I was asked earlier to test via passmark. I did it back on day 1 but I've got the computer and results at hand again.

 

My passmark score was 6541.

 

Here's some further details:

 

6541

General Information
PerformanceTest     Version 8.0.1043
Baseline ID     346003
Operating System     Windows 7 Home Premium Edition Service Pack 1 build 7601 (64-bit)
Submission Date     7th of January, 2015
Benchmark Scores
Passmark Rating     6,541     full starfull starfull starfull starfull star
CPU Mark     16,904     full starfull starfull starfull starfull star
2D Graphics Mark     1,175     full starfull starfull starfull starfull star
3D Graphics Mark     10,261     full starfull starfull starfull starfull star
Memory Mark     3,035     full starfull starfull starfull starfull star
Disk Mark     4,935     full starfull starfull starfull starfull star
Hardware
CPU Type     Intel Core i7-5820K @ 3.30GHz
CPU Measured Speed     3.30 GHz [Turbo: 4.40 GHz]
Motherboard     X99 Extreme4
Memory     16GB , G Skill Intl F4-2400C15-4GRB
Video Card Tested     GeForce GTX 970
Hard Drive Tested     Samsung SSD 840 EVO 250G (250 GB)

 

And a link to the full results:

 

 

Anyone have an i7 4790K or similar and want to post their results?
 

Core usage in P3D:

 

Here's the results when I'm using the sim as I typically would:

 

P3D v2.4

Plane: A2A 182 with Flight One GTN 750

Scenery: Megascenery earth Switzwerland, Aerosoft Night Environment

Weather: ASN, some clouds and haze

Autogen: lots, flying over Geneva using OpenVFR for trees and buildings. Sliders to the right.

Frame rate 35-45 unlocked

 

Screenshot of this - my attempt to add the picture didn't work, but here's the URL:

 

https://www.dropbox.com/s/9dqxlw5qa8ovaah/P3D%20core%20usage%20Open%20VFR%20A2A%20GTN%20ASN%20AS_NE%20MSE.png?dl=0

Comment: I'm actually quite impressed by how much the 12 threads are used. Sure, core 0 is still maxed out but the other 11 cores/threads are all solidly utilised. This suggests to me that in real-world usage, six cores is better than four, and hyperthreading is helpful. If true, these are both very relevant points with regards to cpu chip selection.

 

With regard to core usage, note that I haven't set an affinity mask. Does anyone think that would help in this case?

 

With regards to passmark, I thought my URL would link to the full results which include single thread performance. Looks like it just gives me a badge with my score! Haven't used pasmark before, just ran as requested but I'm quite impressed. Will try and rerun when I get a chance and link to or screenshot the full results. If anyone wants to do the same with one of the main alternative P3D chips I think that would be interesting data.

Thanks. What was the single core score?

As per above not sure, still travelling so only have my computer intermittently - but will try and get back to it on the weekend and find that number for you.

@OzWhitey,

      Thanks much for all the details on the 5820! I haven't seen a better review for us P3Dv2.4 folks. I would also like someone to provide  4790K Passmark and results for  given scenario (fxml FLT file) that we all could use as a quick performance test - somewhere without addons - standard but high density weather theme and all sliders to the right of average but not excessive traffic (which would dominate).

      Also, I found that some payware scenery  PAULO RICARDO FSX - MEGA ROUTE RIO at the following:

http://secure.simmarket.com/paolo-ricardo-fsx-mega-route-rio-sao-paulo-and-natal-city-pack-5.phtml

 

is amazing looking in P3Dv2.4 with volumetric fog and a little weather - AND has excellent FPS - even with all the well textured buildings - maybe you might like it.

 

whitav8

Comment: I'm actually quite impressed by how much the 12 threads are used. Sure, core 0 is still maxed out but the other 11 cores/threads are all solidly utilised. This suggests to me that in real-world usage, six cores is better than four, and hyperthreading is helpful. If true, these are both very relevant points with regards to cpu chip selection.

 

I am not sure that core usage translates into useful performance.

 

The "too many cooks in the kitchen" saying comes to mind.

 

Maybe there is a lot of thrashing, but not necessarily more output.. :huh:

Thank you for this test, 

I actually bypassed the 5820K and went for the 8 core 5960X i7 on the X99 Mobo.

I understand the clock speed stock is only 3.0Ghz but somehow some way the sim still seems far smoother performing than my i7 4820-k or my i7 4770-k at stock speeds.

 

I have never been one to overclock because I have always been afraid to smoke a chip setting a value I know nothing about too high though I would really like to try so I will give it a try using a cheaper chip, I will do it with the 4770-k as it seems to be the lowest performer of all 3 and if I lose that one at least it won't hurt as bad.

 

That said with any of those 3 cpu's frame rates were never the issue,the only issue was scenery loading times and with the 5960X scenery loading is practically instant with each view and each head movement which kind of took me by surprise as I am sure most people notice when changing views you get that momentary washed out cockpit for a second or two then it catches up but with the 5960X that hasn't happened and as I said the scenery is just there no loading wait, no blurry textures .

 

Somehow I wonder if I could have had the same performance just using the 5820K ?

But oh well what's done is done.

I do believe my X99 system is future proof for quite a while. 

I am only running 16 gig of DDR4 @ 2100 Mhz

the GFX is GTX 970 4 gig 

 

I run all high detail aircraft PMDG, VRS Simulations and Iris F15 with M2M add on 

REX

Mega Scenery X  scenery along with countless downloads from Blue Sky Scenery because they're pretty damn impressive,Track IR and probably 5 or 6 other add ons I am forgetting about like Night lighting and things like AI aircraft.

 

For me as long as I am getting 25-30 FPS in Manhattan and my scenery is loading the way it is I don't think I could possibly get any better i see no difference in flight controls or attitude from 25-30 or 40 -60 FPS  at less dense airports in cockpit FPS jump into the 120's which to me is a waste and could probably use this energy elsewhere ?

 

Anyway, that's my 2 cents.

I appreciate the technical aspect of your post because I surely do not have the time to sit there and monitor all of that mumbo jumbo

 

Edited to add :  I have been reading the forums for quite a while picking up scraps of information that I have found extremely useful but have never really been inclined to comment until I read this post so I thank you for pulling me out of the weeds so to speak

Edited January 26, 201511 yr by HesTNTonPMS

Are you still happy with 5xxxk cpus? I have [email protected] and im thinking about upgrading to 5820k.

Hmm well this is an old thread, but as I have been spending a fair bit of time trying to get my 5820K based rig working with P3D 2.5 I thought I would post.

 

I have been really happy with it, mostly flying an A2A Cessna C172 with Orbx addons and regions. I have a GTX980 as well, and generally with the C172 frame rates are enjoyably high, when I bother to look I am usually seeing 50-75 out of town, with maybe dips over densely populated areas down to 35+ if I fly low enough.

 

I have bought myself the aerosoft A319 and paired it with Orbx Palm Springs. I have not put my REX addons in yet, but I have been experimenting with maxing sliders and seeing what happens. With everything absolutely max I am seeing about 20+ FPS, but the dips are annoying. Scaling back some sliders a little boosts the frames up into the 25-30 region. Resolution is a weird custom one, about 1920 x 1500 as I am waiting for a new 2k monitor and am using an old TN panel with NVidia DSR.

 

I don't have software to take trends / data, so if someone can recommend me a good program I will collect some figures. Looking around at YT videos made with 4790k's with the nuts OC'd off them, at least from a subjective 'smoothness' perspective I think the 5820K really does quite nicely.

 

I am running mine at 4.5GHz and with HT off, which really seems to make a difference. Havent bothered with affinity masks, though P3D is using all the cores. Not tried FIBER FRAME FRACTION yet either, but OPTIMIZE PARTS does nothing at all. Core 0 is indeed maxed, but that is true no matter which aircraft I use.

 

As an aside, the sim looks BEAUTIFUL. I cannot wait to get my monitor.

 

EDIT: Just quickly took the airbus to Heathrow for a take-off, with Orbx Open LC Europe handling scenery duties. Still with everything, and I mean everything, at max, min FPS was 13 but mainly hovering around 20. Not where you want to be for simming, but provided here as an indicator of power.

 

Now if I can just get rid of the annoying terrain shadow flashing problem, I will be very happy.

 

SPECS: 

5820k @ 4.5GHz

16GB Corsair Dominator Platinum DDR4 @ 2666GHz

2 x EVGA GTX 980 SC ACX 2.0 (though not SLI'd, am under impression P3D doesnt like it)

Samsung 850 Pro 512GB, at least for sim duties

H105 AIO cooler

Superflower Leadex Platinum 1kw

NZXT H440

 

Various controllers.

 

EDIT 2: Hmm it seems that having hyperthreading off has actually given me some more headroom with my overclock. Just pushed it up to 4.625GHz, and FWIW is stable on a 15 minute realbench test (usually fell over immediately above 4.5GHz before). With P3D that seems to have given me a buffer over 20 FPS with everything still at max (graphics, not tried AI yet).