Sign in to follow this  
Michael Moe

How is terrain loading on those I9?

Recommended Posts

Hi everyone.  Even with 4.4 which is the best sim yet i still struggle a bit with terrain loading under heavy load. Nothing comes for free offcause and its quite natually i guess. 

I am still on the old I7-4770K@4.6 with DDR3 2400 ram and using the FSL A319 in combo with FTX England, FTX German is to much on my system. I need to reload textures on arrival twice (AS EDDK 4K) which made me wanna ask how those I9 handles terrain with a jet flight from one heavy area to the next in a complex aircraft like the FSLA320 series? 

Using FFTF at 8500 feet with a 0.6

Thanks Michael Moe

Edited by Michael Moe
  • Like 1

Share this post


Link to post
Help AVSIM continue to serve you!
Please donate today!

I have been using my 9900k with 16 gb Trident Z at 3600 and a 1tb NVME PCIE  for a couple of weeks now getting P3D tuned in.  I don't overclock but use the XMP 2 (Asus Maximum XI Hero (wifi)), no HT, all 8 cores run at 4.7 and it seems to load a little quicker than my 7700k but I don't run addon airports so YMMV.  I do use global, vector, and landclass.  Now only to get the 2080ti so overcast conditions will not kill my performance, the old 1080ti just cant do it.

 

 

Edited by sgreen91
  • Like 1

Share this post


Link to post

Hi Michael,

I upgraded my CPU from a 4790k to a 9900k. At the same time is switched from Win7 to Win10 and from P3Dv4.3 to v4.4. So there are some more factors that might contribute to the outcome other than the CPU change.

With the old 4790k I had to choose between blurries + smooth and microstutters + sharp textures. Always looking for the best compromise.

Now I‘m totally happy with my new 9900k OC @5GHz (HT=OFF) together with 32GB RAM @ 3200MHz. Since I use a 4K Monitor, 30Hz Refresh is vital to avoid any stutter issues for me (P3D setting: vsync on + triple buffer on). Even with the FSLabs A320 and demanding Airports my system runs smooth with sharp textures almost any time. Some areas like LAX etc. generate single hickups on very rare occasions.

MSI Afterburner shows core zero around 75% in use on average, the other seven cores are evenly active from anything between 20% up to 90%. That‘s with Active sky, TrackIR and UTL (40% Airline, 30% GA).

So I hope this helps to make a sound decision!

Regards Martin

 

 

 

  • Upvote 1

Share this post


Link to post

9700K at 5.1ghz upgraded from a 7700K at 5ghz. Terrain loading much improved.  I ran the 7700K with HT enabled.  Why everyone wastes money on the 9900K is beyond me.

 

  • Like 1
  • Upvote 1

Share this post


Link to post

Would you guys with 8 and 9 series CPU's mind downloading and running the Passmark CPU single Threaded test.

The test can be downloaded and run for free from the PassMark website here:https://www.passmark.com/download/pt_download.htm

Once downloaded you can install and run. The GUI opens and a registration window opens on top of that. Just click continue on the registration window to close it.

In the GUI you wills see a spinning motherboard showing your installed components and on the left of the window you will see a column showing different benchmarks. The second box down which is yellow in color is CPU mark. Click on the lighter colored part of that box. It's the top part and much bigger than the smaller part. A gauge will appear in the GUI and below the gauge 9 buttons for each of the CPU testc.

The CPU single threaded test is the one I am interested in. The one on the right of the GUI.


Make sure there are no other programs running and click run. A window will open indicating that the test is running. It only takes about 20 seconds.

Please post your CPU, The frequency that you are running it at (the main core) it they are not all the same. And of course the Score you got in the test.

My CPU is i7 4770K @ 4.8ghz (all cores) and the score that my CPU got is 2890

It is not specifically relevant to P3D but some lose correlation can be made to performance.

I ran an identical scenario with my CPU at 3.9ghz and 4.8ghz and also ran the above test at both frequency's the test score and frequency correlated around 24% between higher score and lower score and higher frequency and lower frequency. The difference in average FPS was 18%.

I really like to see what score 8 and 9 series CPU's get for the purpose of making apples to apples comparisons between an i7 4770k at 4.8ghz and for example an i9 9700K at 4.8ghz and you specific max stable clock. The i9 9900k would be great to see too. And 8 series result.

I keep putting of pulling the trigger on a new build but its definitely getting close.

But there will probable be a GTX 1180ti soon and 10nm intel cpu's by the end of this year. My I7 4770K at 4.8ghz has server me very well for the last 5.5 years. I think I can stick out the next year with perhaps only buying a used 1080ti or a GTX 1180ti if they show up soon.

Share this post


Link to post

Thanks every one. Nice to see those new CPUs might be mandatory for latest P3D performance. 

With more and more features also in study level aircrafts its nice to know that those cores gets used. 

Also latest SDD drives helps with loading also i guess. They even seems to be at a very resonable price as well. 

Might be the year for an upgrade. 

Thanks Michael Moe 

Share this post


Link to post
1 hour ago, Avidean said:

Would you guys with 8 and 9 series CPU's mind downloading and running the Passmark CPU single Threaded test.....

Dave, pm sent....

Share this post


Link to post
3 hours ago, Avidean said:

Would you guys with 8 and 9 series CPU's mind downloading and running the Passmark CPU single Threaded test.

The test can be downloaded and run for free from the PassMark website here:https://www.passmark.com/download/pt_download.htm

Once downloaded you can install and run. The GUI opens and a registration window opens on top of that. Just click continue on the registration window to close it.

In the GUI you wills see a spinning motherboard showing your installed components and on the left of the window you will see a column showing different benchmarks. The second box down which is yellow in color is CPU mark. Click on the lighter colored part of that box. It's the top part and much bigger than the smaller part. A gauge will appear in the GUI and below the gauge 9 buttons for each of the CPU testc.

The CPU single threaded test is the one I am interested in. The one on the right of the GUI.


Make sure there are no other programs running and click run. A window will open indicating that the test is running. It only takes about 20 seconds.

Please post your CPU, The frequency that you are running it at (the main core) it they are not all the same. And of course the Score you got in the test.

My CPU is i7 4770K @ 4.8ghz (all cores) and the score that my CPU got is 2890

It is not specifically relevant to P3D but some lose correlation can be made to performance.

I ran an identical scenario with my CPU at 3.9ghz and 4.8ghz and also ran the above test at both frequency's the test score and frequency correlated around 24% between higher score and lower score and higher frequency and lower frequency. The difference in average FPS was 18%.

I really like to see what score 8 and 9 series CPU's get for the purpose of making apples to apples comparisons between an i7 4770k at 4.8ghz and for example an i9 9700K at 4.8ghz and you specific max stable clock. The i9 9900k would be great to see too. And 8 series result.

I keep putting of pulling the trigger on a new build but its definitely getting close.

But there will probable be a GTX 1180ti soon and 10nm intel cpu's by the end of this year. My I7 4770K at 4.8ghz has server me very well for the last 5.5 years. I think I can stick out the next year with perhaps only buying a used 1080ti or a GTX 1180ti if they show up soon.

9700K 5.1ghz all cores.  3099

at 4.8ghz 2919

Edited by mpw8679

Share this post


Link to post

Not a 9xxx series CPU, but for a point of reference my 8086K @ 5.4 GHz all cores scored 3275.

Greg

Share this post


Link to post
26 minutes ago, lownslo said:

Not a 9xxx series CPU, but for a point of reference my 8086K @ 5.4 GHz all cores scored 3275.

Greg

Can u run the test at 4.8ghz?

Share this post


Link to post
2 hours ago, mpw8679 said:

Can u run the test at 4.8ghz?

2915

At 5.1GHz... 3094

Edited by lownslo
Add 5.1 score

Share this post


Link to post
22 minutes ago, lownslo said:

2915

At 5.1GHz... 3094

I knew they would be close but wow.

Share this post


Link to post

Gigabyte Z390 AORUS Pro

Windows 10 Pro

i5 9600K@4.8

Score: 2913

 

Edited by TuFun
closed programs rerun test

Share this post


Link to post

Don't read to much into those scores. As I said there is no direct correlation between this test and overall Prepar3D V4.4 performance. My own experience is that P3D V4 is far more multi threaded than it used to be. In my case I use HT off But I regularly see core 0, 2 and 3 at 100%. The upshot of that I think is that 4 cores just doesn't quite cut it for Prepar3D anymore. So regardless of OC to get the best blurry free performance you need those extra cores. I am leaning toward 8 cores so the i9 9700K is appealing. I won't use HT!

Thanks for sharing. As I suspected later generation CPU than mine do score better at the same frequency than my  i7 4770K but not that much more. The real meat and potatoes is the extra cores.

5 hours ago, lownslo said:

Not a 9xxx series CPU, but for a point of reference my 8086K @ 5.4 GHz all cores scored 3275.

Greg

That's a golden chip you have there and you can see that at 4.8ghz there is only 6 point in the difference between the i9 9700K and only 23 point ahead of mine. Of course I cannot get 5.4ghz.🤣

Even so that's only 13.5% ahead of mine. I'd be satisfied with that especially with the extra 2 cores which I think would be more important at this point in the development of P3D. However the odd of getting a chip like that are slim I think.

For now I'll stay put and get a used 1080ti to replace my 980ti. Since I Sim only in VR, Autogen density and LOD and AG draw distance is far less important given the soft image in the Samsung Odyssey+. However, I really need 8xSSAA which is just a little out of reach for the 980ti.

Edited by Avidean

Share this post


Link to post

8700K @ 4.8 ghz

2915

This is interesting, as the 8086K and even 9-series are all showing about the same single-thread.

I think I'm about a year away from a rebuild.  Year and a half maybe.

Share this post


Link to post

8086K no HT; 3084 single threaded. that's on a ROG max code X board on which I ran the auto-overclock software. boot screen says 32% - not exactly sure what that 32% relates to (~5300 clock observed) but... HT on the best it'll do is 25%. not optimized beyond that and it's still cool running which is nice.

Edited by jpf

Share this post


Link to post
On 1/27/2019 at 10:30 AM, lownslo said:

Not a 9xxx series CPU, but for a point of reference my 8086K @ 5.4 GHz all cores scored 3275.

Greg

5.5 3345  bonus 5.5 stuff: cpuz 666 / cinebench 1/1 241 - all cores 1 thread 1405 / Realbench "Imaging" 306 17.3 sec / Aida 4200 c17 62GB read/write 36.5ns

5.4 3285

5.0 3040

4.8 2920

 

Edited by FunknNasty

Share this post


Link to post

correction:

single core cinebench score is actually 245  .....the multi core score is at 5.4 (could not complete run multi at 5.5).  ----sorry

Share this post


Link to post

So it comes down to: is main thread overclock potential of number of cores more important for P3D. Looks like 8086 based on the results above has a better chance of reaching 5.4ghz than the 9 series cpu's but the 9's have 8 cores compared to the 6 cores of the 8's

Makes sense that 6 cores have more OC potential than 8. I wonder are 8800K users getting 5.4ghz too or is it just 8086K users.

Still I am going to hold off until the 10nm chips come along. At 4.8ghz on my current system since I am using VR I think I might be underwhelmed by the return on my investment. But I will treat myself to 1080ti in the next week or so 😀

Thank you all for sharing your results you bunch of Lucky word not allowed.🤣

Edited by Jim Young
Inappropriate language, even when masked, is not allowed here at AVSIM. Last offensive word removed...

Share this post


Link to post
33 minutes ago, Avidean said:

Makes sense that 6 cores have more OC potential than 8. I wonder are 8800K users getting 5.4ghz too or is it just 8086K users.

I'm only familiar with the 8700K -- but I think westman and at least one other got around that with 8700K's but you would need to check with them for sure.  I'm not sure if they were delidded or not, I would think you would have to be to get that high.

Share this post


Link to post

Yeah, I'd be shocked if the my 8086 couln't do P3D stable at 5.5 (1.55v)  with little or no avx offset  -with a real cooler using the mobo's LN2 mode. I did boot to 5.6 {1.6v) today, but cpuz crashed at the bench run ....I do think 5.6 benches are possible if I use the LN2 mode to remove some of the circuit breakers. :-)

Share this post


Link to post

Thought I'd post picts of todays work ....mostly pointed at Westman: <g>

https://imgur.com/Roi0ZeP

 

note:every thing in picture is at 5.5/5.0 ring 4200 C17

Edited by FunknNasty
  • Like 1

Share this post


Link to post
4 hours ago, Mace said:

I'm only familiar with the 8700K

Yeah that's what I meant to type. If I was going to build now that's what I would go for or an 8086K. In VR 45fps is the new 30fps and I am finding that I can manage that on fairly low Autogen  and LOD settings. But with the soft image of the Samsung Odyssey+ the difference between low and high autogen and LOD setting is not as stark as it is on a monitor. However, I find you really need 8xSSAA. I find that high AA setting are much more necessary in VR than on a monitor. Also other GPU intensive features such as shadows, quality and draw distance and water detail and tessellation are much more GPU dependent than CPU dependent. In fact tessellation takes some of the pressure of the CPU. Anyway I digress. Definitely, the smart move for me at this point given my usage is a 1080ti. The 980ti just isn't cutting it. But I am repeating myself.

Edited by Avidean

Share this post


Link to post
4 hours ago, Avidean said:

Yeah that's what I meant to type. If I was going to build now that's what I would go for or an 8086K. In VR 45fps is the new 30fps and I am finding that I can manage that on fairly low Autogen  and LOD settings. But with the soft image of the Samsung Odyssey+ the difference between low and high autogen and LOD setting is not as stark as it is on a monitor. However, I find you really need 8xSSAA. I find that high AA setting are much more necessary in VR than on a monitor. Also other GPU intensive features such as shadows, quality and draw distance and water detail and tessellation are much more GPU dependent than CPU dependent. In fact tessellation takes some of the pressure of the CPU. Anyway I digress. Definitely, the smart move for me at this point given my usage is a 1080ti. The 980ti just isn't cutting it. But I am repeating myself.

Just wondering whether you mean you need 8xMSAA? 8xSSAA seems an impossible target in VR?

Share this post


Link to post

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
Sign in to follow this