Jump to content

Archived

This topic is now archived and is closed to further replies.

simmerhead

FSX vs. P3D benchmarking in progress

Recommended Posts

I just happen to have a brand new computer with a fresh Windows 7 install so I'm taking the trouble to benchmark FSX Acceleration vs. P3D.I've set up an autopiloted 25 min. flight @ 1000 ft. in the default King Air 350 starting at on approach to KJFK RW 4R, circling KJFK, heading towards La Guardia, then Central Park, shaving roof off skyscrapers over MIdtown and Downtown, circle Statue of LIberty, back to Central Park, circle Central Park, then return to KJFK and circkel the KJFK terminals one more time.Each flight is done with a freshly booted Windows 7. The computer (2700K, 16GB RAM, Fast SSDs and NVidia GTX 580) runs everything up to date in terms of Windows and drivers. No overcklocing or BIOS tweaks. The only software that has been installed is 3D Mark Vantage and a few other benchmark and stability testing software - To be 100% sure the computer has no driver or hardware errors.I run everything maxxed out - every slider. In P3D I rund 1024 textures and have disabled bathimery (or whatever it's called :) ). There are some sliders missing in Prepar3D compared to FSX, so the comparison is not 100% scientific. I have added the average fps count to the cfg's and take 12 measurements. Each measurement representing the lowest average fps I got f.ex. circling KJFK terminals. I've done each flight twice in both sims and there are slight variations on fps in each flight, but usually it is less than 1 fps difference.The result: FSX runs on average 0,5-1 fps faster on my system than P3D. FSX average for the flight is 9 fps, while P3D is just above 8 fps.Tonight I'm going to set up a flight in the Alps or Seattle and make the same kind of test there. Good region for testing water, mesh, urban and rural/forest areas in one flight.

Share this post


Link to post

Funny. Back when I was testing it, I ran with same settings and saw performance boost on average fsxmark11. Definitely had 10-15% better performance.Even if you have to set sliders lower, make sure they are the same. Be careful with water, p3d offers refraction, which should be off for fsx comparison as it doesn't exist in fsx.

Share this post


Link to post

Refraction, clarity and bathimery are disabled in P3D. I have even desabled the Bathimery layer in the scenery library. Frames are set to unlimited.And I should add, I fly in gray and rainy conditions to get the weather engine working a bit.All settings will be explained in my blog once the tests are finnished. I only posted it here now to get some feedback of things I might have overlooked while doing the testing, so thanks for the feedback Word Not Allowed!

Share this post


Link to post

Alright, looking forward to some comparisons!

Share this post


Link to post

Hi,Have you compared your FSX/P3D config files?. Make sure are no differences between the two such as Texture_Max_Load=4096 / Texture_Max_Load=1024.

Share this post


Link to post

Thanks Mike. Will have a look, but what I am after is "out of the box" performance with no tweaks or addons installed. Might be an interesting comparison nevertheless.

Share this post


Link to post

Hi,I'm not sure if the settings within P3D are set higher than FSX out of the box, if they are it wouldn't be a fair comparison. Same holds true if FSX's settings are set higher than P3D's out of the box. By checking the config files you will ensure a fair and even comparison.I wouldn't want to see you spend all that time working on P3D/FSX comparisons only to find out there were some config differences that may make the entire project moot.

Share this post


Link to post

With "out of the box" I mean tweaking everything to maximum settings from the custom menus - no manual cfg edits. Otherwise it would be a totally futile comparison of corse.

Share this post


Link to post

Here are the settings which differ from each cfg. FSX to the left. Please share som input if you think any of these will have significant impact on performance. Mind you. These are values created by the software, they have not been edited by me:

Share this post


Link to post

Only difference I see, which impacts performance, is water. I remember slider for reflection in P3D being different than water reflection in FSX.Otherwise nothing.You Nvidia CP is also reset for both apps?

Share this post


Link to post

Yep Nvidia is set to "Let 3D application decide".Yep. The watre settings are different, so I really don't know if they are directly comparable. I belive the two extra sliders in P3D only controls water settings when bathimery is active.

Share this post


Link to post

Well. I've finnished the benchmark. The conclusion:I see no difference in fps from FSX to P3D running at maximum settings. I even had P3D crash twice during the benchmarks, but FSX remained stable. Both programs have the same problems with stutters, blurries and black texture tiles when overloaded.The benchmark was done on a brand new computer with a fresh windows 7 install and no overclocking or tweaking to hardware or software whatsoever. Various other benchmark programs have been used to ensure maximum performance and stability of the computer.I did three 25 minute criss cross flights on each platform - all flights on full autopilot, except for take off. Before each flight I rebooted Windows to start fresh.I did 12 measurements at various waypoints during the flight. Best average fps was 15, lowest was 7. I'll write a full article on this in my blog when I get the time.So, on my computer P3D shows no improvement over FSX in terms of performance. In terms of stability and bugs FSX wins. With P3D I had two CTD, and experienced a few other bugs during testing, two of which I have reported to LM directly on their forum.

Share this post


Link to post

I thank you for this benchmark and look forward to results.

Share this post


Link to post

Hi,My system runs 30 FPS and rarely dips into the lower twenties. I'm surprised your FPS are dipping down to 7, guess it must be my overclock that is making such a difference.

Share this post


Link to post
Hi,My system runs 30 FPS and rarely dips into the lower twenties. I'm surprised your FPS are dipping down to 7, guess it must be my overclock that is making such a difference.
I'm running every slider to the right for benchmark purposes. Also, these are clean installs with no tweaks or addons. My benchmark flight is between JFK, La Guardia and Newark and the amount of AI traffic, dense autogen and weather brings the system to its knees!I will do a test with my computer OC'd, just for the fun of it, but I am interested in stock performance on stock systems. I don't OC my computer normally as it puts too much strain on my system which I use for other things than FS and I don't want to find some of my valuable photo and video files corruped becuase of a glitch in stability when processing them.

Share this post


Link to post

Simmerhead: but if I tried that on FSX it also would shudder to a crawl.I would really like to know at what settings you would judge P3D and FSX to be running smoothly. Possibly by holding traffic settings at some mid point, or lower, and then playing with the Complexity and autogen sliders.

Share this post


Link to post

Just some thoughts I had when I read the dismal results from both sims:The problem with running FSX without certain fixes and performance adjustments is that it does not take advantage of the knowledge that provides the much better performance we have gained over the years. When I (rather loosely) tested P3D I had the same performance enhancement using the same cfg and other adjustments. As long as both FSX and P3D had the benefit of having the same improvements it would be a fair and meaningful comparison.Having such low frames like 8 and 9 FPS, even under those demanding settings shows the futility of trying to run either sim "out of the box". It would seem to me that running the FSXMARK11 Benchmark test using high performance overclocked machines with the exact same parameters specified would yield interesting findings.http://forum.avsim.n...9116-fsxmark11/Kind regards,

Share this post


Link to post
Simmerhead: but if I tried that on FSX it also would shudder to a crawl.I would really like to know at what settings you would judge P3D and FSX to be running smoothly. Possibly by holding traffic settings at some mid point, or lower, and then playing with the Complexity and autogen sliders.
I have run the same test three times on BOTH FSX and P3D. It is a comparative test to see if one performs better and more stable than the other.

Share this post


Link to post

I appreciate what you are trying to do, Simmerhead, it's just that I am at the moment I am getting in my small opinion better performance out of P3D but not a lot more.But surely P3D has the potential to improve, FSX does not because it is metaphorically 'locked'.

Share this post


Link to post
I appreciate what you are trying to do, Simmerhead, it's just that I am at the moment I am getting in my small opinion better performance out of P3D but not a lot more.But surely P3D has the potential to improve, FSX does not because it is metaphorically 'locked'.
Yep, I heard that many got better performance out of P3D so I had to test it for myself on my own computer and I had no such luck I'm afraid...

Share this post


Link to post
Just some thoughts I had when I read the dismal results from both sims:The problem with running FSX without certain fixes and performance adjustments is that it does not take advantage of the knowledge that provides the much better performance we have gained over the years. When I (rather loosely) tested P3D I had the same performance enhancement using the same cfg and other adjustments. As long as both FSX and P3D had the benefit of having the same improvements it would be a fair and meaningful comparison.Having such low frames like 8 and 9 FPS, even under those demanding settings shows the futility of trying to run either sim "out of the box". It would seem to me that running the FSXMARK11 Benchmark test using high performance overclocked machines with the exact same parameters specified would yield interesting findings.
I'm not sure I would interpret the results as being dismal. This is a stresstest, not a test on optimizing performance. I wouldn't fly with all settings maxed out as that isn't very realistic, especially in terms of AI traffic.As for "out of the box" many knowledgable individuals like John Venema, Mathijs Kok and Nick N has a "less is more" approach to cfg tweaks and the effect of tweaks seem to vary greatly between different hardware setups.I was in a unique position to test the simulators on a brand new computer with a fresh Windows 7 install and fresh installs of both FSX and P3D. If you have tips for another test I'd be hyppy to try it before I clutter Windows and the sims with addons and get back to my normal flying. Again, I'm not looking for optimal performance, but comparing the two as fairly as possible.

Share this post


Link to post
I'm not sure I would interpret the results as being dismal. This is a stresstest, not a test on optimizing performance. I wouldn't fly with all settings maxed out as that isn't very realistic, especially in terms of AI traffic.As for "out of the box" many knowledgable individuals like John Venema, Mathijs Kok and Nick N has a "less is more" approach to cfg tweaks and the effect of tweaks seem to vary greatly between different hardware setups.I was in a unique position to test the simulators on a brand new computer with a fresh Windows 7 install and fresh installs of both FSX and P3D. If you have tips for another test I'd be hyppy to try it before I clutter Windows and the sims with addons and get back to my normal flying. Again, I'm not looking for optimal performance, but comparing the two as fairly as possible.
Hi Simmerhead,Yes, but there are a number of other prominent voices clearly heard above the noise, not just the 3 you mentioned. BTW, of those 3, only Nick's makes much sense, and he has a few things wrong too, like we all do. As I said, I recommend you run both simulations in the test program already developed in the AVSIM hardware section called FSXMark11. You could do it in both standard and an overclocked state to really demonstrate the differences under very high, but realistic settings and conditions. The real clincher is that the results could be compared with those of other systems as well.http://forum.avsim.n...9116-fsxmark11/Kind regards,

Share this post


Link to post

It is what it isn't. Its still a high priced FSX and maybe not even that! No one believes the proofs in the pudding? For us who would try it for an increase in performance this is very good to know. Saved us alot of time and money . Thank you for the test.

Share this post


Link to post
I'll see what my schedule permits Stephen!
Thank you Sir. I look forward to the results.Kind regards,

Share this post


Link to post
×
×
  • Create New...