Anyone tried with dual CPU´s

[aaman 0]

Has anyone tried to play Arma2 with dual CPU´s and could tell about the experiance? :)

[larsiano 12]

I expect you mean two actual fiscal CPU's like most servers use? This does not add anything spectacular to 3D games cause its meant to handle more then just one application (treat) at a time. Best performance gain is usually on the GPU front. But there are exceptions:

link:

[aaman 0]

I expect you mean two actual fiscal CPU's like most servers use? This does not add anything spectacular to 3D games cause its meant to handle more then just one application (treat) at a time. Best performance gain is usually on the GPU front. But there are exceptions:

link:

So you say so, i actually know a guy who had a very old motherboard with two CPU´s (And there are many new out there)

My question was not if it was good or bad for 3d games.

The question was if anyone has tried it with ArmA2 and how big was the improvement in preformance.

There would be interesting to know if the improvement was 100% or only 50%.

[Leon86 13]

I can assure you arma2 cant effectively use much more than 3 cores. A quad is never fully loaded. There's always 1 core on 100% and the rest 40-70% or so. There's absolutely no point in buying anything more than a quadcore for A2.

[tpw 2314]

Unfortunately, as anyone who has done cluster computing will tell you, N x CPUs does not equal N x performance. The overhead of parcelling out portions of computation to different CPUs (or cores) eats into the performance gain. Same goes for SLI graphics cards.

[msy 22]

Quad cores are better than dual cores in arma2

[aaman 0]

Not cores!

CPU´s.

Ok let´s put it like this.

If you have one CPU at 4000 Mhz in one computer.

And 2 CPU´s with 2000x2 Mhz in the other computer.

Witch computer will run ArmA2 best, and how much is the difference?

[larsiano 12]

Oké more info:

I have a dual quadcore server system (2 x E5420 @2,5Ghz) that is faster then my gaming computer (runs dual core E8600 @ 3,33Ghz) when it comes to processing power & memory bandwidth. But servers are not optimized to run games and therefore any normal PC will be faster then a dual CPU workstation when it comes to gaming. The question you are really asking is: Is a computer that costs 5000$ faster then a computer that costs only 1000$ and the answer is yes :) Hind have a look at http://www.cpubenchmark.net/ to see the different calc. power of different processors and there (dual) cpu performance

Comparison:

E8600:

calc. sec: 6,7

mem. ops. sec: 6,8

desktop perf: 7,3

3D perf: 7,3

dual cpu E5420:

calc. sec: 7,6

mem. ops. sec: 7,6

desktop perf: 6,8

3D perf: 6,8

Note: server was running 8800GT GPU, gaming system had a: EN285GTX

[tpw 2314]

Not cores!

CPU´s.

Ok let´s put it like this.

If you have one CPU at 4000 Mhz in one computer.

And 2 CPU´s with 2000x2 Mhz in the other computer.

Witch computer will run ArmA2 best, and how much is the difference?

Depends on how many cores each CPU has.

Let's say you had a 2GHz single core CPU in one computer, and 2 x 1GHz single core CPUs in another computer. All else being equal, you'd probably get better performance out of the 1 x 2GHz computer because it is not wasting cycles splitting computations to different CPUs and recombining results. However, if you compare a 1 x 2GHz SC with a 2 x 2Ghz SC, you might reasonably expect an increase in performance (especially if the software is written to take advantage of it), but not a 100% increase.

I'm unsure what happens if you have 1x 4GHz quad core vs 2 x 2GHz quad cores :) Any comp sci PhDs care to chime in?

Edited December 7, 2010 by tpw

[aaman 0]

Oké more info:

I have a dual quadcore server system (2 x E5420 @2,5Ghz) that is faster then my gaming computer (runs dual core E8600 @ 3,33Ghz) when it comes to processing power & memory bandwidth. But servers are not optimized to run games and therefore any normal PC will be faster then a dual CPU workstation when it comes to gaming. The question you are really asking is: Is a computer that costs 5000$ faster then a computer that costs only 1000$ and the answer is yes :) Hind have a look at http://www.cpubenchmark.net/ to see the different calc. power of different processors and there (dual) cpu performance

Comparison:

E8600:

calc. sec: 6,7

mem. ops. sec: 6,8

desktop perf: 7,3

3D perf: 7,3

dual cpu E5420:

calc. sec: 7,6

mem. ops. sec: 7,6

desktop perf: 6,8

3D perf: 6,8

Note: server was running 8800GT GPU, gaming system had a: EN285GTX

So now if you set every possible setting on low in game but the viewdistance and object detail, keep it at 10.000 meters.

And now measure the game preformance, witch system will be fastest?

CPU´s are not cheep, i know that.

When the hunt goes for preformance i think the best is to have 2 or even 4 i7´s. Or E8600´s

And yes, i know they are not cheap.

Edited December 7, 2010 by aaman

[Pyronick 21]

No, the overhead would be too big.

Dual CPU's communicate over the front-side bus. I used to have a dual Slot 1 system, it was nice in Quake 3 in the beginning, but I only gained about 5-7% framerate.

It's ineffectiveness lead to the MCM setup (NetBurst dualcores, Core 2 quadcores) to somewhat improve latency and memory issues and eventually to the superior multicore (AMD and Intel Core 2 dualcores and newer) chips.

[aaman 0]

Common there are lots of 2 and 4 CPU space moterboards out there.

Did not find a 4 x i7 yet, but im sure it will come soon.

I have worked a lot with model´s past months and i know how CPU´s are affected by this. (2D or 3D does not matter, it´s the total number that´s important)

So now i just asked fore some honest preformance numbers before i buy a new motherboard with 2 cpu´s.

I could say that i got a little sad (pissed off) when i spent 400$ on a GPU and noticed it made no difference in preformance, ok you could have higher graphics settings but the preformance did not improve.

And also i was forced to buy a new mother board to this F*****g GFX card.

Edited December 7, 2010 by aaman

[TechnoTerrorist303 10]

There would be little or no performance gain on Arma2 or any game if you had more than one multicore cpu. A quad core cpu will always (I would have thought) run quicker than 2 dual core cpus on a dual board, since nothing really stresses a 4 core cpu (except prime 95) then there is no advantage to having more than one cpu.

[aaman 0]

Well, the tests i tried with single and quad cores of the same mhz, the quad even has 200 mhz more.

I tell you there was no difference in preformance at all.

[TechnoTerrorist303 10]

How were you measuring performance? What tests did you run? If you're talking about just FPS in a specific mission then it's quite possible you wouldn't notice any change.

Also, if you have run tests with single and quad cores and not noticed any difference, what twisted logic are you applying to conclude that more cpus will increase performance?

[aaman 0]

Do you really think a core in a quadcore is the same as a CPU?

I also tought that, but that was 6-7 years ago, when the seller of a quadcore CPU convinced me that togheter all cores would give me 9000 mhz.

Dumb as i was i thougt the computers really had did that progress so fast.:)

Same time 10 years ago i had a laptop with 3 ghz CPU, now im sorry i throw it away then.

[TechnoTerrorist303 10]

Do you really think a core in a quadcore is the same as a CPU?

I also tought that, but that was 6-7 years ago, when the seller of a quadcore CPU convinced me that togheter all cores would give me 9000 mhz.

Dumb as i was i thougt the computers really had did that progress so fast.:)

Same time 10 years ago i had a laptop with 3 ghz CPU, now im sorry i throw it away then.

I don't really understand what you think you know about the insides of CPUs or how Windows uses them...

1 CPU = 1 CORE for the sake of this argument since that's how Windows sees them. As far as Windows is concerned I have 4 CPUs in my PC. I am not an idiot, I know that 4 cores running at 3ghz doesn't mean I have a 12ghz machine, that would be stupid but I do know that 1 quad core chip is faster than 2 dual core chips of the same speed running on the same board.

I have a PC sat behind me that has a 3.4ghz CPU. Would I use it to run Arma2? NO! I have a quad core 3.15 ghz cpu in my main rig that runs the game faster.

[Rewindlabs 10]

In short you are not onto a goldmine of performance by any means nor an area that hasn't already been trekked to death by gamers/performance seekers before you

[spangg 0]

And to get some constructive feedback as well:

invest your money in a decent quad-core, fast memory(you have a new motherboard so hopefully it supports ddr3?) and you may get a slight performance gain from a faster hdd or maybe an ssd? These are still will be cheaper than build a server of CPUs just to learn that it didn't improved as expected. If you give exact spec of your system some of us may help to point out the weak point. If necessary....

[SWAT_BigBear 0]

and you may get a slight performance gain from a faster hdd or maybe an ssd?

Yes, that is what he should aim for.

I have an old 2x2 setup and if I disable 1 cpu, I can't really notice a difference in ArmA2.

Where I seen the big gain, was the fast hd's with raid-0.

[bangtail 0]

Dual CPUs are great for benchmarking but that's it with things even remotely related to gaming.

Faster CPU (single), SSD, GPU, etc are much better places to spend money

[kyfohatl 10]

I've got dual CPU. It's a 2.33 GHz, so its getting a bit old and slow, however it does run ArmA 2 just fine ON NORMAL SETTINGS. When I set it to high, or there are too many troops on the map (large battles), you can get significant lag. I don't think it has that much to do with my graphics card, as When I check with task manager when the game is lagging, I get 95% or so CPU usage (so it probably means it is a CPU problem). However, newer and stronger Dual CPUs should run the game fine.

[binkster 0]

From my expierence with overclocking I get major performance boost with higher frequency.

But it was with a quad. For example I have Q9650 running at 3.0ghz. I could almost increase FPS by 50% when I overclocked it to 4GHZ so people that say buy a better GPU should really do some research.

Check this thread out and you can get a better idea on where performance gains are.

[jw custom 56]

But it was with a quad. For example I have Q9650 running at 3.0ghz. I could almost increase FPS by 50% when I overclocked it to 4GHZ

I barely get any difference if my core i720 is running @ stock 2.66ghz or it's overclocked to 3.6+

[binkster 0]

I barely get any difference if my core i720 is running @ stock 2.66ghz or it's overclocked to 3.6+

Thats weird I get huge results. What OS are you using? I remember having no difference between clocks on VISTA. As soon as I went back to XP I was getting 20fps more from stock clocks to a minor overclock of 600mhz. Alot of memory tweakage too.

Edit... Also I didnt mention that I am using a gtx285. Im assuming you have a decent gpu if you have an I720?

Edited December 12, 2010 by binkster