2016 Hardware Benchmark - DCS World 1.5.x

I just tested out all the graphics options and setting water and civ traffic from "high" to "medium" raised my fps in critical situations from 30 to 60 fps.

In case somebody else is also struggling with fps at times and doesn't know this already.

I hope some testing with the new gtx 1000 series.

Rx480 as well

Just a quick note of thanks for doing these tests and sharing the results, they've answer a bunch of questions I'd been wondering about.

Cheers

Peter

Well, I'm grateful for the comprehensive tests. This explains the underwhelming performance of my triple monitor, dual 980 Ti, liquid cooled Skylake configuration. Hopefully they figure out a way to punch it up a bit, shame to see my system not reaching its potential.

I think there is something really wrong with the engine/drivers/or whatever manages objets, cars, etc...I did a simple test which consist on set all options to the max except civ. traffic and trees (both at ZERO)...I got very high fps no matter where I go, so problem here are traffic and trees.

Ok, then, I tried rising the trees slider to 5000...again, very high performance everywhere.ç

Then a little bit more trees, went to 12000 trees...nice performance again, then 16000 trees AND...performance went to HELL!!!!!:huh::huh: and here is where things got interesting. As soon as I set trees at 12000+ for some reason, my GPU usage drops to absurd levels. At 12000, I see GPU working some times to 99% but with 12000+ gpu usage is like 40% and thus performance is crap. The same goes for the civ traffic, as soon as I set it at any level above zero, same behaviour.

I have had the chance to test this with both, a gtx 970 and my new gtx 1070, same results on both but with different performance level (1070 gives more fps)

Of course, this shouldnt be a problem with the CPU being maxed out, as my 6700K @4,5Ghz wasnt never maxed along the entire test. So basically, I see my fps dropping to 30/25 some times while my CPU and GPU are working like 50%. I think there is definitely something wrong here (Im talking about 1.5)

I hope some testing with the new gtx 1000 series.

If given the track I could give numbers for a 6700K running at 4.5Ghz and a Gainward GTX 1080...

I think there is something really wrong with the engine/drivers/or whatever manages objets, cars, etc...I did a simple test which consist on set all options to the max except civ. traffic and trees (both at ZERO)...I got very high fps no matter where I go, so problem here are traffic and trees.

 

Ok, then, I tried rising the trees slider to 5000...again, very high performance everywhere.ç

 

Then a little bit more trees, went to 12000 trees...nice performance again, then 16000 trees AND...performance went to HELL!!!!!:huh::huh: and here is where things got interesting. As soon as I set trees at 12000+ for some reason, my GPU usage drops to absurd levels. At 12000, I see GPU working some times to 99% but with 12000+ gpu usage is like 40% and thus performance is crap. The same goes for the civ traffic, as soon as I set it at any level above zero, same behaviour.

 

I have had the chance to test this with both, a gtx 970 and my new gtx 1070, same results on both but with different performance level (1070 gives more fps)

 

Of course, this shouldnt be a problem with the CPU being maxed out, as my 6700K @4,5Ghz wasnt never maxed along the entire test. So basically, I see my fps dropping to 30/25 some times while my CPU and GPU are working like 50%. I think there is definitely something wrong here (Im talking about 1.5)

From what I have read, you are running into a DirectX limiter, no matter what monster GPU you come along with, the logic underneath just BLOCKS iirc

From what I have read, you are running into a DirectX limiter, no matter what monster GPU you come along with, the logic underneath just BLOCKS iirc

There must be something like that...as you said, its like no matter how powerful your system is, there seems to be some kind of "limit" we cant surpass. So even with my 1070 I cant increase trees amount because as soon as you go over 12000, fps fall in. Nevada seems to make it better as gpu usage tends to be as high as it needs but has also some weird problems like getting very close to trees...

I have this configuration : gtx980 oc, i3570k @4.2Gz , 120Hz lcd

In this game i never heard my gpu or cpu fans.

So i've check with HWiNFO64 Program and that strange:

cpu is ok , gpu is ok , memory and so on are ok.

But fps are not so good. (around 60); That with gazelle. (vsync off)

Then if i power on lcd screen on the gazelle fps drop very quickly..but same gpu/cpu and others things are ok.

I have try with/without mods , with low or high settings : same

seems like the bootlegs is not hardware ? or related to how memory works (between gpu/main/virtual) ?

Guys I just upgraded to an AMD fx 8370, 16g ram, gtx 1050 windforce 4g. I am new to nvidia drivers. Do i need to.set anything specific on the nvidia software for dcs?

Sent from my HUAWEI G510-0100 using Tapatalk

I would certainly switch energy management mode from adaptive to max performance - You don't want Your card to down- and up-throttle in the middle of a mission. The rest is subjective fps-vs-eyecandy matter, probably not very different from ATI card settings.

Where is tiborr it's been a long time. How are you buddy?

Reading from this I guess a i5 6600 non k and a gtx 1060 would be good for DCS World 2

Invest in some good fast DDR4 RAM, many tests say it wont matter but I honestly think it matters a lot how fast your RAM can talk to the rest of the system in certain scenarios. You can scale from about 28-29GB/sec to almost 45-50 GB/sec with fastvRAM...which is where X99/2133/2400 boards settle in when not running the same ram, then those pull off towards 75+GB/sec.

Wish my system was up and I could test at 2133 and 3000. They ran till 3600 with a Samsung profile at even lower values...which makes me wonder how far you can push those RAMs without getting into the red arena with VCC-IO and VCC-SA..and DRAM Voltage directly. The Samsung profile wanted 1.60V for DDR4...well...that is too much for the CPU over time.

Damn, how did I miss this thread??!? I've been flying in VR and was hoping that moving to a 1080ti when it comes out would help, but looking at the VR results I doubt it would much.

The one area it might help with is pushing the pixel density / resolution oversampling a bit. Hmm, I will look forward to an update if you try this.

Perhaps in this game not syncing all cores would help with heat when over clocking. 45,44,43,42mhz.

I

Unfortunately, when I do that on mine it makes it goofy for placing the values and I get confused :helpsmilie:..,, so I just place mine at 4.5- 1.2686 volts and call it a day with my i7 4700k. I am going to sell it and get the last updated processor before having to switch MB's. Forgot what it was called but runs at 4.3 ( I think or 4.4) MHz w/o overclocking.

To be honest, it is very disheartening, to hear this info about this game/sim.

I would purchase double the price or triple if it would utilize crossfire and the cores more appropriate! Screw " free" game. Give me greatness and I will pay for it gladly.

........

 

..... Give me greatness and I will pay for it gladly.

Some things even money can't buy !

MasterCard should maybe include DCS Pilots in their commercials...LoL

I would certainly switch energy management mode from adaptive to max performance - You don't want Your card to down- and up-throttle in the middle of a mission. The rest is subjective fps-vs-eyecandy matter, probably not very different from ATI card settings.

Yes that helps, but the new cards still down throttle in DCS when it goes under 30% GPU utilization, it's pretty sad. Wish there was an app out there to dynamically load GPU an extra few % only when needed to prevent going under 30% :) (if not proper vbios etc.)

looking forward to see Ryzen processors included in the comparison, I'm on the fence and would like to know if thats a good option for DCS or should I wait for skylake-X 6 cores (I will use multithreading as well and need a good blend of single core and multithreading)

Actually, when I look at IPC alone, tbh, it is not the sole story and not the only thing that makes a difference.

COming from my 4.8 Sandy with 2800 IPC in Passmark and having 2900 now makes we wonder if those 100 points really are responsible for that HUGE step forward in DCS this new CPU brought for me.

There is a good chance that Ryzen will just perform quite well in DCS just because it is a more modern CPU than what we used to have. The benefits under the hood are somewhat hard to quantify, but they are easy to qualify when I grab the stick and fly as smooth as never before.

It just cannot be that 100 score + in Passmark. New CPU's have many improvements over older generations and that seems to pay off big time, at least with my rig and what I had before.

I wish I could grab a Ryzen and lock myself into a room for 2 weeks with Pizza Cola & Coffee :)

Like in old days without kids & wify LOL

In other words, I bet I could downclock my 7700k to the speed of my old sandy in IPC and still surpass it in DCS ! I am 100% sure that is what woulod happen....gonna try out tonight when I get to my machine,,,just to see if it is really like that.

2800 vs 2800 old vs new not a bad thing to do

The only difference will be newer RAM ( cant avoid that part ) and newer architecture

While people claim that max and avg fps is not that great with Ryzen, the minimum fps seems to be much better. Also, despite general complains on gaming performance, it goes across the board that people find gameplay with Ryzen much smoother. I think for things like DCS smooth gameplay is very important, hence I'm looking forwards to the results.

Hopefully someone pulls the trigger on that one and shows how is it performing so far here.

Bitmaster, you mean clock, not IPC. IPC = instructions per cycle. Nothing is getting anywhere near that much IPC. According to my calculation based off Wikipedia (*cringe*) data, the 1800x has 5.2 IPC per thread or 10.4 IPC per core. I could find no data on the 7700k but the 4770k gets 8.57 IPC per physical core. Regardless, total processing power is IPC*clock speed (over simplified, but gives a rough idea).

Nitpicking aside, I thought Hiromachi bought a 1700x. Or is our Hiromachi different from the one on the BoS forums? In regards to minimums, I noticed that too in a lot of the posted benchmarks. I agree that they are more important. That said, DCS loves its single thread performance and I don't think any of the R7's are made to complete with the I7-7700k in that regard. That's why I'm mainly looking at the R5 1600x. It might boast better IPC with that same 4GHz clock.

I preordered 1700X indeed. But since my country is always at the end of shipping chain I wont have CPU until next week and who knows when I will get motherboard. Hence why I look forward to what other people can report.