have 40% more fps for € 260 ( extra GTX 680 )

 

BTW: nice oc results

I72600k @4.8 Ghz, GTX680, DDR3@2133 CL10, SSD SATA III................couldn't be a happier simmer. Not planning any upgrade in the short term, unless I have to replace some hardware components for failure. 

Regards

http://forum.avsim.net/topic/412162-cpus-compared-clock-for-clock-fsxmarkcpu/

 

There you can see how the different CPUs perform at the same clock speed in a total CPU bound scenario in FSX.

 

Being CPU bound is what FSX is most of the time for most people in most scenarios. In particular the lowest FPS scenarios. There's a lot more than 1% difference between Haswell and Sandy Bridge in FSX then.

 

Most people say that SLI is broken in FSX. It's not. It works. But its only in very specific scenarios it helps. You've proven it yourself. You just need to be very GPU bound for it to help. SLI can also cause worse performance in other scenarios though.

 

As far as I'm aware FSX can be either CPU, GPU or PCIe bound. In scenarios when you are very PCIe bound for example, huge changes in CPU or GPU makes no difference. The same is valid for GPU and CPU bound scenarios as well. Out of the 3 I'd say that FSX is in general most of the time CPU bound followed by PCIe bound and the least of the time GPU bound. But your own specific use case may vary.

In terms of autogen performance, I believe that is PCIe bound. Faster bus means less stuttering (see link I posted on the first page). Autogen is the hardest hitter in terms of performance I believe. Water/shaders are somewhat, but DX10 helps that. 

 

Someone told me that in FSX the CPU calculates, GPU renders, and the PCIe bus is how fast the information gets passed to the GPU. 

 

If the PCIe bus is slow (< PCIe 3.0), then bus may be backed up because it can't push stuff fast enough to the GPU to render. Therefore, you'll have stuttering and decreased performance. I guess that is one reason to get Haswell over Sandy Bridge. 

You have taken your first remark right out of the sky,

 

Your second remark is totally incorrect. This has been tested and SLI proved to be working in certain circumstances ( spotplane view and 2d without instruments ).

That is seperate from high AA and multi monitor setups.

Word Not Allowed has published about it a year ago.

Word Not Allowed? NV? Total BS please dont make me laugh!

Word Not Allowed? NV? Total BS please dont make me laugh!

 

This sort of aggressive badgering contributes little to the discussion.  Voicing disagreement isn't a problem...doing it disrespectfully IS.  Please dial it back a notch.

 

Regards

 

 

My 4.5Ghz 4770K outperforms my 4.9Ghz 2600K.

 

The problem w/ all of this is that the 'outperforms' while absolutely correct is just not very relevant--the outperform is relatively small, and that equates to very little practical difference in running the simulation.  By comparison, for example, one can migrate to SSD and experience a huge difference in boot times & responsively, or if you take for example the T7 and run it wide open (no limiters of any kind except VSYNC ON, you will discover mouse pointer visible hovering over the VC will ding you by up to 35% over mouse pointer invisible.  That is a big difference whereas 4770K over 2600K is an effectively small difference.  I understand people want the best they can get but that's not the same as having effectively big improvements w/ these sorts of upgrades.