Jump to content
Placebo

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

Recommended Posts

@verstand3n

You're using an outdated YAAB version. The latest (with hotfixes) is 1.00.

Your 3200 MHz CL14 has a latency of 8.75 nanosecondes whereas mine (2400 MHz CL10) has a latency of 8.33 nanosecondes (+ very very agressive secondary and tertiary timings).

Your CPU runs at only 4.4/4.1 GHz whereas mine runs at 4.8/4.4 GHz.

Nothing strange.

 

Midrange DDR4 has more bandwidth than high end DDR3 due to higher frequency, but has worse latencies.

Games benefit rather from lower latencies than higher banwidth (to a certain point).

 

So my DDR3 2400 MHz CL10 corresponds to DDR4 3600 MHz CL15 (in terms of latencies).

Share this post


Link to post
Share on other sites

I am using the one from the Workshop and it actually says v1.0 in the result screen but when I mouse over it states what I copied into my results. Is that maybe the correct one after all?

 

I did not expect the Core Frequency and timings to make such a big difference in conjunction with the admittedly small increase in IPC during generations but after all I am happy with the results and gameplay experience coming from a Xeon 1230v3 with <30 fps in most situations.

 

Should I stop feeling lazy about optimization I will try to improve Ram timings and cache frequency.

 

 

Share this post


Link to post
Share on other sites

@verstand3n

If it's the one directly from the Workshop, then it's 1.00, even if it says 0.99.

 

ArmA is all about frequency and latencies, not as much about cores.

 

i7-6700K is a really good CPU.

No problems to reach stable 4.8 GHz on air (once delidded) and paired to some 3600 MHz CL15 RAM or higher.

  • Thanks 1

Share this post


Link to post
Share on other sites
7 hours ago, Tankbuster said:

No, what does that do?

 

Sets everything to 'standard', that way all players get a benchmark they can compare.

  • Like 1

Share this post


Link to post
Share on other sites

For future YAAB results submissions/comparison, please use this topic.

Share this post


Link to post
Share on other sites

i7-5775C @ 4.2/3.8/2.2 GHz core/cache/eDRAM | 8 GB DDR3 2400 MHz 9-11-11-31-T1 | GTX 1080 Ti

On one hand, FPS is higher (slightly), because of GTX 1080 Ti vs. RTX 2060 / GTX 1080 / GTX 1070 Ti / RX Vega 64.

On the other hand, FPS would have been even higher (slightly), if files' swapping (RAM -> paging file) didn't happen, because of only 8 GB of RAM.

And frame time must have suffered (slightly) as well, because of only 8 GB of RAM.

 

1080p ultra:

+5 FPS vs. i5-8600K @ 4.8/4.5 GHz core/cache | 16 GB DDR4 3200 MHz 15-17-17-35-540-2T | RTX 2070

+4 FPS vs. i7-4790K @ 4.8/4.4 GHz core/cache | 32 GB DDR3 2400 MHz 10-11-12-18-204-1T | GTX 1070 Ti OC

-2 FPS vs. i7-7700K @ 5.0/4.4 GHz core/cache | 16 GB DDR4 4200 MHz 17-17-17-36-260-1T

-3 FPS vs. i9-9900K @ 4.7/4.3 GHz core/cache | 16 GB DDR4 3600 MHz 16-16-16-36-631-2T | RTX 2080 Ti

-7 FPS vs. i7-8700K @ 5.0/X.X GHz core/cache | 16 GB DDR4 4200 MHz 16-XX-XX-XX-XXX-XT

 

1440p ultra:

+6 FPS vs. i5-8600K @ 4.8/4.5 GHz core/cache | 16 GB DDR4 3200 MHz 15-17-17-35-540-2T | RTX 2070

+6 FPS vs. i7-4790K @ 4.8/4.4 GHz core/cache | 32 GB DDR3 2400 MHz 10-11-12-18-204-1T | GTX 1070 Ti OC

-0 FPS vs. i9-9900K @ 4.7/4.3 GHz core/cache | 16 GB DDR4 3600 MHz 16-16-16-36-631-2T | RTX 2080 Ti

Share this post


Link to post
Share on other sites

Will 100% buy it on AliExpress for 350$ + shipping for 3 additinal years of ArmA on same platform.

Because @ 4.4/4.0 GHz core/cache and with my 32 GB DDR3 2400 MHz CL10, I will get almost 10 FPS more, which will be fine (in ArmA) till 2022.

Vs. buying a very expensive CPU (i9-9900K 5.2 GHz) + very expensive RAM (32 GB 4000 MHz CL17) + expensive mobo + custom watercooling.

For less than 10 FPS, average, singleplayer.

 

Q2 2020 still 14 nano, but 10 cores Comet Lake (confirmed)

Q2 2021 still 14 nano, but 10 cores (refresh) Rocket Lake (confirmed)

Share this post


Link to post
Share on other sites

At the other end of the spectrum, still looking for the minimum hardware actually required to play Arma3, I continue my tests with an AMD R3 2200G.

I am getting interesting results while upgrading the rig up to 16 GB @ 3000 MHz and using the ASUS EZ mode [OC for the noobs] ~ 28 FPS range

I have open a topic Playing Arma3 on an AMD R3 2200G APU - Rogue One

 

 

 

  • Like 1

Share this post


Link to post
Share on other sites

With an iGPU, RAM amount, frequency and timings are even more critical for FPS and frame time.

So having only 8 GB of RAM and then share these poor 8 GB with the iGPU is an additional performance to the hit of having only 8 GB.

So with an iGPU, having more RAM, better frequency and timings is even more critical.

Share this post


Link to post
Share on other sites

I've been keeping an eye on this laptop with the intention to buy in the near future:

 

https://www.amazon.co.uk/dp/B07HH5XR8K/?coliid=I49C4WSH9H7HZ&amp;colid=20PHMYNGG1ACA&amp;psc=1&amp;ref_=lv_ov_lig_dp_it

 

However, for the same price, the below is about to come on to the market with the new mobile 1660 and AMD Ryzen. It also has a 512GB SSD minus the 1TB HD and 8GB RAM v 16GB in the above model.

 

https://www.amazon.co.uk/dp/B07QKHHRLK/?coliid=I3VK2BI0LVC32O&amp;colid=20PHMYNGG1ACA&amp;psc=0&amp;ref_=lv_ov_lig_dp_it

 

My question is, on balance, what is the better option? To go with a 256GB SSD, 1TB of storage, i7-8750H, 16GB RAM and a 1060 6GB. Or to go for a 512GB SSD, with a newer GPU and Ryzen?

Share this post


Link to post
Share on other sites

@-Snafu-

8 GB RAM have no future.

256 GB SSD is to low.

GTX 1660 Ti has performance level of GTX 1070.

 

Have checked Enfusion (DayZ) today and the load across threads is pretty equal. Framerate very high and stable and the gameplay feels/looks very smooth.

Have very high hopes for A4 (updated Enfusion).

dayz_x642019-04-2716-eaj3e.jpg

And this is even without DX12, where it's possible to spread load as well across threads. Hope Bohemia will fully implement DX12 (with all its advantages) and not only its visual part!

  • Like 1

Share this post


Link to post
Share on other sites
1 hour ago, Groove_C said:

Have checked Enfusion (DayZ) today and the load across threads is pretty equal. Framerate very high and stable and the gameplay feels/looks very smooth.

Have very high hopes for A4 (updated Enfusion).

 

I would think that DayZ has much less to think about than an Arma title, therefore would naturally do better with performance.

AI, bullet calculation, to mention just a few things that are not used in DayZ to the same degree.

Share this post


Link to post
Share on other sites

@-Snafu- Here and now we don't know how the incoming AMD Ryzen R7-3750H will perform in Arma 3, so it' not possible to give you a full answer.

- RAM, I agree with Groove_C about the fact that, now, you must look for 16 GB. I will add, that you will need the fastest RAM you can use on the laptop

- SSD, 256 GB is the minimum, 500 GB is better, as the SSD price are still going down fast, you must keep an eye on new bargains

 

  • Like 1

Share this post


Link to post
Share on other sites

Thanks, I'll hold off for a while and keep an eye out for deals.

Share this post


Link to post
Share on other sites

YAAB results comparison updated with an i9-9900K.

 

"Extreme" OC on an Asus Maximus XI Apex (Z390).

2x420 mm 30 mm thick + 1x420 mm 60 mm thick radiators with 9 140 mm fans.

i9-9900K 5.2/4.8 GHz core/cache

RTX 2080 Ti 2145/7800 MHz CUDA/vRAM (KFA2 Hall of Fame)

RAM 2X16 GB DDR4 4200 MHz 16-17-17-32-320-2T

  • Like 2

Share this post


Link to post
Share on other sites
On 5/5/2019 at 2:06 AM, Groove_C said:

YAAB results comparison updated with an i9-9900K.

 

"Extreme" OC on an Asus Maximus XI Apex (Z390).

2x420 mm 30 mm thick + 1x420 mm 60 mm thick radiators with 9 140 mm fans.

i9-9900K 5.2/4.8 GHz core/cache

RTX 2080 Ti 2145/7800 MHz CUDA/vRAM (KFA2 Hall of Fame)

RAM 2X16 GB DDR4 4200 MHz 16-17-17-32-320-2T

Looks like wrong results. It is 22 FPS (32%) more compared to other 9900K result, but overclocked only 10% on CPU and 16% on memory higher. That is impossible result. May I ask for video proof with all settings set on video plus YAAB run?

Share this post


Link to post
Share on other sites

It's not just 10% CPU OC and 16% RAM OC. It isn't as simple as you think.

There is also 11% CPU cache OC, which adds slightly more FPS, significant GPU cores and vRAM OC, which adds slightly more FPS as well and 16% RAM frequency OC is not the most important thing to look at, because higher RAM frequency provides only higher bandwidth in GB/s. RAM bandwidth doesn't help as much FPS-wise as primary timings do. But most significant improvement is brought by secondary and tertiary timings, which aren't even shown here.

It's a as low as possible cache and RAM latency that really boosts FPS in ArmA, not just raw bandwidth (RAM frequency).

 

By just optimizing evry single timing of my DDR3 RAM I gained 7 FPS @ 1080p standard (vs. XMP) at same RAM frequency. So don't blindly look at RAM frequency/bandwidth only.

 

This setup is not mine, but its owner has proposed to me to record a video to prove his YAAB result.

Share this post


Link to post
Share on other sites

I think, that tester misunderstood the settings. Probably, he tested on LOW and STANDARD instead of STANDARD and ULTRA. Then 3 - 4 FPS boost is appropriate compared to old results (69FPS vs 66FPS).

Share this post


Link to post
Share on other sites

I will see once I have the video from him as proposed by him.

Share this post


Link to post
Share on other sites

Bah, posting videos that agree with your own, minority viewpoint. When youtube is alive with similar videos saying AIOa are dead, then I'll take notice. When youTube is alive saying the 5775 is the way forward for CPUs, I'll sell mine, but we both know that these aren't going to happen.

AIOs are very reliable and work cooler and quieter than air-fan coolers. The technology is mature and proven. But they are expensive and we know that you put the cost criteria higher than most of us. So you found the only video on the internet that agrees with you.

Linus moves the goalposts as regards noise and then he moves the goalposts as regards heat removal. He's got form for this, as have many people who earn their money from traffic. Say something controversial to drive traffic. Doesn't have to be true, just controversial.

I hope this post doesn't come across as too combative or argumentative because that's not my intention at all. I do think you are a little blinkered, though.

 

  • Like 1
  • Haha 1

Share this post


Link to post
Share on other sites

Cooler? May be by few degrees, but this is achieved by letting fans and pump spin at higher rpms, which generates more noise + density of radiator fins also increases noise level.

Reliable? More than several years ago, sure, but not more than air coolers.

Water only if going custom or MO-RA, but AiO...

Pay >50€ more for a few degrees less with more noise, more risk of failure.

I've never affirmed that the i7-5775C is as powerful as the i9-9900K (in ArmA). See YAAB list.

 

It's not my point of view, it's a fact, proven by a lot of tests if you know which are correct. Not those (paid/sponsored majority), where only a 4 cores CPU with Hyper-Threding is used, with conservative OC and voltage and stressed for a short period of time, to not let water heat to the level it can, after extended game session or after rendering, thus not reflecting the reality, which I've already explained earlier in this thread. But people just see/read, what is more convenient to them.

I don't care if it's Linus or XYZ. I've linked his video, because it reflects the reality (surprisingly). It could have been an other video or screenshot or review with somebody else involved.

 

Linus could have been sponsored by Noctua, sure. But if what he's showing reflects the reality, it's the only reason I've linked his video. If it wouldn't have been true, I wouldn't have linked it, even if it's Linus.

You know I've been overclocking since 2005 and have seen/compared a lot of hardware and coolers in person and I know what I'm talking about.

I don't link stuff and affirm it to be the truth because of a youtuber(s), screenshot(s) or review(s).

 

So if the industry pumps a lot of $$$ into marketing and pays youtubers and sites for certain results and the majority just eats all of this, because they know even less than those who promote all this stuff, so it has to be true? lol...

If 10 persons will tell you that a car is black despite it being white in reality, you will stick to their "opinion" to be part of the majority to avoid being outsider because of having alternative point of view, especially if this car is actually white?

Share this post


Link to post
Share on other sites

Please sign in to comment

You will be able to leave a comment after signing in



Sign In Now

×