Will-my-pc-run-Arma3? What cpu/gpu to get? What settings? What system specifications?

[Reezo 45]

Thanks for the additional info, I just ran my good old Valley (I have a weak spot for it) to see what difference I could have with the two setups and it shows that:

 

2 sticks @ 3600 (SOC 1.1v DRAM 1.40v)

= Score 4020, FPS 96.1 Average (Max 229.6)

 

3 sticks @ 2800 (SOC Auto DRAM Auto Voltages)

= Score 3623, FPS 86.6 Average (Max 206.6)

 

Both tests with the same XMP timings..

 

Cinebench r15 shows 1840 vs 1868 with 3 sticks, but I think they're practically negligible..

 

Obviously 4 sticks would be good to know, but since I doubt RAM will maxed out, I am still thinking higher speeds and/or tighter timings = better performance.

 

@Groove_C you were saying that 4x8 at 2800 will be better for ArmA 3 instead of 2x8 at 3600? Maybe I should then tighten the timings for 2800 to be fair? This goes with what @oldbear wrote than RAM speed itself gave ArmA 3 only 1/3 fps more but the fact that it has twice as  much memory could help clogging in longer missions etc. after hours of playing? Sorry if I did not get it right..

 

[Groove_C 267]

15 minutes ago, oldbear said:

Reezo : the gap in performances benching Arma3 between Corsair LPX 3200 C16 and G.Skill TridentZ NEO 3600 C16 is 1/3 FPS in YAAB, and invisible in game.

Sure, but the goal is to buy cheaper RAM (with Samsung B-Die chips) and to CO it to something like 3600 14-15-14 or 14-15-15 or same timings, but 3800 MHZ, which works.

The FPS difference is then more significant.

It's not really the goal to compare XMP of one frequency vs. XMP of another frequency. XMP is just junk, that is used by average users, that don't have a clue about RAM and should be avoided. Manually set timings, primary, secondary and tertiary is best way to do, for a specific combo of mobo and CPU.

 

3 minutes ago, Reezo said:

Thanks for the additional info, I just ran my good old Valley (I have a weak spot for it) to see what difference I could have with the two setups and it shows that:

 

2 sticks @ 3600 (SOC 1.1v DRAM 1.40v)

= Score 4020, FPS 96.1 Average (Max 229.6)

 

3 sticks @ 2800 (SOC Auto DRAM Auto Voltages)

= Score 3623, FPS 86.6 Average (Max 206.6)

 

Both tests with the same XMP timings..

 

Obviously 4 sticks would be good to know, but since I doubt RAM will maxed out, I am still thinking higher speeds and/or tighter timings = better performance.

 

@Groove_C you were saying that 4x8 at 2800 will be better for ArmA 3 instead of 2x8 at 3600? Maybe I should then tighten the timings for 2800 to be fair?

 

 

Score is lower with 3 sticks, because your RAM is not running in dual channel, but in single only, which happens if you have 1 or 3 sticks instead of 2 or 4 sticks.

Never said anything about 2800 MHz.

Are you sure, that 1 sticks is defective?

Have you checked it separately with only it only in one of the slots?

[Reezo 45]

Score is lower with 3 sticks, because your RAM is not running in dual channel, but in single only, which happens if you have 1 or 3 sticks instead of 2 or 4 sticks.

 

I noticed CPU-Z saying "Dual" with 3 sticks but it was probably wrong anyway.. it seemed weird and I thought maybe one pair was dual and one single?

 

Never said anything about 2800 MHz.

I read around that it will be impossible to have 3600 with 4 sticks on a 2700x? I might have just misinterpreted it..

 

Are you sure, that 1 sticks is defective? Have you checked it separately with only it only in one of the slots?

I'll do more tests, there might be a slot it likes.. it seems that nothing worked so far..

 

 

In other news I think 3600 1.40V/1.1V is still not very stable with the XMP, I'll try the manual tuning of tertiaries, primaries and secondaries and then see..

 

[Groove_C 267]

Just go for 3433 MHz with tight timings. That's it.

If one of the sticks is defective, then it's not necessarily the CPU/mobo combo that can't handle 4 sticks.

4 sticks should be doable, even if not @ 3600 MHz.

But if one is defective, then you have to use only to sticks. Never use 1 or 3 sticks.

[Reezo 45]

Thanks @Groove_C I am about to start testing.

 

I wanted to show you guys the place I'll be streaming from! I am very happy about it, I gotta say!

 

 

 

 

[Reezo 45]

Thanks again @Groove_C

 

Bad news and good news. The bad news is simply that one stick is defective. A replacement is on the way. It just does not work on any slot, makes the motherboard halt even before posting.

 

The good news is good, though. I could run a 3600 Fast Preset, exactly as recommended by DRAM Calculator, on my first try. I just filled the parameters and off I go. Here is what I am currently sitting at.

 

Spoiler

 

I did bench against Valley once again and got:

 

SCORE: 3641

FPS Average: 87.0

 

against the XMP profile (still at 3600):

 

SCORE: 4020

FPS Average: 96.1

 

So it seems the XMP wins against DRAM Calculator at fast? Woops! Interesting.

[Groove_C 267]

DRAM Calculator is not reliable.

As I told you before, choose XMP and try 3600 15-15-15-35 1.35 or if not 1.4V. Or even 3600 14-15-15-35 or 14-15-14-34 or 16-16-16-36.

[Reezo 45]

1 hour ago, Groove_C said:

DRAM Calculator is not reliable.

As I told you before, choose XMP and try 3600 15-15-15-35 1.35 or if not 1.4V. Or even 3600 14-15-15-35 or 14-15-14-34 or 16-16-16-36.

 

I changed RttPark to /5 (as reported in Alt.1) and it's now stable and matching the numbers in performance that I had on the XMP profile (still the XMP was faster of maybe some decimals over a 1000th, nothing too important).

 

I'll run these and take it from there!

The best guide I've ever used in the past is this one: https://github.com/integralfx/MemTestHelper/blob/master/DDR4 OC Guide.md

 

I'll save a profile and try now..

 

UPDATE:

 

After manual tightening, here is my current settings (blue) vs what DRAM Calculator said to be ok in FAST mode..

I am benching higher in Cinebench but lower than before in Valley. Funny 🙂 It almost seems XMP with all auto and 2 sticks was king 😉

Spoiler

 

I am benching higher than before, this seems to be close to the sweet spot!

 

[Groove_C 267]

Well, this is not much. Why don't you try tighter timings?

[Reezo 45]

 

4 minutes ago, Groove_C said:

Well, this is not much. Why don't you try tighter timings?

 

Anything tighter in the first five will not post.. even at 1.25v and 1.46v

[Groove_C 267]

Ok, then this is the CPU.

Try 3433 with tighter timings.

[Reezo 45]

More testing today, new sticks just came in..

https://imgur.com/a/dcR49de

 

My biggest thanks to the people from the floor above ours that made this possible! Berlin is a great city.

 

Fun fact: the RAMs booted up instantly with the previous preset, which was 3600 16-17-17-17-34. Not that it means anything but they woke up and like "woot? ya I can do dis". boom.

[Reezo 45]

New testing.. I got here at the moment... @Groove_C do you think I should push more? 🙂

Spoiler

I was able to fit all the red numbers (suggested) and the RAM booted instantly and it works. Benchmarks are kinda the same but so far I cannot beat the Valley score done yesterday with the Ripjaws V.. but I haven't done as much testing.

 

UPDATE:

 

Wow I got here:

 

 

 

That's pretty insane and I kind of a personal record. This + the RGB tells me I can return the Ripjaws V.. The TridentZ RGB 3200 CL14 seem to be more stable and get up to superior timings. I am not sure I should bench the two because 3733MHz and CL14 is kind of extreme, while I was at 3600 on CL16 before.. AND WITHOUT RGB. I mean.. what are we ta'lkin' abbout, right?

 

UPDATE 2:

 

Boy has my New Year's Eve gotten more insane.. I put the Ripjaws V back in and loaded a BIOS profile where I had the bench before.. and I am getting at least 100 FPS-Average compared to the TridentZ that go to 89. something, in Valley.

 

Now I am not sure how I should proceed, I might go over Reddit and ask for more opinions.. this is really cool and interesting.. basically the Ripjaws V at 3600 CL16 are killing the TridentZ at 3733 CL14. Both are set on the FAST preset of DRAM Calculator.. Cinebench r15 scores are similar, but obviously I am interested in FPS.. we are talking over 11 fps more average for the Ripjaws V. Wow.

 

UPDATE 3:

 

Further adventures led me to this, which holds the same score as above but it's more stable. Cool stuff. Notice how my settings in blue are even tighter than the recommended FAST Preset in red.

 

 

 

 

 

UPDATE 4:

More testing today.. I got here (see the numbers in BLUE). I am at 3600 from an initial of 3200 (these sticks are 3200 CL14). It seems pretty solid but it's CL16..but it's also 3600 lol.. I played some Borderlands 3 and nothing went wrong. So at least it's stable!

 

Moar testing!

 

 

[SpartanOfficial 0]

Hi I have recently been experiencing EXTREME frame lag on multiplayer and single player. Changing graphics doesnt affect anything and I used to be able to get over 80 fps consistently. My computer hasnt had this same effect in other games either so Im not sure whats going on. I have tried moving arma to SSD, clean install, changing graphics. 

 

Here are my specs: https://imgur.com/a/LPisONJ

[oldbear 390]

Welcome 😎

 

Arma3 is based on the  "aged 10 years" Real Virtuality Engine design, so you can't expect 2020 AAA game performances.

Bohemia Devs have done an incredible job to bring it at the level it's still working now doing thing it hasn't been built for.

Arma* has turned multicore with Arma2: Operation Arrowhead but is not much multithreaded.

RV Engine does not allows any kind of parallelism in the processing, all is strictly dependent on the simulation running on a limited number of cores on the processor.

 

So, in order to play Arma3 on the high end level, you need first a very fast CPU, the faster the better, a Intel Core i9-9900K @ 5.0 GHz on all cores [or a Ryzen 9 3900X on PBO], so your Intel i9-9920X is not the best suited but must do the job at a good level especially if the cooling solution used allows it to access the Turbo Boost Max 3.0 @ 4.5 GHz. Nevertheless, it should be right to check if the performances are not enhanced by disabling the multi threading, 12 cores seems more than enough.

A GPU allowing to play without limitations, probably a GTX 2070S or a RX 5700XT in order to display the simulation in 1440p, here the GTX 2080Ti is an overkill, Quality can be set at "Ultra" level.

A good  and fast SSD hosting Windows 10 64bits and Arma3 giving high level read performances in order to allow the textures streaming, best option to get rid of stuttering and textures flickering.

A 16 GB (2x8 GB) kit of fast DDR4 @ 3600 MHz C16.

 

So, first, tell us tell us which disk is hosting Arma3.

Second try to tweak AA&PP section such as :

 

Note : these are the settings of my #1 gaming rig [i7-7700K-RTX 2060], it's not the rule to apply, just a track for different settings. To turn off what makes the game blurry is part of my personal fight as a myopic guy against the lack of sharpness, but the rotation blur setting is purely a gameplay setting to keep the feeling of the character in game just as keeping a minimum  head bob.

[Groove_C 267]

What we have so fare in YAAB standard.

 

R7 3700X 4.2 GHz | 16 GB 3600 MHz 14-15-15-35 1.35 V | GTX 1080 | SATA SSD

min. 42.5 avg. 64.2

i9-9900K 5.0/4.3 GHz core/cache | 32 GB 3600 MHz 15-15-15-35 1.4 V | GTX 1080 | SATA SSD
min. 42 avg. 64.2

i9-9900K 4.8/4.3 GHz core/cache | 32 GB 3200 MHz 16-18-18-36 1.35V | GTX 1080 Ti | SATA SSD
min. 35.3 avg. 53.7

 

Windows and BIOS up to date + no unnecessary known background processes/programs running.

 

The guy, which has the first place with his 9900K 5.3 GHz + 4200 MHz CL16 + RTX 2080 Ti has somehow faked his results, 88.7 FPS avg., since past 4.7-4.8 GHz the frequency doesn't bring anything anymore (Intel), his RTX 2080 Ti has nothing to do @ 1080p standard and his 4200 MHz CL16 is not even 1 FPS better than 3600 MHz CL14.

 

The best setting is 3800 MHz CL14/15.

 

 

[Groove_C 267]

YAAB standard

i9-9900K 5.0/4.3 GHz core/cache | 32 GB 3600 MHz 15-15-15-35 1.4 V | GTX 1080 | SATA SSD

min. 42, avg. 64.2, max 98

CMA x64 AVX2 memory allocator:

min. 44, avg. 68.7 max 101.5

vs. i7-5775C 4.2/3.8 GHz core/cache 128 MB L4 cache | 32 GB 2200 9-9-11-18 1.65 V | GTX 1070 Ti OC | SATA SSD (also with CMA x64 AVX2 memory allocator)

min. 42, avg. 69, max. 103

[Groove_C 267]

@Reezo

It's better to stream using your GPU, its unused by Arma overpower, with Shadowplay.

 

Because OBS is very heavy on the CPU, especially on your by now "old" 2700X and can't do 60 FPS without FPS drops/skipping at preset slow with constant and 10K bitrate, unlike 3900X.

 

See video here, frames skipped/dropped in % with OBS 1080p 60 FPS lock, constant bitrate, preset slow and 10K bitrate and see FPS drops in % in the video 3900X vs. 3700X vs. 9900K 3600 MHz 16-16-16-36 (all CPUs stock).

No need to understand russian. Just watch.

 

[Reezo 45]

Thanks! But does OBS use hardware when using the NVIDIA ENC (new)? I am not using x264 which is indeed on the software. I believe Shadowplay should be using the same encoder as well and going off the GPU..

 

Also I am not sure I will stream at 60p as ArmA will never reach those FPS reliably enough for the stream to be 1:1 - my guess.

 

Heck I was even thinking to stream at 24p movie-style..

 

It's important to note that, past that, I am not willing to give out an organ for one single software title like ArmA 3, no matter my love for it. If a game cannot be run on a system that was born SIX years after its release date (and with the hectic pace at which the computers world fares), the issue is not with the build. It's clearly somewhere else. I think the same degree of maturity that is a virtue among some gamers of this niche genre (milsim) should be put to use once more when approaching the choices involved in investing in a system to run the genre itself.

 

UPDATE: The 3700x just got here! Thanks to my most amazing colleague Gregg for lending it for the cause! You will be allowed to play with me, mate.. maaaaybe (shush: he lurks these forums).

 

[SpartanOfficial 0]

On 1/1/2020 at 8:06 AM, oldbear said:

Welcome 😎

 

Arma3 is based on the  "aged 10 years" Real Virtuality Engine design, so you can't expect 2020 AAA game performances.

Bohemia Devs have done an incredible job to bring it at the level it's still working now doing thing it hasn't been built for.

Arma* has turned multicore with Arma2: Operation Arrowhead but is not much multithreaded.

RV Engine does not allows any kind of parallelism in the processing, all is strictly dependent on the simulation running on a limited number of cores on the processor.

 

So, in order to play Arma3 on the high end level, you need first a very fast CPU, the faster the better, a Intel Core i9-9900K @ 5.0 GHz on all cores [or a Ryzen 9 3900X on PBO], so your Intel i9-9920X is not the best suited but must do the job at a good level especially if the cooling solution used allows it to access the Turbo Boost Max 3.0 @ 4.5 GHz. Nevertheless, it should be right to check if the performances are not enhanced by disabling the multi threading, 12 cores seems more than enough.

A GPU allowing to play without limitations, probably a GTX 2070S or a RX 5700XT in order to display the simulation in 1440p, here the GTX 2080Ti is an overkill, Quality can be set at "Ultra" level.

A good  and fast SSD hosting Windows 10 64bits and Arma3 giving high level read performances in order to allow the textures streaming, best option to get rid of stuttering and textures flickering.

A 16 GB (2x8 GB) kit of fast DDR4 @ 3600 MHz C16.

 

So, first, tell us tell us which disk is hosting Arma3.

Second try to tweak AA&PP section such as :

 

Note : these are the settings of my #1 gaming rig [i7-7700K-RTX 2060], it's not the rule to apply, just a track for different settings. To turn off what makes the game blurry is part of my personal fight as a myopic guy against the lack of sharpness, but the rotation blur setting is purely a gameplay setting to keep the feeling of the character in game just as keeping a minimum  head bob.

I understand that I cant expect AAA speeds but it used to be a lot higher than this. Its performance is worse than my old laptop! I have arma on my games disk which is the d: Drive. The drive is 1TB SEAGATE BARRACUDA 2.5" SSD, (upto 560MB/sR | 540MB/sW). Given all my games are all on that disk and every other game seems to run smoothly I doubt it is the drive. 

I have tweaked AA&PP accordingly and still experience the same issue.

[oldbear 390]

@ Spartanofficial could you please run YAAB bench mission  3 times using Standard settings and post here a screenshot of the results ?

You can get YAAB [Yet Another Arma Benchmark] on the Steam Workshop : https://steamcommunity.com/sharedfiles/filedetails/?id=375092418

How to :

- Disable all the mods and addons you are using.

- Set your rig to "Standard" settings, then start the mission and press "S" key .

- Let the mission run to the end, save the result in the results UI.

- Start it again ...

[SpartanOfficial 0]

On 1/7/2020 at 6:16 AM, oldbear said:

@ Spartanofficial could you please run YAAB bench mission  3 times using Standard settings and post here a screenshot of the results ?

You can get YAAB [Yet Another Arma Benchmark] on the Steam Workshop : https://steamcommunity.com/sharedfiles/filedetails/?id=375092418

How to :

- Disable all the mods and addons you are using.

- Set your rig to "Standard" settings, then start the mission and press "S" key .

- Let the mission run to the end, save the result in the results UI.

- Start it again ...

https://imgur.com/R9fnjVF Sorry for delay have been busy! Appreciate the help!

Edited January 11, 2020 by SpartanOfficial

[oldbear 390]

Well ... something went wrong.

We can try to understand what is wrong but from a distance it is almost impossible in particular if it is a hardware problem.

I am telling you something is not right, because, here you can have a look at what I get when I had done some YAAB runs with a R5 2600/RX570/16GB_3200_C16 rig on ASRock AB350ITX Mobo + IIyama 1080p.

 

 

It will be necessary to make a complete examination both from the hardware and software point of view.

You will have to check the system, all the drivers and possibly the presence of malware, viruses and trojans.

 

[SpartanOfficial 0]

13 hours ago, oldbear said:

Well ... something went wrong.

We can try to understand what is wrong but from a distance it is almost impossible in particular if it is a hardware problem.

I am telling you something is not right, because, here you can have a look at what I get when I had done some YAAB runs with a R5 2600/RX570/16GB_3200_C16 rig on ASRock AB350ITX Mobo + IIyama 1080p.

 

 

It will be necessary to make a complete examination both from the hardware and software point of view.

You will have to check the system, all the drivers and possibly the presence of malware, viruses and trojans.

 

The system still performs fine in all the other games I've tested it on. I don't know what to do, I've scanned for viruses with my antivirus, CCleaned everything several times, clean installed several times, cleaned my computer of dust (after all this happened), validated steam files, launched as an admin, changed graphics, updated graphics drivers fully, optimized the game using nvidia experience, I would defrag but it seems pointless since I tried it on two different SSDs and same issue also it would damage them. Should I contact arma devs - if so how? Should I be contacting my computer manufacturer since it is under warranty and is 6months old? How do I properly examine both software and the hardware to find any answers? 

 

Sidenote: I get 200+ fps in the menus but I guess that's just a still image usually. 

[oldbear 390]

Rhoo !

Ok, understood!

You said  "optimized the game using nvidia experience" and that is wrong. Nvidia experience tend to apply to your rig the parameters used by other players.

It's not based on tests done by Nvidia, it's based on statistics, for short, if the majority of players set up their rig helter skelter, the proposals from GeForce experience will be non sense.

 

Could you post here screenshots showing the 3 screens from the Arma3 Video Tab.

 

Here, my own parameters, loosely based on the Old Bear method ™, I am currently using on my #1game rig [i7-7700K/RTX 2060 6Go/16 GB DDR4 3200/SSD M.2 500GB + SSD M.2 500GB/Z270] + Iiyama 144Hz 1080p, you had already seen my AA&PP video parameters.

 

 

- never use "SAMPLING" over 110 % even if you want to count all the bolts from the enemy tank hull before firing,

- never use "SHADOW" : Low settings if you get a real GPU

- setting "VISIBILITY" > "OVERALL" to 12000m is "insane" as a well known Veteran Arma Dev officially declared a long time ago. Please use 3800m as a maximum if you are playing as a grunt.

 

 

- disable VSYNC : with a high end rig as a test.

- playing on a 4k monitor may be a highly disappointing experience because of the way the game engine works.