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

[jumpinghubert 48]

thanks for ur time.i think ill go with 760

??....but its 12% slower...:confused:

[Sneakson 1]

I'd buy the 770...

I haven't seen any recent GPU benchmark and I really think we need one but there's this old one from version 1.0:

http://www.techspot.com/review/712-arma-3-benchmarks/page3.html

It says the 760 and 7970 are identical (so buy the cheaper one) but the 770 is 35% stronger and in Sweden/Amazon.com 16%/24% more expensive... so it's worth it.

With a 770 you'll max everything except for object distance as usual, in singleplayer, with high fps.

[jumpinghubert 48]

It says the 760 and 7970 are identical

my second :confused: for today :p

[Sneakson 1]

my second :confused: for today :p

May very well be a dated benchmark.

I did see another benchmark from last year too though that said the 770 is 14% better than the 7970...

In a Battlefield 4 benchmark the 7970 scores 10% better than the 770 though...

I'm confused. AMD lol...

Googling a bit more before purchase might be a good idea.

[jumpinghubert 48]

760 = 770? my third :confused: for today....look one or two postings above AND read them please :p

[Sneakson 1]

760 = 770? my third :confused: for today....look one or two postings above AND read them please :p

One benchmark (above) says 760 = 7970 < 770...

Anther one says 7970 < 770

And the Battlefield 4 benchmark says 7970 > 770...

:p

[jumpinghubert 48]

One benchmark (above) says 760 = 7970 < 770...

Anther one says 7970 < 770

And the Battlefield 4 benchmark says 7970 > 770...

:p

your own linked arma3 benchmark says clearly 760 < 7970. The higher the resolution the more the difference between both cards. My only question is why he sold his 7970 (ghz?).

[KrazyBee 8]

Thanks for ur info guys.i got 770.hope to run everything maxed out:)

[Sneakson 1]

your own linked arma3 benchmark says clearly 760 < 7970. The higher the resolution the more the difference between both cards. My only question is why he sold his 7970 (ghz?).

Oh, I was looking at 1680x1050. In 1920x1200 the 7970 is 10% better indeed.

[oldbear 390]

I have done some tests about RAM on the "i3-4130/HD 7770/256Go SSD" rig I am testing at the moment.

RAM used for the tests are all from Kingston brand :

RAM 1600 : Kingston KH1600 C9D3B1K2

RAM 1866 : Kingston KH1866 C9D3B1K2

All the tests have been done using Arma3Mark by Helo on ArmA 3 Benchmark v0.51.Stratis

Results :

4 Go 1600

- 2000 m : 46/46/46

- 4000 m : 41/41/40

8 Go 1600

- 2000 m : 46/47/46

- 4000 m : 40/41/40

16 Go 1600

- 2000 m : 46/46/46

- 4000 m : 39/40/40

16 Go 1866

- 2000 m : 45/46/45

- 4000 m : 41/40/40

In fact most of the variations in results are related to AI behavior.

Based on these tests, it seems the RAM amount as well as the small increase of RAM speed from 1600 to 1866 has no effect on FPS.

Video settings General : http://www.armedassault.info/ftp/pics/news/pics1/video_settings_mem_test1.jpg

Video settings Display : http://www.armedassault.info/ftp/pics/news/pics1/video_settings_mem_test2.jpg

Video settings AA&PP : http://www.armedassault.info/ftp/pics/news/pics1/video_settings_mem_test3.jpg

[tonschuh 3]

Haswell Real World Performance: DDR3-1600 is Not Enough

The prevailing wisdom in the enthusiast community has been, for generations, that DDR3-1600 is the sweet spot and that faster memory offers at best extremely limited performance improvement and that at worst, it’s snake oil. There’s an element of truth to that; AMD’s Bulldozer architecture and its derivatives see arguably minimal benefit from faster memory, and Ivy Bridge and its predecessors actually were just fine at DDR3-1600. So the idea that the paradigm might have shifted is tough to swallow because it goes against wisdom that’s been ingrained for years, a veritable lifetime in our industry.

Except that it has. DDR3-1600 is quite simply no longer enough for modern chips outside of Ivy Bridge-E and Vishera. That Kaveri benefits from faster memory (at least on the GPU side) is a foregone conclusion that was confirmed by our testing. AnandTech already exhaustively detailed performance scaling with different memory speeds on Haswell, and I’ve studied the effect of memory speed on Battlefield 4’s performance. Between our work and AnandTech’s extremely thorough research, you’d think there would finally be a pervasive understanding of the benefit of faster memory on Haswell, but that hasn’t been the case.

I originally went into this testing specifically trying to determine whether or not overclocking would increase the strain enough on Haswell’s memory controller to justify higher speed memory. In testing, I discovered fairly conclusively that DDR3-1600 essentially leaves performance on the table even at stock clocks.

For testing I ran Intel’s Core i7-4770K at stock speeds and overclocked to 4.5GHz. A 32GB (4x8GB) kit of our Dominator Platinum DDR3-2400 was used to scale from DDR3-1600 CAS 9 to DDR3-2400 CAS 10. Test system specs are as follows:

- Intel Core i7-4770K CPU

--- Stock Speed (3.5GHz nominal, turbo to 3.7GHz on four cores or 3.9GHz on one core)

--- Overclocked (4.5GHz, 45x100 BClk, 4GHz Northbridge

- Gigabyte G1.Sniper 5 Motherboard

- 4x8GB Corsair Dominator Platinum DDR3-2400

--- DDR3-1600 (9-9-9-24 CR2)

--- DDR3-1866 (9-9-9-24 CR2)

--- DDR3-2133 (10-11-11-31 CR2)

--- DDR3-2400 (10-12-12-32 CR2)

- NVIDIA GeForce GTX 780 Overclocked (980MHz nominal, boost to 1150MHz, 7GHz GDDR5)

- 240GB & 480GB Neutron GTX SSDs (for Adobe testing)

I very deliberately chose a mixture of synthetic and real world benchmarks. Cherry picked synthetics can admittedly overstate the importance of higher speed memory; I wanted tangible, demonstrable, practical benefits.

Overclocking the CPU itself had virtually no effect on memory bandwidth, producing results essentially within the margin of error. We’re just going to measure the raw amount of bandwidth made available to the i7 as memory clocks increase.

There’s a very steady increase in bandwidth going from step to step, but read speed tapers off moving from DDR3-2133 to DDR3-2400. This isn’t surprising; Kaveri also started to get shaky around DDR3-2400. Discovering Haswell’s memory controller’s “breaking point†may be worth looking into in the future. Nonetheless, we’re seeing roughly 18% improvements in raw memory bandwidth at each step until DDR3-2400.

Now we’ll see if that translates at all in the two synthetic benchmarks I’ve included. First up is the x264 HD 5.0 benchmark, which absolutely hammers the CPU.

This benchmark is almost entirely CPU limited, but there are trends to point out: overclocking increases the effect memory bandwidth has on performance (proving the initial hypothesis), and the jump from DDR3-1600 to DDR3-1866 is the most significant. The second pass sits almost entirely on the CPU and absolutely hammers it, but the first pass is able to eke out small gains.

The next synthetic is the built-in benchmark in 7-Zip. WinRAR has long been a stronghold of fast memory, but I haven’t used it in ages; 7-Zip is free, fast, and it works.

7-Zip shows modest but steady increases in performance as memory speed goes up, with an unusual jump at DDR3-2400 at stock. These aren’t the kinds of massive gains you might see in WinRAR, but they’re definitely present.

Where things really get interesting are with my two big practical benchmarks: Adobe Media Encoder CC and StarCraft II: Heart of the Swarm.

Both of these benchmarks were run with CUDA acceleration enabled in Mercury Playback Engine, and both of them see substantial improvements in running time. The AVCHD encode is able to shave off about 10 seconds, while you can save a full minute with the HDV encode. In fact, on the HDV encode, running the stock CPU with DDR3-2133 or DDR3-2400 instead of DDR3-1600 actually gets you pretty close to the 4.5GHz overclock with DDR3-1600.

The Hail Mary in the group is StarCraft II: Heart of the Swarm. Testing was done at 1080p with all of the settings maxed out, so it at least looks the same way it might on your home system. This is also one of the few games that could theoretically benefit from a framerate higher than 60fps, and any kind of demonstrable performance benefit on the CPU/memory side is valuable.

If you run the i7 at stock with DDR3-2400, it’s basically as fast as a 4.5GHz i7 with DDR3-1600. Bump the 4.5GHz chip’s memory to DDR3-1866 and it starts to soar. Getting roughly 10fps out of something as simple as faster memory is staggering.

When we take a look at the performance results holistically, we see our real world benchmarks are able to eke out a minimum of 5% improved performance by going up to DDR3-2133; even going up to DDR3-1866 still nets a solid jump. DDR3-2400 still has plenty to offer in some cases. Add overclocking to the mix and the performance gaps widen even more.

As far as I’m concerned, the conclusion is simple: DDR3-1600 isn’t enough for Haswell, and it leaves performance on the table. This testing was done with DDR3-1600 CAS 9, when DDR3-1600 CAS 11 is actually exceedingly common and thus even slower. If we’re willing to overclock every component in our system to extract as much as five or ten percent more performance, it seems absurd at this point to cheap out on memory. You can get 16GB of DDR3-1866 CAS 9 for nearly the same price as 16GB of DDR3-1600 CAS 9 on our web store, and DDR3-2133 CAS 10 or CAS 11 isn’t much more than that.

With Haswell I continue to be convinced that DDR3-1866 is the new entry level, and that DDR3-2133 is really the sweet spot. Performance improvements from DDR3-2133 to expensive DDR3-2400 are less consistent, and Haswell’s IMC itself starts to get a little shaky there. I may do more testing in the future to determine where the IMC’s limit is; my experience was that under some circumstances, there’s a slight performance regression from DDR3-2400 when you get to DDR3-3000 (which is an absolutely ridiculous speed). In the meantime, the conclusion remains clear: for modern and especially high performance Haswell and Kaveri systems, DDR3-1600 isn’t enough.

Source: Click

:)

[oldbear 390]

Well it's an undisguised attempt to advertise Corsair Dominator Platinum DDR3-2400, but it's desperately missing informations about how this overpriced RAM is working Arma3 wise.

Around me there are a lot of youngster without much money wanting to play.

So my main concern is : how to build an Arma3 dedicated rig at the lower price ?

Kingston ValueRAM 8 Go (2 x 4 Go) DDR3 1600 MHz CL11 SR X8 = 74€

Kingston HyperX Genesis XMP Grey 8 Go (2x 4Go) DDR3 1600 MHz CL9 = 82€ (the one I have used in the test)

Corsair Dominator Platinum 8 Go (2 x 4Go) DDR3 2400 MHz CL10 = 160€

[jumpinghubert 48]

I have done some tests about RAM on the "i3-4130/HD 7770/256Go SSD" rig I am testing at the moment.

RAM used for the tests are all from Kingston brand :

RAM 1600 : Kingston KH1600 C9D3B1K2

RAM 1866 : Kingston KH1866 C9D3B1K2

All the tests have been done using Arma3Mark by Helo on ArmA 3 Benchmark v0.51.Stratis

Results :

4 Go 1600

- 2000 m : 46/46/46

- 4000 m : 41/41/40

8 Go 1600

- 2000 m : 46/47/46

- 4000 m : 40/41/40

16 Go 1600

- 2000 m : 46/46/46

- 4000 m : 39/40/40

16 Go 1866

- 2000 m : 45/46/45

- 4000 m : 41/40/40

In fact most of the variations in results are related to AI behavior.

Based on these tests, it seems the RAM amount as well as the small increase of RAM speed from 1600 to 1866 has no effect on FPS.

Video settings General : http://www.armedassault.info/ftp/pics/news/pics1/video_settings_mem_test1.jpg

Video settings Display : http://www.armedassault.info/ftp/pics/news/pics1/video_settings_mem_test2.jpg

Video settings AA&PP : http://www.armedassault.info/ftp/pics/news/pics1/video_settings_mem_test3.jpg

do you have used exactly same timings?

i have quit different results, locked ram timings, helo´s bench, all ultra, 1920x1200, 3000view distance, 3000object distance, 200m shadow:

cpu 4.5Ghz/ram 1600Mhz: 36 fps

cpu 4.5Ghz/ram 2400Mhz: 42 fps

cpu 4.5Ghz/ram 2666Mhz: 43 fps

cpu 4.8Ghz/ram 1600Mhz: 37 fps

cpu 4.8Ghz/ram 2666Mhz: 45 fps

cpu 4.84Ghz/ram 2694Mhz: 46fps (Bclk @100.9)

sorry, made only 1x run

@ 11/14/13/35/1T

http://geizhals.de/patriot-viper-3-black-mamba-dimm-kit-8gb-pv38g240c0k-a962829.html

Edited March 24, 2014 by JumpingHubert

[oldbear 390]

Some more infos about the RAM used during the tests :

RAM 1600 : Kingston KH1600 C9D3B1K2 DDR3 @ 1600 MHz 9-9-9-27

RAM 1866 : Kingston KH1866 C9D3B1K2 DDR3 @ 1866 MHz 9-11-9-27

In fact I have done more than a hundred runs and cross match results after

- restarting game

- rebooting PC

My own feeling is that the 2 RAM sets characteristics are too close.

I will like to have some more demanding testing environments, a bit like the "Benchmark 02" we had on Arma2.

[tonschuh 3]

Well it's an undisguised attempt to advertise Corsair Dominator Platinum DDR3-2400, but it's desperately missing informations about how this overpriced RAM is working Arma3 wise.

Around me there are a lot of youngster without much money wanting to play.

So my main concern is : how to build an Arma3 dedicated rig at the lower price ?

Kingston ValueRAM 8 Go (2 x 4 Go) DDR3 1600 MHz CL11 SR X8 = 74€

Kingston HyperX Genesis XMP Grey 8 Go (2x 4Go) DDR3 1600 MHz CL9 = 82€ (the one I have used in the test)

Corsair Dominator Platinum 8 Go (2 x 4Go) DDR3 2400 MHz CL10 = 160€

As @JumpingHubert already pointed out, ARMA-3 usually benefits from faster RAM.

I had 16GB 1600MHz-Corsair Vengeance (2x CMZ8GX3M2A1600C8), which cost me back then AU$ 300,- and didn't come down much in price.

Now I use 16GB 2400MHz-G.Skill-TridentX (1x F3-2400C10Q-16GTX), which were only AU$ 250,- and didn't come down much in price either.

When you have a look at @JumpingHuberts RAM, then you will see, that you are able to get fast RAM at a bargain-price, if you shop around a bit.

If you can only afford "faster" RAM with high latencies, then you would be most likely better of to buy just a slightly "slower" Kit, but with a lower latency.

But back to my post:

I only pointed out, that if you go down the Haswell-Path, which is possible not in your price-range, that you should not go for 1600MHz-RAM-Kit's, as you would waste some of your CPU-Power via RAM-Bottleneck.

I'm not saying that you need or have to go down the Haswell-Path to be able to play ARMA-3, but some people don't build / upgrade a rig only for one game or one purpose.

:)

[jumpinghubert 48]

yep tonschuh, i had to pay for my 8GB 2400er ram 99 Euro.

@OldBear

Low/sharp timings are very important in ram overclocking. With good cooling its possible to give the ram 1.775V to get good results. I used memtest to check ram-errors and maxmem2 to check ramspeed. In arma3 a better gpu give you "bad" performance gains (better maximum fps and better "editor"-fps), higher cpu/ram speeds gives you in arma3 "good" performance gains (higher min. fps in ai heavy situations).

[Sneakson 1]

Some more infos about the RAM used during the tests :

RAM 1600 : Kingston KH1600 C9D3B1K2 DDR3 @ 1600 MHz 9-9-9-27

RAM 1866 : Kingston KH1866 C9D3B1K2 DDR3 @ 1866 MHz 9-11-9-27

In fact I have done more than a hundred runs and cross match results after

- restarting game

- rebooting PC

My own feeling is that the 2 RAM sets characteristics are too close.

I will like to have some more demanding testing environments, a bit like the "Benchmark 02" we had on Arma2.

It's possible that your specs are too weak... the i3 is one of the weaker Haswells, which I would guess means it's comparable to previous generation CPUs. Also 1600-1866 may not be significant enough.

Maybe next week I'll have time to order some 2400s and do a 4770K (or disable HT and viola: 4670K) test with both a 2013 Haswell and 2013 graphics card :p

Not sure what to hope for but at least RAM is cheap :)

[oldbear 390]

@ Sneakson : I have built this "i3-4130/HD 7770/256Go SSD" rig as a base for some experiments and it's "weak" on purpose.

As I have already stated, my main concern are about guys being on a budget and desperately trying to play Arma3 alas very often in order to play Altis Life afterwards.

Go get a look at this chart :

Source : http://www.hardware.fr/articles/913-1/kaveri-amd-a10-7850k-a10-7700k-test.html

As it looks the "weak" i3-4130 was -RealVirtuality Engine wise- more or less on par with all the AMD APU and FX I had decided to stop building my "Haswell top notch" config to replace my actual i7-3770 and install a i3-4130 in lieu of a i7-4770K on the brand new MSI Gamer socket 1150 MoBo.

The Kingston KH1866 C9D3B1K2 DDR3 asset history is a kind of an accident, I have bought it on Amazon at a great bargain price the day before the garbage bin was set on fire in Hynix factory and it was waiting on a shelf to be used somewhere!

As far as I can said ATM, the "i3-4130/HD 7770/256Go SSD" rig is performing quite well in Single as well as in Multi. It can be an excellent alternative to somewhat cheaper but less effective AMD based rigs.

[jiltedjock 10]

Why would anyone buy an i3-4130 when they could buy an unlocked 2500k and overclock it easily to the mid-4s where it would be at least 33% quicker than the 4130 ? Surely that is the sensible choice for a "budget player" ?

[oldbear 390]

@ jiltedjock : because working "out of the box" rig is what the average gamer is looking for, not for a makeshift rig based on outdated hardware.

Your so-called "easy" solution means you find a 2500K to buy as well as a LGA1155 MoBo and then go the OC way on this "old" CPU ... not so easy in fact.

[jiltedjock 10]

Outdated hardware? My 2500k runs at 4.7Ghz on air, 16GB RAM at 8.8.8.24 1600Hz, 4GB GTX680. Do you think me, or the muppet who buys an i3-4130 on your "advice", is going to trade up first because they feel their hardware is outdated in a couple of years? Only 2 physical cores, locked at 3.4Ghz....pretty sure I know the answer to that. All to save 100 Euros?

As for your RAM tests - try running it at 1333 and test again. I bet you get similar results, because the CPU is bottlenecking it.

And the chart on the preceding page? - those are at stock clock speeds. Anyone who buys a CPU based on that chart...well, a fool and his money are easily parted.

Edited March 25, 2014 by jiltedjock

[jumpinghubert 48]

its so easy to overclock a 2500k......and still the ultimative arma3 hardcore fan will play other games too with real 4-core utilization. I think the i3-4130 isn´t a good choice for a gaming-pc.

[Sneakson 1]

I'd buy a 4670K under all circumstances. $100 over the i3 is nothing over the course of two-five years.

[DrSues 10]

win 8

i7 2.8-3.8ghz

2xGTX 870m

16g ram

is that enough specs to get a idea at what settings it will run on?

[KrazyBee 8]

ok so i went and bought 4670k with msi tf 770 but some how im getting shitty fps like 15-30..any idea why??should i replace and get asus 770?