i9 9900k

[Richy93]

On 8/12/2019 at 11:39 PM, Doug Sawatzky said:

 

Of course we test our sim after upgrade...do you? Why wouldn't they translate into P3D performance? Are you implying FS is optimized only for Intel’s single threaded performance?

 

Where some Ryzen CPUs score better in CPU-Z and Cinebench single thread benchmarks, Intel (9700k / 9900k) produce more FPS in most games. So unfortunately, CPU-Z & Cinebench benchmark scores are not directly comparable with P3D performance.

 

I’m really interested to see a comparison in P3D V4 performance, between the Ryzen 3800x/3900x (with PBO auto OC) and Intel’s 9700k/9900k (at 5GHz).

[AnkH]

47 minutes ago, Richy93 said:

 

Where some Ryzen CPUs score better in CPU-Z and Cinebench single thread benchmarks, Intel (9700k / 9900k) produce more FPS in most games.

 

That is exactly what I meant. And for certain games, it seems to be still the case:

 

https://www.guru3d.com/articles_pages/amd_ryzen_5_3600x_review,9.html

 

While the Ryzen 3600X is just lousy 5% slower in the Cinebench15 single-thread run, but better in the IPC run and even slightly better (0.6%) in the Cinebench20 single thread run, the same processor is significantly slower in certain games (FarCry: 18% slower, Shadow of the Tomb Raider: 30% slower) and slightly faster in some other games (Strange Brigade: 3% faster).

 

Means: those numbers basically do tell us nothing in regard of Prepar3d.

[lawrence aldrich]

On 8/13/2019 at 1:51 AM, AnkH said:

 

Of course not. But it would be not the first time in processor release history that a processor is by far not that capable as benchmark results suggest. That is why. Those benchmark numbers are sometimes representative of the performance in Prepar3d, sometimes absolutely not.

 

Besides that, the FS community still lacks a lot of knowledge and experience with AMD processors. The more people like you "publish", the better this situation gets.

Well...as a suggested solution, one can always just buy one and check it out.

Just saying.....

 

[AnkH]

Maybe if one has to much time and money, yes. Here in Switzerland, there is in contrast to the EU no possibility to buy computer hardware and return them within 14d in any case. So, no, I can not just buy a Ryzen 3000X build just to test if P3Dv4.5 runs well on this hardware, especially not if people are around that DO have such a hardware combo.

 

So, I disagree.

 

Better than "buy one and check it out" is ask those that have one...

[lawrence aldrich]

54 minutes ago, AnkH said:

Maybe if one has to much time and money, yes. Here in Switzerland, there is in contrast to the EU no possibility to buy computer hardware and return them within 14d in any case. So, no, I can not just buy a Ryzen 3000X build just to test if P3Dv4.5 runs well on this hardware, especially not if people are around that DO have such a hardware combo.

 

So, I disagree.

 

Better than "buy one and check it out" is ask those that have one...

OK.... But you are still assigning somebody else to positively evaluate a subjective experience.

Cheers!! 

[Doug Sawatzky]

On 8/15/2019 at 8:39 AM, AnkH said:

Better than "buy one and check it out" is ask those that have one...

 

What would you like to know specifically? Usually anything I suggest, you dispute and berate anyway   

Like I have stated many times, my 1800X, 2700X, and now 3800X have all ran all the flight simulators very well.

 

Do you consider the 7700K and 8700K able to run P3Dv4 very well (they are not ahead in gaming)? Is the 9900K faster in most games? Yes it is, but as you know gaming benchmarks don't necessarily tell the whole story all the time either because they are GPU bound at different levels.

So I hope you will find the below videos helpful to compare where things are at.

 

https://www.youtube.com/watch?v=GpE2UAlhzVM

https://www.youtube.com/watch?v=UsdKnnX4se0

 

Ryzen clock for clock IPC performance has already surpassed Intel with this latest generation.

I can assure you the upgrade to the 3800X (with the single threaded performance increase) has translated into performance gains in P3Dv4.

And yes, the situation will be getting better very soon.

 

https://www.youtube.com/watch?v=Eg4G4YjWH5w

 

 

 

[Richy93]

1 hour ago, Doug Sawatzky said:

Do you consider the 7700K and 8700K able to run P3Dv4 very well (they are not ahead in gaming)

 

The 7700k is litterly from Q1 2017 whilst the 3800x is from Q3 2019, not really a fair comparison. Take the 9700k instead..

 

1 hour ago, Doug Sawatzky said:

Is the 9900K faster in most games? Yes it is, but as you know gaming benchmarks don't necessarily tell the whole story all the either because they are GPU bound at different levels.

 

Very true. But CPU tests are mostly performed with an overpowered GPU at “low resolution” (read: 1080p), because at higher resolutions, the GPU becomes the bottleneck. This way the CPU is the bottleneck during the benchmark.

[Doug Sawatzky]

2 hours ago, Richy93 said:

The 7700k is litterly from Q1 2017 whilst the 3800x is from Q3 2019, not really a fair comparison. Take the 9700k instead..

 

Of course, this is only intended as help to answer AnkH's question about single threaded performance gains of the new Ryzen.

 

https://www.youtube.com/watch?v=phKLyU2ObEw

 

[Bullfox]

Doug,

I am considering a build using the Ryzen 3800, or maybe the Ryzen 3900 but not sure if the 3900 is worth the extra $150.00.  Also are you using the Wrath cooler that comes with the Ryzen even with overclocking.  Whats your best overclocking speed so far with the 3800?

[Doug Sawatzky]

Hi

 

I am not using the Wrath cooler, I already had a 280mm Thermaltake 3.0 AIO with temps never exceeding 70c under any condition, and I am just using a memory XMP profile and PBO enabled in the bios, so no manual overclocking, and I am getting a 4.5ghz as advertised boost on one or more cores at a time. I personally don't think the 3900x is going to provide any more single threaded performance for FS than the 3800x. 

 

 

 

[Dusterman]

And now Intel's Cascade Lake-X to compete with AMD, coming to a computer supplier near you in 2020. Maybe I will have to upgrade my Intel X79 setup:)

[Doug Sawatzky]

Unfortunately it’s just a regurgitation of the same old 14nm architecture, just a couple more cores to use more power and make more heat.

[Dadtom65]

Hi guys. Myself also have been thinking of updating my comp and did look at the i9 9900K but read some of the remarks of the people that bought said Cpu and they said that they had trouble with the temperature of it. Also looked online and there are a lot of people saying about keeping the dam thing cool. Also Doug read about the new Cpu but that’s not coming for awhile yet. So not quite sure what to do yet. Derek.

[Doug Sawatzky]

Hi Derek

 

There is a full lineup of Ryzen 3000 series CPU's already available that perform as good if not better than anything else, and I have been able to use my same mother board for the 1000, 2000, and now 3000 series of Ryzen CPU's, and yes the 4000 series Ryzen is still in the near future.

[Dusterman]

And to make matters more complicated, DDR5 motherboard RAM is expected to become available this year. I am a platform agnostic, but if you want to use this RAM it looks like AMD will becoming out with CPUs using it. I have data intensive applications that don't use a large number of cores, so more interested in speed than number of cores.

[Doug Sawatzky]

Also, in regards to speed, there is the PCIe 4.0 being offered as future proofing from AMD.

[ozboater]

On 1/4/2020 at 10:07 PM, Dadtom65 said:

Hi guys. Myself also have been thinking of updating my comp ...   <snip> ...   read about the new Cpu but that’s not coming for awhile yet. So not quite sure what to do yet. Derek.

@Dadtom65

I am always thinking about my next upgrade - I think it's a common Flight Sim malady.

 

Hopefully not to confuse you, when I posted many months back, I was planning to upgrade my R1700 (now quite old tech) to an AMD 3900X. The pace of development has been so fast, and my old R1700 (yes @AnkH at just 3.4 GHz) is still doing P3Dv4.5 so well, and with a GTX1080  - not a 2080Ti - that I'm going to wait one more generation and wait for Gen4 Ryzen, and see if there is a 2080Ti Super, or 3080 also in the market.

 

Many people oversimplify (and some mis-understand) that GHz is not the only parameter in a successful P3Dv4+ setup.

This next is empirical data (yes @AnkH I too am a scientist), but with my own eyes, to this date I have yet to see an Intel powered P3D setup that does not have some amount of stutter regardless of resolution. I think the Intel powered guys just get used to it and stop seeing it. My old setup (close to 3 years now) still runs smoothly and without the 'Intel stutter' when I restrict max FR to 30fps synced at 4K. Complete stutter free smoothness, and nowhere near the GHz that many people claim is necessary.

This is because Intel and AMD cpus don't work the same. You can't just compare chips on GHz and cores alone.

 

Now I'm not talking low-res gaming - that's a different thing entirely. I'm referring to P3D4+ at 4K and maximum settings.

I don't muck around with 'tuning' and affinity masks and all that stuff as I like to keep things as simple as possible. My P3Dv4.5 setup is stock without any tweaking except putting up the sliders to max. I have a full 1 Tb NVMe M.2 ssd for P3dv4.5 alone, ORBX for P3d4 using up almost 600 Gb by itself.

 

BTW, I will add, that with the impending release of MSFS2020 later, things will be different again. Asobo (the developers) say that the sim will download massive amounts of data to your computer and then on your machine, it will RENDER the data to produce the graphics. It doesn't matter whether you go Intel or AMD, you will want a lot of cores and threads. Again, today there is no dispute that the latest AMD chips way outperform - at rendering - any equivalent Intel chip. This is not a single thread process.

 

I mention this, as a herring (I won't specify a colour) if you are considering any sort of 'future proofing'.

 

Cheers, A.

[ozboater]

@Doug Sawatzky

Thank you, I have enjoyed seeing your P3D performance with the 2nd and then 3rd generations of AMD cpu's. Particularly, I am heartened by the performance of your R3800X.

Should I break down, and can't wait to do an upgrade, I might just pick up one of these. It will massively outperform my R1700. And as you say, it is a drop-in upgrade (Ive already updated the BIOS in case).

 

As you also stated, PCIe 4 will be important for the future, and the PCIe drives are just beginning to take off with increased performance. We'll see what is announced at CES . All that rendered data for MSFS 2020 will surely benefit greatly from speedy data transfer.

 

Cheers. A.

[AnkH]

4 hours ago, ozboater said:

Many people oversimplify (and some mis-understand) that GHz is not the only parameter in a successful P3Dv4+ setup.

This next is empirical data (yes @AnkH I too am a scientist), but with my own eyes, to this date I have yet to see an Intel powered P3D setup that does not have some amount of stutter regardless of resolution. I think the Intel powered guys just get used to it and stop seeing it. My old setup (close to 3 years now) still runs smoothly and without the 'Intel stutter' when I restrict max FR to 30fps synced at 4K. Complete stutter free smoothness, and nowhere near the GHz that many people claim is necessary.

This is because Intel and AMD cpus don't work the same. You can't just compare chips on GHz and cores alone.

 

Because you mentioned me so many times in your post: the reason I was insisting on "proof" was the simple fact that your screenshots on imgur simply do not proof anything relevant, as the important thing is (you write that yourself...) stutter free experience, something you can simply not show by showing screenshots.

 

Summary of what those screenshots lack/not show:

- settings are not shown completely, some of the not shown settings turned off will provide a lot of "spare" performance, e.g. all shadowing options...

- screenshots where ground is shown in detail (slow and low) are taken in complete clean weather with not a single cloud

- screenshot with 141FPS only shows basic weather (clear skies) and from 24000ft, load on hardware is exteremly little in such a situation

- screenshot with 100FPS and "slow and low" is taken with view direction open water, load is minimized in such a view direction and I guess again with clear skies

- a screenshot can never give any information on the FPS fluctuation (e.g. "smooth simulator"). If I take a screenshot in the right moment, I can basically provide you a screenshot with whatever FPS you want... Enabling the average FPS indication via config would be the least thing to do, that we can see the avg. FPS number on the screenshot...

- screenshots never tell you the full story. Example: when overloading the CPU with too high settings, blurries are usually the case after flying around for some minutes. A screenshot can never recapitulate this...

- your task manager shows that although P3D obviously uses all available cores, your core0 is hammered to almost constantly 100%. Certainly nothing you want to have for longer flights... 

- using the task manager to show CPU clock rate is just plain wrong. It does not show overclocking properly, so nobody actually knows if your Ryzen is running on 3.33GHz or any higher

 

That is all I am saying. I am into simming for a very long time now and honestly, every single claim on how well the simulator runs on "maximal settings" was debunked sooner or later. You do not use maximum settings, plain fact. That is why you are easily maintaining 30FPS on not so heavy situations and more than 100FPS on totally low load situations. Nothing special about this.

 

If you have time: show me a screenshot from your Majestic Dash 8 sitting on a runway of EHAM, view direction terminals, some AI traffic and really bad weather. And as you claim "maximum settings" being used: please show this screenshot with all reflections ticked on and a screenshot from the lighting page with all shadow options turned on, including dynamic lighting. If you then still get your 30FPS locked, we can actually start to "believe" that you get 30FPS locked super smooth in any situation on 4K with max settings. Until then, it is as always: hot air with no proof... Sorry for being this rude, but I went through this kind of discussions so many times before, nobody ever was able to really "proof" their initial faboulous claims...

 

Just to underscore my sceptisim: Rob A., a very well known simmer and an expert in many aspects continuously posts videos on YT showing various aspects in P3D. He has a monster of a rig and guess what? Not even him is claiming that he uses "maximal settings", as he is simply not capable of. Not even with such a monster rig...

[F737MAX]

5 hours ago, AnkH said:

If you have time: show me a screenshot from your Majestic Dash 8 sitting on a runway of EHAM, view direction terminals, some AI traffic and really bad weather. And as you claim "maximum settings" being used: please show this screenshot with all reflections ticked on and a screenshot from the lighting page with all shadow options turned on, including dynamic lighting. If you then still get your 30FPS locked, we can actually start to "believe" that you get 30FPS locked super smooth in any situation on 4K with max settings. Until then, it is as always: hot air with no proof... Sorry for being this rude, but I went through this kind of discussions so many times before, nobody ever was able to really "proof" their initial faboulous claims...

Running FlyTampa EHAM, Orbx TE NL, ASP4 weather, AI limited to 70 aeroplanes and FFTF Dynamic. (Hardware in signature).
I get 19 FPS on the ground. It drops when airborne obviously.

 

Screenshot - https://i.imgur.com/JX3aA0j.jpg

Settings - https://i.imgur.com/ObVkKgk.jpg

With those CPU-intensive settings turned down, in the same scenario, I have 30 FPS with only the occasional drop to 28 FPS.
 

It appears to me from watching and reading other simmers' experiences that Intel's 9900 and 9900K are able to run P3D and XP at higher settings than AMD, however where the Ryzens shine is at running sims smoothly (no infamous Intel stutters) with lower settings and doing it a cheaper price.

 

5 hours ago, AnkH said:

He has a monster of a rig and guess what? Not even him is claiming that he uses "maximal settings", as he is simply not capable of. Not even with such a monster rig...

Completely agree. 'Maximum' settings isn't even possible on a super computer like his.

 

On 1/4/2020 at 8:14 PM, Dusterman said:

And to make matters more complicated, DDR5 motherboard RAM is expected to become available this year. I am a platform agnostic, but if you want to use this RAM it looks like AMD will becoming out with CPUs using it. I have data intensive applications that don't use a large number of cores, so more interested in speed than number of cores.

Intel have been moving towards more cores and away from 5.0+ GHz o/c capability (see the trend of the highest recommended overclock on their CPUs for the last 3 years). More multi-threaded applications are the future.
 

As for DDR5, I don't think that's going to appear on most people's radar until next year. Only people with a professional need for DDR5 or gamers wanting to show off will be buying it this year. To maintain sales, I feel that only the Zen 3 threadripper and whatever Intel offers as their competitor CPU will be DDR5-ready this year.
There will be a lot of AMD users hoping that the Zen 3 is still a AM4 socket CPU.

 

On 1/4/2020 at 11:07 AM, Dadtom65 said:

Hi guys. Myself also have been thinking of updating my comp and did look at the i9 9900K but read some of the remarks of the people that bought said Cpu and they said that they had trouble with the temperature of it. Also looked online and there are a lot of people saying about keeping the dam thing cool. Also Doug read about the new Cpu but that’s not coming for awhile yet. So not quite sure what to do yet. Derek.

If you are not in any rush, I would wait a couple of months to see what AMD is offering later this year. Zen 3 (4000 series Ryzen) CPUs are anticipated this summer, based on the now 3 year-old AM4 socket architecture.
The next Intel range (10th Gen) has been launched (no Core i9 version of them yet though), but real-world TDP is notoriously high on their CPUs. Cooling them sufficiently will likely need a very good AIO water or excellent air cooler to get the best sim experience.

One thing I will say is that for simming, hyperthreading (HT) or simultaneous multithreading (SMT) should be turned off / forced off using a batch process for your sim.
Neither P3D nor XP are designed for using multiple threads on the same core. Just leads to hotter temps and an increased chance of stutters.

 

[ozboater]

15 hours ago, AnkH said:

 

Because you mentioned me so many times in your pos.... < snip>

 

Yeah mate, you were a target, and I was (rudely) poking fun at you - it's an Aussie thing, because you were so darned serious, and I'm on holiday (relaxed) and couldn't resist. (A guilty pleasure is reading the occasional flame war purely for amusement).

I do appologise for any ruffled feathers.

 

Also, I think you miss the point.

 

Context - my Imgur series, nearly a year ago,    https://imgur.com/gallery/LuU1ruC     was in response to another gentleman's response to a newbie's enquiry about what CPU to purchase, where over several posts he insistently claimed that you 'must' have an overclocked, watercooled 8900K or 9900K to run a flightsim, and that only single thread performance matters because there is no multithreading. It shows this is clearly not true.

 

For the record, all of the pics were screenshots with the Task Manager overlaid on the sim window to show real time in-sim cpu usage. To show the settings, I had to paste an image of the settings used, onto the screenshot. Took me quite a while to figure this out, having never bothered doing anything like this before, and I doubt I will again. Only 4K resolution is used.

 

My setup remains an 'out-of-the-box' P3Dv4.5 install with no tweaking apart from increasing sliders. I install ORBX (563 Gb), VFR France and more, many brands and types of of aircraft, and environmental programs (Toposim, REX SF and EF,  Navionics, Flight1 Garmins, and RXP to go with my RealSimGear GNS430 and 530). This completely fills a 1Tb SSD. It is what I run. Many keen simmers will have a variation like this. XP11 is on another SSD, and I will give the new Vulcan update a run sometime to see how it is progressing, but I rarely use XP.

 

My priority is SMOOTH flight without stutters, and I achieve this admirably with the oldest, lowest (and cheapest) Ryzen 8 core in the stack, running nowhere near 5 GHz, and using the included free stock cooler. I make it quite clear that I run my sim locked at 30 fps synced, as on my 4K 65" monitor my old eyes cannot tell the difference with higher frequencies. The high fps shots were, again, in response to a claim that only Intel can produce high fps. Also, clearly not true.

 

And if P3Dv4 hammers the first thread - so what? Any higher performing CPU will give increasing performance, as expected, and there are plenty of more powerful CPUs than an R1700 to choose from. But just an R1700 gives perfectly acceptable (and indeed good) P3Dv4 results and that is my aim to show. The smoothness is most enjoyable, and something I have not yet seen, irl, on an Intel system.

 

Now I have more insight into how the smoothness probably is achieved with Ryzen. I think (and don't know yet for certain) that the other 15 threads are RENDERING the data for the graphical screen image, and this is a process Ryzen is especially adept with. Again, hundreds of YouTubes demonstrate that Ryzen multicore renders way better than any comparable but more expensive Intel CPU. Of course, there are other things that an Intel CPU does better (like low-res gaming that is not my priority). However, if smooth flight is a requirement, then this should be borne in mind. High rendering performance will become increasingly necessary for MSFS 2020 to deliver detail smoothly.

 

Anyone who can shed more light on why Ryzen sims more smoothly than a comparable Intel CPU will be most keenly read by me.

 

Anyway, @AnkH you are not wrong, and equally, I am not wrong. My apologies again, and also my wish for all simmers to enjoy what looks like an exciting sim year ahead in 2020 in both software and hardware.

 

 

[AnkH]

8 hours ago, ozboater said:

where over several posts he insistently claimed that you 'must' have an overclocked, watercooled 8900K or 9900K to run a flightsim, and that only single thread performance matters because there is no multithreading. It shows this is clearly not true.

[...]

The high fps shots were, again, in response to a claim that only Intel can produce high fps. Also, clearly not true.

[...]

My priority is SMOOTH flight without stutters, and I achieve this admirably with the oldest, lowest (and cheapest) Ryzen 8 core in the stack, running nowhere near 5 GHz, and using the included free stock cooler.

[...]

And if P3Dv4 hammers the first thread - so what?

 

I do get your point, no worries. But the reason I was answering that extensively is: while I totally agree that you do not NEED a 5GHz 9900K to properly run P3D, your statement was that you can achieve "maximal settings on 4K" with your old Ryzen and then you underscore your statement above with inconclusive screenshots. Inconclusive due to the fact that they do not show maximal settings and furthermore even rather low demanding scenarios. Or vice versa: no wonder you get locked 30FPS or above 100FPS when using unlimited FPS on scenarios like that with settings like that.

 

This does not render your statements wrong, of course, but just imagine someone reading it as it was written: "oh, this guy can run P3D on 4K with maximal settings on a Rzyen 1700, so I should be totally fine with my Ryzen 3600." Then he turns on his P3D, puts maximal settings and 4K and bummer, sitting on EHAM with bad weather and AI traffic, his FPS are nowhere near 30 but rather around 20 or below and he is totally disappointed and thinks: "oh gosh, I should have get an Intel"... Without knowing that:

 

- A: an Intel would suffer from the same low FPS in such a heavy scenario with max. settings.

- B: you simply did not use maximal settings although claiming it.

- C: 20FPS on EHAM with bad weather and AI traffic is totally fine on max. settings.

 

To me, giving statements about how well (or bad) P3D runs on its own machine are only usefull with the full set of settings. And no exagerations, because they certainly do not help anyone but instead leading to wrong decisions of users that are not that experienced.

 

Means: if your "priority is SMOOTH flight without stutters", fine, that should be anybody's priority, I agree. But then tell the people how you achieve this rather than making ridiculous claims with false information about using max. settings on 4K. Be honest and tell the people that yes, you can achieve such reasonable results even with your "old" Ryzen if you put those settings right. THAT is it what it should be about: use those settings wisely and yes, you do not need a 5GHz 9900K or a Ryzen 3950X.

 

Last: if you think that hammering the first core with P3D does not matter, you should do some reading in other forums. It does matter, especially when using (maybe too) high settings (what you obviously don't do) and as soon as you do longer flights (medium to long haul), as hammering core0 is also something that renders the whole system more susceptible for blurries.

[ozboater]

@AnkH

The pictures certainly are honest, and very clearly show that this is an 8c/16t cpu, and also show cpu use (in real time) and my CPU, Gx card and settings used.

 

Yes, Chris, I agree that some of the wording is not accurate like the images are, and many months ago was written quickly and in heat as a rebuff to repeated strident claims that P3Dv4 does not multithread (it clearly does), and that you need  5 Ghz to run the sim (you do not, and can still get great performance at lower frequencies). If you look at the pics, I believe there can be no confusion.

 

This never was an AMD vs Intel thing, and never intended be that way, but over time, some people have made it so, and in light of the confusion you also suggest (though why a 4 or 6 core user would expect the same, I don't know), I now feel that more accurate text will help. I will attend to that, but not tonight.

 

My colloquial colleages and I habitually refer to max settings as on the Settings\World page, all sliders in the left column are all the way to the right, and I have certainly 'slid' into that habit. Clearly not everyone is that loose. I do like the precision in thought and process of Germans and Swiss. It's precisely why for 30 years I've chosen Stuttgart engineering for my cars.

 

Almost totally, I sim-fly singles and light twins, and mostly in total  ORBX Aust and NZ, but also photoreal Tahiti and Hawaii, and yes, I predominantly get the full 30 fps locked  with whatever weather REX serves up, though I sometimes reset weather to 'Fair'. Also, except for some takeoffs and landings, I enjoy external views. By contrast, one of my friends only uses cockpit views, and only flies tubeliners. We all do it differently. However I was aware that some aiports pose a challenge, and high complexity aircraft adds to this.

 

My thanks to you, because the only time I had flown the Majestic Q400, since purchase, was for those 2 pics. Your response prompted me to get re-acquainted with it, and use it at some of these big airports currently wasted on my drive. It was fun, and in Baltic livery, now I shall try to emulate YouTube's Captain Marina with Drzwieki Riga as a base, like her. (She is a great watch and a serious flier).

 

I do not own EHAM, so I tried Flightbeam EDDN, then FB KMSP, then ORBX ESSA from the Gate to take-off with cockpit and external views. And yes, the fps is much as you suggest. With all World page settings to the right, I got 16 to 20  with occasional drop to 14, but 17-18 predominantly, in the cockpit, and 22-24 with external views at EDDN, a few more at KMSP and a few more still at ESSA (29 externally).

 

Next, sliding all World sliders back just 1 notch, I got 27 in the cockpit, and solid 30 for external views. EDDN was about 2fps lower than KMSP and ESSA, but all external views were 30 and that is the experience I'm satisfied with. Kindly note - no traffic. I am saving my purchased Global Traffic for my next upgrade. No doubt we have other stuff set differently, but this is what I am using and enjoying. It makes me no longer in a rush for my next upgrade, but am still looking forward to doing so - gonna wait for next gen cpu's and gx.

 

[not appropos of anything, at my usual (and more correctly) generally large World settings, I unrestricted the FR in ORBX England TE v2 at the Scilly Isles in a JF Arrow III and got 140 to 170, mostly 160, before eventually dumping it back to my old 30 fps and have left it there. My monitor did not know what to do with 160. Southhampton was a solid 30 in cockpit and external.      FWIW]

Cheers, A.

 

PS, I probably should no longer call myself a scientist, as I am now retired, but boy, my whole career was driven by an overbearing need for accuracy and perfection (medical), and now I bask in freedom to enjoy my newfound looseness.  Truly, I hope you do not have to always carry such a demanding yoke.

 

Oh Yeah - first core stuff.My primary interface when simming is not with my cpu, but with my monitor. Having 15 threads serve up smooth graphics is where I found my priorities satisfied, and I don't worry about the rest. No doubt, should I try trans-atlantic tubeliner flying to a schedule, my priorities could change, as you state.

 

And maybe if many years ago someone had thrown a handful of chlorine in the gene pool, the world might have been a nicer place...

[Alex2378]

Hello ,i’ll like to upgrade my computer ,i’m coming from i7 7700k with asus strix gtx 980 4gb ,32gb ddr4 ram, to 9900k but I’m not sure if is better 9900ks for my P3d 4.5,and off course for the next fs2020 will be ok this cpu or will be better amd solutions?

I’m waiting to upgrade the gpu for the next 3080 solution.

Any help please?

thanks

[recores]

On 5/20/2019 at 9:46 AM, AnkH said:

However, for P3D alone, I would rather go with the 9700K instead (8 cores no HT). The 9900K and its hyperthreaded cores do not help you much in P3D (if at all).

In this sense, you are right. The 9700k is also noticeably cheaper, so the choice is yours. Comparing both processes suggests that it is better to take 9900K, since it has more L3 cache by 4 MB and 2 times more threads. Otherwise, they are identical to each other.

 

https://nanoreview.net/en/cpu-compare/intel-core-i9-9900k-vs-intel-core-i7-9700k

[AnkH]

Time flies by and no surprise, what was said yesterday in regard of hardware is not true nowadays anymore. Personally, I would not suggest to go for a 9700K/9900K build anymore unless you really, really only use the rig for P3Dv4.x. In any other case, my suggestion would go to AMD, the Ryzen 3700X/3800X or even, if some dollars spare to spend, the 3900X. By far the more "future-proof" processors.

[aceridgey]

On 6/10/2019 at 12:16 PM, ozboater said:

@AnkH

You are now quite wrong, and out of touch with P3Dv4 when you state " I repeat, single-core power is still more important than anything else ". I am replying to you as your posts seemed so vehement regarding what is now out of date, and old thinking.

 

P3Dv4 multithreads extremely well, and if you have enough cores and threads, you NO LONGER need high GHz or high IPC.

 

Proof - https://imgur.com/a/LuU1ruC

 

This incredible performance is done at 'just' 3.4 GHz on a cpu of modest IPC. But it runs hard on ALL 16 threads giving fantastic smooth sim performance. And this is 2 year old tech with no Ti graphics.

Personally, I'm looking forward to getting a R3900X with 24 threads and 50% more clock speed and much higher IPC than my 'old dunger' that still performs so well.

 

Thank heavens for Lockheed Martin's programmers and engineers, that now let you use all your hardware's potential, rather than it just snoozing in your case.

 

Cheers

.

I've also noticed for p3d v5, enabling hyperthreading and reducing the clock speed on my 6700k brought much more stability and better performance. 

 

I'm upgrading to a 10700k and have no intention to turn off HT nor clock it yet. 

[Condor]

Sam,

What kind of cooling do you use on your system?

 

Cheers

Condor