On 5/12/2017 at 4:23 AM, GEKtheReaper said:

Hi @TechguyMaxC and thank you for sharing your experience. It's almost unbelievable that you gained 20 FPS, and this on avarage, using some of the tweaks. I tend to believe that there was something else causing this jump.

To clarify something: I'm in my 40`s, studied informatics and my dayly job is driving the future forward (so I'm not a kid trying to get some likes or hearts). I have to admit though, that I have no idea how FSX engine works and therefore decided to test known "tweaks" by my own. You can find my experience here:
 

If you read it you will see that all the testing has been done on clean OS installations, vanilla FSX:SE, same place, same time, same weather, same settings.... .

On my setup, FSX has it's own 64GB partition (which is also the first partition on my WD Black HDD) and I even formated my partition with a 64K allocation table (following the "FSX Bible"). IMO, you can't get a cleaner test environment then this!

Furthermore you might have noticed that even the framecounter FRAPS had a big influence on the core load.

To summarize it all up: I have not been able to notice any dramatic change in AVG FPS using any of the tweaks (or in combination). I noticed big changes in the CPU CORE load, some changes in smoothness (min/max FPS jumps), some stutters...but again...far from finding the holy grail of FPS increase.

***At this point I will again underline that I'm using FSX:SE. The boxed version is a mess!***

Since you find tons of users reporting tweaking the hell out of FSX:SE and tons of those who state the contrary, given my own test, I decided to take the unbelieving side. Furthermore, you will find developers of addons, shaders, etc. who also don't use that stuff in FSX:SE even though they allow you (through their interface) to easily add some of the tweaking lines to your cfg.

Today my RIG is fully setup with all addons installed and get with a newly build .CFG file, almost the same AVG FPS as with any of the tweaks above. But IMO the test with the fully loaded sim is not relevant because you have many variables which you can't realy control (how much traffic is loaded, how many clouds, are always the same type and amount of clouds loaded, etc.).

It even makes a big deal if you run DX9 or DX10 for testing, with or without the DX10 fixer.

 

Regards,
Gerald

Gerald, as someone who is intimately familiar with benchmarking procedures being a top contributor to the FSX Mark thread and having reviewed PC components professionally for a brief period, my methodology is sound.  http://semiaccurate.com/2010/11/22/amd-radeon-6800-series-review/

As I said, I have demonstrated the effects of these tweaks on 2 different hardware configurations, and have tested repeatedly to ensure accuracy.  The results are not anomalous.

22 hours ago, GEKtheReaper said:

@TechguyMaxCI wanted to edit last post but choose to write another reply since the other post was already too long.

Forgive my Q: Do you own FSX:SE? I will elaborate why I ask this Q...

You see, I'm kind of open minded so I went ahead and read some pages of your shared link to the FSMark. After 5 pages I decided to download the ZIP and have a look at the setup guide. Then I said: Oh DARN...this is all about the boxed version of FSX!

So another 30min and more of my time wasted...

The FSX Mark 11 benchmark thread is rather old at this point in time (the title of the thread and the date stamps on the earliest posts make this apparent).  The methodology established therein is still viable today as it is in no way constrained to any particular version of an ESP-based flight simulator, but rather applies to them all.  My recent testing (for which I reported results in my previous post) has been conducted with FSX:SE.  

 

The tone conveyed by the second half of your post is unwarranted.  You admit to not being an expert on this subject matter, you then make an assumption that I do not know what I am talking about because I linked you to a thread that began 6 years ago but is still active today.  What I actually did was link you to the most recent page of the thread and pointed to specific posts on that page which have been made in the past few weeks.  

Gents,

let's keep to the topic and don't make it personal please. 

Cheers,

Pete

@TechguyMaxCMate, I'm still trying to get from you the 1 piece of information you fail and fail to respond to. Do you have FSX:SE or do you have FSX Boxed? The same information was also missing from the other linked thread: I was unable to find anyone stating that they are using FSX:SE. Since this thread is related to FSX:SE, it was a waste of time to read it that far (so I was not picking on you).

 

As a matter of fact I realy did use the FSMark to test my system but I did not revert to a Vanilla setup (I already had to undergo this step 4 times in the last 3 weeks)!

Regardless of Vanilla or not, if the tweaks would work, they would be offering different results, which unfortunately they did not.

I ran each test at least 3x (5x max.) without any noticeably effect.

Tested: AffMask: 20, 84, 85, STD (no entry in cfg); Bufferpools: STD, 0, 1; With and without ChasePlane to the test (CP);

All my settings were maxed out without traffic enabled but with DX10 + ACFT shadows + following entries in cfg:

[GRAPHICS]

MultiSamplesPerPixel=4
MultiSampleQuality=8

 

[Display]

ForceVCShadowMap=1

[TERRAIN]

LOD_RADIUS=5.500000

 

Some of these entries were created with the DX10Fixer control panel, some of them inserted manualy.

The result:

Between each run with same config the AVG FPS fluctuates ~2FPS. All other configs are in the same AVG FPS range from 33 to 35. I will try to send you the FRAPS log via PM since I'm unable to load it up in this thread.

I will go one step further and offer you the opportunity to tweak my FSX:SE live via TeamViewer.

 

Regards

 

I can't send you the file so here the contents so far !!!PLEASE DON'T QUOTE THIS SINCE IT WILL FILL THE FORUM!!!

>>>FRAPSLOG<<<

Standard CFG

2017-05-13 10:14:39 - fsx
Frames: 10502 - Time: 300000ms - Avg: 35.007 - Min: 21 - Max: 47

2017-05-13 10:21:11 - fsx
Frames: 10413 - Time: 300000ms - Avg: 34.710 - Min: 22 - Max: 46

2017-05-13 10:27:15 - fsx
Frames: 10031 - Time: 300000ms - Avg: 33.437 - Min: 22 - Max: 45

2017-05-13 10:33:22 - fsx
Frames: 10349 - Time: 300000ms - Avg: 34.497 - Min: 22 - Max: 45
*****
AM=84

2017-05-13 10:41:58 - fsx
Frames: 10259 - Time: 300000ms - Avg: 34.197 - Min: 21 - Max: 45

2017-05-13 10:50:04 - fsx
Frames: 9889 - Time: 300000ms - Avg: 32.963 - Min: 21 - Max: 45

2017-05-13 11:02:42 - fsx
Frames: 10095 - Time: 300000ms - Avg: 33.650 - Min: 21 - Max: 45

2017-05-13 11:13:19 - fsx
Frames: 10426 - Time: 300000ms - Avg: 34.753 - Min: 21 - Max: 46
*****
AM=20

2017-05-13 11:22:41 - fsx
Frames: 10618 - Time: 300000ms - Avg: 35.393 - Min: 22 - Max: 45

2017-05-13 11:30:06 - fsx
Frames: 10565 - Time: 300000ms - Avg: 35.217 - Min: 22 - Max: 45

2017-05-13 11:36:22 - fsx
Frames: 10740 - Time: 300000ms - Avg: 35.800 - Min: 24 - Max: 46
*****
AM=85

2017-05-13 11:43:59 - fsx
Frames: 10236 - Time: 300000ms - Avg: 34.120 - Min: 21 - Max: 45

2017-05-13 11:50:17 - fsx
Frames: 10118 - Time: 300000ms - Avg: 33.727 - Min: 19 - Max: 45

2017-05-13 11:56:30 - fsx
Frames: 9988 - Time: 300000ms - Avg: 33.293 - Min: 21 - Max: 44
*****
AM=20 + CP
2017-05-13 13:28:24 - fsx
Frames: 10065 - Time: 300000ms - Avg: 33.550 - Min: 21 - Max: 44

2017-05-13 13:34:30 - fsx
Frames: 9854 - Time: 300000ms - Avg: 32.847 - Min: 20 - Max: 44

2017-05-13 15:33:22 - fsx
Frames: 10223 - Time: 300000ms - Avg: 34.077 - Min: 20 - Max: 45

2017-05-13 15:40:38 - fsx
Frames: 9983 - Time: 300000ms - Avg: 33.277 - Min: 21 - Max: 44

2017-05-13 15:47:10 - fsx
Frames: 9900 - Time: 300000ms - Avg: 33.000 - Min: 22 - Max: 43

2017-05-13 15:53:53 - fsx
Frames: 9976 - Time: 300000ms - Avg: 33.253 - Min: 22 - Max: 43
*****
AM=20+CP+Pools0

2017-05-13 16:46:47 - fsx
Frames: 10453 - Time: 300000ms - Avg: 34.843 - Min: 11 - Max: 44

2017-05-13 16:53:14 - fsx
Frames: 10124 - Time: 300000ms - Avg: 33.747 - Min: 22 - Max: 44

2017-05-13 17:00:13 - fsx
Frames: 9892 - Time: 300000ms - Avg: 32.973 - Min: 22 - Max: 43
*****
AM20+CP+Pools std (no entry)

2017-05-13 17:08:28 - fsx
Frames: 10240 - Time: 300000ms - Avg: 34.133 - Min: 22 - Max: 45

2017-05-13 17:14:19 - fsx
Frames: 10083 - Time: 300000ms - Avg: 33.610 - Min: 22 - Max: 43
*****
from here AM20 + CP only
*****
Watter effects lowered to 2xMid

2017-05-13 17:27:19 - fsx
Frames: 13860 - Time: 300000ms - Avg: 46.200 - Min: 27 - Max: 58

2017-05-13 17:34:02 - fsx
Frames: 13410 - Time: 300000ms - Avg: 44.700 - Min: 27 - Max: 57
*****
Above + Scenery lowered to Very Dense

2017-05-13 18:12:27 - fsx
Frames: 14137 - Time: 300000ms - Avg: 47.123 - Min: 27 - Max: 61

2017-05-13 18:18:49 - fsx
Frames: 13589 - Time: 300000ms - Avg: 45.297 - Min: 27 - Max: 59

>>>END OF LOG TILL NOW<<<

3 hours ago, GEKtheReaper said:

@TechguyMaxCMate, I'm still trying to get from you the 1 piece of information you fail and fail to respond to. Do you have FSX:SE or do you have FSX Boxed? The same information was also missing from the other linked thread: I was unable to find anyone stating that they are using FSX:SE. Since this thread is related to FSX:SE, it was a waste of time to read it that far (so I was not picking on you).

 

As a matter of fact I realy did use the FSMark to test my system but I did not revert to a Vanilla setup (I already had to undergo this step 4 times in the last 3 weeks)!

Regardless of Vanilla or not, if the tweaks would work, they would be offering different results, which unfortunately they did not.

I ran each test at least 3x (5x max.) without any noticeably effect.

Tested: AffMask: 20, 84, 85, STD (no entry in cfg); Bufferpools: STD, 0, 1; With and without ChasePlane to the test (CP);

All my settings were maxed out without traffic enabled but with DX10 + ACFT shadows + following entries in cfg:

[GRAPHICS]

MultiSamplesPerPixel=4
MultiSampleQuality=8

 

[Display]

ForceVCShadowMap=1

[TERRAIN]

LOD_RADIUS=5.500000

 

Some of these entries were created with the DX10Fixer control panel, some of them inserted manualy.

The result:

Between each run with same config the AVG FPS fluctuates ~2FPS. All other configs are in the same AVG FPS range from 33 to 35. I will try to send you the FRAPS log via PM since I'm unable to load it up in this thread.

I will go one step further and offer you the opportunity to tweak my FSX:SE live via TeamViewer.

 

Regards

 

As I said in my last post, the most recent testing I have conducted WAS PERFORMED WITH FSX:SE.  See the last line of the first paragraph:

6 hours ago, TechguyMaxC said:

The FSX Mark 11 benchmark thread is rather old at this point in time (the title of the thread and the date stamps on the earliest posts make this apparent).  The methodology established therein is still viable today as it is in no way constrained to any particular version of an ESP-based flight simulator, but rather applies to them all.  My recent testing (for which I reported results in my previous post) has been conducted with FSX:SE.  

 

The tone conveyed by the second half of your post is unwarranted.  You admit to not being an expert on this subject matter, you then make an assumption that I do not know what I am talking about because I linked you to a thread that began 6 years ago but is still active today.  What I actually did was link you to the most recent page of the thread and pointed to specific posts on that page which have been made in the past few weeks.  

As for your offer to tweak your config file for you, I'm not particularly interested, thanks.  I perform that service for clients and friends.  At the present point in time you are neither.

In your testing you cite separate entries for lowering water effects from maximum, which implies they were in fact at maximum in previous tests.  If you read the configuration guide for FSX Mark you can see on page 11 that the water effects are to be lowered to mid 2x in order to run the benchmark.  

Also, your average FPS never exceeds the magic 48 FPS "soft cap" that is present whenever Bufferpools is still enabled, that tells me you either don't have a fast enough CPU to take advantage of disabling Bufferpools or you did not disable it correctly.  Your signature states you have a 4790, but you do not list the clockspeed or specify if it is an unlocked "k" series chip.  If you are running a non-k 4790 that means your processor is running at between 3.6-3.8GHz during FSX Mark 11 testing.  If it is the k chip it will run at between 4.0-4.2GHz.  You can find this information here: http://www.intel.com/content/www/us/en/support/processors/000005523.html  The k chip should exceed 50 FPS average with a properly tweaked config file, reasonably fast graphics card, and no overclock.  With an OC close to 5GHz you can get into the mid-high 50's FPS-wise, given fast enough RAM.  Though Haswell/Devil's Canyon is constrained to the use of DDR3, which limits both minimum and average FPS compared to newer chips that can use DDR4.  Your highest minimum FPS result (27) is only about half that of mine (50) on Kaby Lake with 3600MHz RAM.  

"As for your offer to tweak your config file for you, I'm not particularly interested, thanks." --> My offer was not to help me but to see for yourself that the tweaks don't work.

"I perform that service for clients and friends.  At the present point in time you are neither." --> Your attitude speaks for your proffesionalism (maybe you should have read @Peter Z KCLE_EDDN post). To @Peter Z KCLE_EDDN: This will be my last answer to this post. Feel free to delete any parts that you find offending. I apologize to you and the users who have to read this.

"In your testing you cite separate entries for lowering water effects from maximum, which implies they were in fact at maximum in previous tests." --> correct!

"If you read the configuration guide for FSX Mark you can see on page 11 that the water effects are to be lowered to mid 2x in order to run the benchmark."  --> I started my post in saying that I did not test acc to the test settings. I actually have a fully loaded sim and I wanted to show you the effects of tweaks in this condition. Nobody flyes a Vanilla FSX.

"Also, your average FPS never exceeds the magic 48 FPS "soft cap" that is present whenever Bufferpools is still enabled, that tells me you either don't have a fast enough CPU to take advantage of disabling Bufferpools or you did not disable it correctly." --> I ran the test in DX10 mode. Default Bufferpools with no cfg change is at 1 and leads to ~30% off CPU CORE 1 load. UsePools=0 unloads first CORE to 20%.

"Your signature states you have a 4790, but you do not list the clockspeed or specify if it is an unlocked "k" series chip." --> Well if it's not followed by the "k" then it's not a "k" chip ;). 

"If you are running a non-k 4790 that means your processor is running at between 3.6-3.8GHz during FSX Mark 11 testing.  --> It goes up to almost 4GHz depending on CPU, No. of loaded CORES and Temperature

"If it is the k chip it will run at between 4.0-4.2GHz.  You can find this information here: http://www.intel.com/content/www/us/en/support/processors/000005523.html  The k chip should exceed 50 FPS average with a properly tweaked config file, reasonably fast graphics card, and no overclock.  With an OC close to 5GHz you can get into the mid-high 50's FPS-wise, given fast enough RAM.  Though Haswell/Devil's Canyon is constrained to the use of DDR3, which limits both minimum and average FPS compared to newer chips that can use DDR4.  Your highest minimum FPS result (27) is only about half that of mine (50) on Kaby Lake with 3600MHz RAM. " --> It's not a "k" and it's not OC. If I find the time and space, I will revert to Vanilla FSX and run FSMark as intended.

 

PS: I have overseen the fact that you have tested FSX:SE. I should have opened my eyes. Apologize for that.

 

Here's what you've done in your latest post, for anyone curious:

-Ignored official Intel documentation

-Misconfigured MULTIPLE test parameters (water effects, DirectX level)

-Questioned my professionalism

 

Nope, not interested.  Arguing with people on the internet isn't worth my time.  

And with that, I think this thread has run it's course.