CPU and GPU overclocking does sweet f a.

[windies 11]

and what exactly do you want to say? overclocking have limits? Or: to use arma3-benches is for arma3 better than synthetic benchmarks? Hell.....YES! :rolleyes:

I'm saying overclocking is not the answer to performance issue's with ArmA or really any piece of software. If you have to overclock just to get good performance and there's no proper scaling within the program for modern hardware, something is wrong with the program not the hardware. It's become some sort of answer to every problem. Poor performance? Overclock everything in you system!

And yes, overclocking has limits. Continue on with your "debate" though. :rolleyes:

Edited March 5, 2014 by Windies

[jumpinghubert 49]

I'm saying overclocking is not the answer to performance issue's with ArmA or really any piece of software. If you have to overclock just to get good performance and there's no proper scaling within the program for modern hardware, something is wrong with the program not the hardware. It's become some sort of answer to every problem. Poor performance? Overclock everything in you system!

And yes, overclocking has limits. Continue on with your "debate" though. :rolleyes:

the problem is you can´t distinguish between two complete different things:

a) to excuse arma3

b) to give a workaround for an existing problem.

your position: oh no this workaround supresses the truth of a bad programmed arma3!

The "debate" is only in your brain. LOL

[Sneakson 1]

4770K 3.5 GHz vs 4.4 GHz was 40 to 50 fps while flying in a chopper over empty Stratis and 36 to 44 fps while flying in a chopper over empty Altis however in Showcase: Infantry the difference was only 53 to 56 fps...

However, I have already compared 4.0 GHz vs 4.4 GHz which I believe is more interesting because correct me if I’m wrong but the 4670K/4770K automatically spin up to 3.9 GHz (basically 4.0 GHz) whenever it is needed, don’t they?

In 4.0 vs 4.4 GHz I haven’t noticed any big differences. I did the flying helicopter over empty Stratis thing the exact same and ended up with 49 instead of 51 fps.

Which leads me to my conclusion that overclocking on the Haswells is quite useless.

Everything just tried once.

[windies 11]

the problem is you can´t distinguish between two complete different things:

a) to excuse arma3

b) to give a workaround for an existing problem.

your position: oh no this workaround supresses the truth of a bad programmed arma3!

The "debate" is only in your brain. LOL

Overclocking is NOT a workaround and you're extremely naive to think that. That is the point whether you can grasp it or not.

[Ezcoo 47]

Overclocking is NOT a workaround and you're extremely naive to think that. That is the point whether you can grasp it or not.

I don't think using known incorrect arguments (such as ad hominem) would make your point valid. It rather tells that you're trying to cover the lack of proper arguments with unrelated personal attack.

How is overclocking not a workaround if it can increase the performance of the game by multiple tens of percents? Or is it again the logical fallacy of perfectionism, that overclocking isn't acceptable solution as it doesn't fix everything even though it would still increase the performance? Or are you afraid of that if people overclocked their CPUs, devs would start to think that "Hey, we don't have to optimize the engine anymore because people are apparently overclocking to increase performance?" I don't understand you.

[jiltedjock 10]

OC'ing is just a complete waste of time and defeats the purpose of an intel CPU anyway which is to get a good performance out of a chip that is power-efficient. It is risky, and reduces the life of the CPU considerably. Because you are running it at higher voltages and higher clock speeds than what it is designed for.

The OP seems to be unaware of what the specs of typical chips are. The i5 2500k has a maximum multiplier of 57, for example.

That means that my overclocked CPU - i5 2500k @ 4.7Ghz, air cooled, max operating temp of 69 degrees, is operating entirely within spec. Which perhaps explains why it has been running like this for 3 years with no issues - it's designed to.

Edited March 6, 2014 by jiltedjock

[jumpinghubert 49]

Overclocking is NOT a workaround and you're extremely naive to think that. That is the point whether you can grasp it or not.

then i am naive and have fun with my bad bad workaround with 25-30% performance gain :D

[Polymath820 11]

The OP seems to be unaware of what the specs of typical chips are. The i5 2500k has a maximum multiplier of 57, for example.

That means that my overclocked CPU - i5 2500k @ 4.7Ghz, air cooled, max operating temp of 69 degrees, is operating entirely within spec. Which perhaps explains why it has been running like this for 3 years with no issues - it's designed to.

69.0*C... you are kidding the actual heck...

http://ark.intel.com/products/52210/intel-core-i5-2500k-processor-6m-cache-up-to-3_70-ghz

Tcase = 72.6°C you are 69.0 - 72.6 = 3.6*C off the maximum temperature of your CPU... you are looking to fry it getting that close to the T-case threshold. Lets not forget ambient temperature. affects your overall systems temperature... as well a 30*C room will result in much higher temperatures than that. And it may not be the "CPU" operating at all but how long you have been running it, for 3 years running it that close to it's maximum operational temperature you have already probably caused damage. Without even noticing.

CPU BBQ anyone?

Edited March 6, 2014 by Polymath820
Additional information

[CaptainObvious 95]

Tcase = 72.6°C you are 69.0 - 72.6 = 3.6*C off the maximum temperature of your CPU... you are looking to fry it getting that close to the T-case threshold.

The cpu will throttle down before frying. I suggest you study the subject a bit more before stating false info as facts.

[jiltedjock 10]

69.0*C... you are kidding the actual heck...

http://ark.intel.com/products/52210/intel-core-i5-2500k-processor-6m-cache-up-to-3_70-ghz

Tcase = 72.6°C you are 69.0 - 72.6 = 3.6*C off the maximum temperature of your CPU... you are looking to fry it getting that close to the T-case threshold. Lets not forget ambient temperature. affects your overall systems temperature... as well a 30*C room will result in much higher temperatures than that. And it may not be the "CPU" operating at all but how long you have been running it, for 3 years running it that close to it's maximum operational temperature you have already probably caused damage. Without even noticing.

CPU BBQ anyone?

You are very ignorant on this subject and starting to look like someone who needs to read less and understand more. As CaptainObvious says, it is difficult to "fry" one of these chips due to to the thermal throttling.

69 degrees is its peak under maximum load and extreme stress eg Prime 95. It is seldom running like that in normal use. And 4.7 is obviously the boost clock, at idle it is 1.6Ghz. I didn't detail any of this in my earlier post because if you actually understood this subject, you would have assumed it.

As for not noticing damage, if a CPU degrades it becomes unstable, because it requires more voltage for a given clock rate. Instability is pretty noticeable. Mine has not changed in 3 years, as expected.

So, define the CPU "damage" you refer to, that wouldn't be noticed. If you can.

Edited March 6, 2014 by jiltedjock

[kklownboy 43]

You cant over heat a modern CPU. It will just shut down. Now maybe that is Overheating...But the point is, a Modern CPU can hit 100c and not be damaged. You could run a CPU at 85c forever...and it would be not bothered at all. What does Damage a CPU is Voltage. To much, too little, or some combo of "sub voltages" that are not in sync/ Balance.

Been running a Old i7 965x at 4.1~ since 2008. No issues. Just wish it would have been like the newer SBs 2600k which with just two adjustments in the Bios will get you 4.6/4.7 on air...Or even better save more money and a 2500k same, tho there where golden samples in retail that could hit 5.0, on air, stock cooler. There isnt a Intel or AMD CPU made in the last five years that couldn't SAFELY hit 4.0z+and run at that forever. And most will be a easy 4.4.

What goofs people up is the RAM, they think they need to max there ram and its all fubar... 1600 is just fine for all gaming needs, tho you can tweak away for small gains. The OP is not correct any any of his comments.

---------- Post added at 12:42 ---------- Previous post was at 12:33 ----------

While it's not fact, depending on heat levels and heat tolerance of the actual CPU, you may achieve a high overclock but get worse performance due to erring. This is why overclocking isn't some blanket fix for poor performance. I know with the 2 Phenom II processors I had, a 940 BE and a 965 BE now, going over 3.6 ghz actually caused a large drop in performance, especially in ArmA. ...

And that would be a AMD issue with the DDR2.?.And I would also get reduced frames if I OC with the game running and OCing on the fly thru windows. If I set my OverClock in Bios with a Reboot(of course) I would get a gain in FPS. As for errors... No corruption in my HDDs or SDD, and My Memtests are just fine. Six years of OC....

[das attorney 858]

You are very ignorant on this subject and starting to look like someone who needs to read less and understand more. As CaptainObvious says, it is difficult to "fry" one of these chips due to to the thermal throttling.

I'm with you jiltedjock on this one. I've been OCing for years now on my i7 with no issues (idle 35 deg - max 70 deg).

On the BIOS, you have option to shut off CPU if temps get too high. The only time I've tripped it is when I drunkenly ripped out the cable for the CPU cooler with my oafish drunken foot.

No problems other than that.

If readers don't believe that or w/e then fine. I'm not posting to convince anyone otherwise, just to add my experience. Oh and OC does ramp up performance on this game and lots of other CPU bound games/programs. If it doesn't on your machine, then you should re-examine how you are doing it.

[ratszo 17]

then i am naive and have fun with my bad bad workaround with 25-30% performance gain :D

Yeah. If it works ..., really works well, the guy saying, "NOO! It CAN'T work!"., sounds alittle nutty.

Wish i had one of those "K" series --I'd spank it like a naughty school-girl.

[Polymath820 11]

http://lifehacker.com/a-beginners-introduction-to-overclocking-your-intel-pr-5580998

They clearly state there is a lot of risk. I know about the voltage potentials of the CPU and if I am going to screw around with the voltage I'd be stepping in micro-volts,nano-volts and maybe pico-volts. As said, it's safer to just go buy a new CPU. Than run-risks of causing damage and is even explicitly stated that overclocking can be dangerous, and can shorten the life of or permanently damage some of your components if something goes wrong,

So if I have an I5-2400 which is running at about 1.24V -> 1240milli-volts you should step in values of 0.05 which is 50 milli-volts. And again. more voltage you add the more heat you will add. Also, adding more voltage to the CPU will result in more stress placed on the electro-plated wires. namely under peak voltage 0.01326 amps. More current draw along those already stressed electroplated cables will generate "more overall heat" how do I know this, had a GTS 450 running on a 12V rail and it decided to draw 12.8V under maximum load with my little I5-2400 caused my motherboard to goto 80*C.

And killed the HDD

Lastly, remember that no two systems will overclock the same—even if they have the exact same hardware.

And the resistance generated by higher voltage potentials and increased current draw you can't escape.

The old adage is "Just because you can and you are able to does not mean you should". Voiding your CPU warranties is not cool. But that being said getting someone "professional" to overclock your system who are "qualified" will likely result in much better outcomes as if they cause your CPU to BBQ they wear the cost not you.

Edited March 7, 2014 by Polymath820
Additional information

[Grumpy Old Man 3546]

http://lifehacker.com/a-beginners-introduction-to-overclocking-your-intel-pr-5580998

They clearly state there is a lot of risk. I know about the voltage potentials of the CPU and if I am going to screw around with the voltage I'd be stepping in micro-volts,nano-volts and maybe pico-volts. As said, it's safer to just go buy a new CPU. Than run-risks of causing damage and is even explicitly stated that overclocking can be dangerous, and can shorten the life of or permanently damage some of your components if something goes wrong,

So if I have an I5-2400 which is running at about 1.24V -> 1240milli-volts you should step in values of 0.05 which is 50 milli-volts. And again. more voltage you add the more heat you will add. Also, adding more voltage to the CPU will result in more stress placed on the electro-plated wires. namely under peak voltage 0.01326 amps. More current draw along those already stressed electroplated cables will generate "more overall heat" how do I know this, had a GTS 450 running on a 12V rail and it decided to draw 12.8V under maximum load with my little I5-2400 caused my motherboard to goto 80*C.

And killed the HDD

Lastly, remember that no two systems will overclock the same—even if they have the exact same hardware.

And the resistance generated by higher voltage potentials and increased current draw you can't escape.

You make it look like Overclocking is new to you and/or the cause for all the bad things happening in the world.

Yeah. If it works ..., really works well, the guy saying, "NOO! It CAN'T work!"., sounds alittle nutty.

Wish i had one of those "K" series --I'd spank it like a naughty school-girl.

I bought an i7 4770k last year and after some testing at the stock clock I overclocked it to 4.5ghz and arma3 feels like it's running on an entirely different cpu, the performance increase is ridiculous. (at least for me having upgraded from an i7-920@4.2 which was running at that speed since it came out)

Especially for arma3 overclocking your cpu and ram is totally worth it.

[Polymath820 11]

You make it look like Overclocking is new to you and/or the cause for all the bad things happening in the world.

I bought an i7 4770k last year and after some testing at the stock clock I overclocked it to 4.5ghz and arma3 feels like it's running on an entirely different cpu, the performance increase is ridiculous. (at least for me having upgraded from an i7-920@4.2 which was running at that speed since it came out)

Especially for arma3 overclocking your cpu and ram is totally worth it.

I have been warned off it several times. I had a long discussion with an IT technician that said I don't know how many PC's come to the shop who have been overclocked and died.

Overclocking shortens the life of components A LOT.

Come to think of it an interesting question to test if arma 3 benefits from mainly CPU clock speed. Why don't people try overclocking one of the cores of the CPU and only 1 core and make arma 3 run on it? It's been said that programs can run faster on a 1 single thread processor than a quadcore if the processor is high enough clock speed.

Method:

1. Overclock a single core

2. Run arma 3 as -cpuCount=1

3. Check FPS in-game

4. If FPS is higher compared to quad-core we can definitively say that arma 3 does not use mutli-core CPU's very well.

5. If performance is worse then maybe using 1 CPU core faster than the other ones arma 3 does not use could offer benefit? e.g Async'd clocks

6. Repeat 5 times

7. Draw conclusions based on averaging data.

Edited March 7, 2014 by Polymath820
Question?

[Greenfist 1863]

Could you guys test with some simple CPU-heavy benchmarks the effect of overclocking? Like if you already have a stable OC, drop the clocks to stock for a moment and run the test. Just some arbitrary scenario that pushes your CPU to the limit.

I'm interested in how directly proportional clock speeds are with fps. Just for sake of the argument, no need to get in the scientific details. But please mention how many cpus or other components you have fried in the process.

edit. oh Polymath just posted this while I was typing.

Why don't people try overclocking one of the cores of the CPU and only 1 core and make arma 3 run on it? [/b] It's been said that programs can run faster on a 1 single thread processor than a quadcore if the processor is high enough clock speed.

I don't think it works like that. Arma still uses all cores. Forcing it to only one core degrades performance immensely. But just for measuring, sure.

Edited March 7, 2014 by Greenfist

[Polymath820 11]

I don't think it works like that. Arma still uses all cores. Forcing it to only one core degrades performance immensely.

Not what I meant' CPU's are always running Async just make it so those cores run faster than the rest. Namely the cores arma 3 is hogging all the time.

And it has been stated in computer science fields that a single core single thread CPU can outperform a multi-core

this explains it quite well

http://www.extremetech.com/computing/116561-the-death-of-cpu-scaling-from-one-core-to-many-and-why-were-still-stuck/3

So if we keep going we are going to reach a "flat lining of moores law" You want to make systems faster? Learn how to design a new architecture. (Completely Parallel based)

This is where you delve into the area of memresistors:

http://phys.org/news190016024.html

Graphene Transistors:

http://phys.org/news/2013-02-graphene-transistor-principle.html

Photonic CPU's:

http://phys.org/news/2011-09-fujitsu-compact-silicon-photonics-source.html

3D transistor Lithography:

http://phys.org/news/2013-03-transistors-dimension.html

And last but not least single-atom transistors:

http://phys.org/news179331125.html

Regardless of chip organization and topology, multicore scaling is power limited to a degree not widely appreciated by the computing community… Given the low performance returns… adding more cores will not provide sufficient benefit to justify continued process scaling. Given the time-frame of this problem and its scale, radical or even incremental ideas simply cannot be developed along typical academic research and industry product cycles… A new driver of transistor utility must be found, or the economics of process scaling will break and Moore’s Law will end well before we hit final manufacturing limits

Edited March 7, 2014 by Polymath820
We have no event scratched the surface

[jiltedjock 10]

You still don't seem to understand that people are clocking their CPUs within the manufacturer's specs, both clock speed and voltages. For that reason, it is pointless arguing with you, your statements are so generic, meaningless, and unsupported. As for your "test", it's laughable. Arma is not a single threaded application.

Enjoy the game, those of us with high clock speeds are enjoying it a LOT more. I am not going to argue with you anymore as I recognize the signs of your condition

Edited March 7, 2014 by jiltedjock

[CaptainObvious 95]

Could you guys test with some simple CPU-heavy benchmarks the effect of overclocking? Like if you already have a stable OC, drop the clocks to stock for a moment and run the test. Just some arbitrary scenario that pushes your CPU to the limit.

I'm interested in how directly proportional clock speeds are with fps. Just for sake of the argument, no need to get in the scientific details. But please mention how many cpus or other components you have fried in the process.

If you could kindly provide me with a benchmark mission, I can run it on stock and 4.7GHz.

Fingers crossed I don't fry anything :pet13:

[Greenfist 1863]

If you could kindly provide me with a benchmark mission, I can run it on stock and 4.7GHz.

Fingers crossed I don't fry anything :pet13:

I don't have any one particular mission handy. You could just dump a bunch of fighting AI down in the editor to get lower fps. That should be analogous enough to common real life scenario where the CPU is mostly the bottleneck. Then run a minute or two benchmark with something like FRAPS to get average fps.

[Sneakson 1]

If you could kindly provide me with a benchmark mission, I can run it on stock and 4.7GHz.

Fingers crossed I don't fry anything :pet13:

Start Showcase: Infantry and Showcase: Helicopters and play them both through twice while going the same way every time and you should have quite accurate measurements actually.

Do it more times or with more missions with more/less action going on if you want.

What's your CPU?

[CaptainObvious 95]

I don't have any one particular mission handy. You could just dump a bunch of fighting AI down in the editor to get lower fps. That should be analogous enough to common real life scenario where the CPU is mostly the bottleneck. Then run a minute or two benchmark with something like FRAPS to get average fps.

Start Showcase: Infantry and Showcase: Helicopters and play them both through twice while going the same way every time and you should have quite accurate measurements actually.

Do it more times or with more missions with more/less action going on if you want.

I'd prefer something automated without player interaction to minimize the possibility to influence the FPS by doing something different between runs, I once had a benchmark mission which had a bird's eye camera roaming in the battlefield and AI doing it's stuff, in the end the mission printed the avg.fps, I can't seem to find it atm :icon_redface:

What's your CPU?

2500K

Edited March 7, 2014 by CaptainObvious

[trnapster 12]

Overclocking shortens the life of components A LOT.

1. I'm overclocking since about 15 years and never had a problem with a dead CPUs/GPUs

2. Even if the CPUs are dying faster... Who cares if it dies in 5 years instead of 8. If you want to hold on to a CPU longer than 3-4 years then I know why you have performance problems...

[Sneakson 1]

I'd prefer something automated without player interaction to minimize the possibility to influence the FPS by doing something different between runs, I once had a benchmark mission which had a bird's eye camera roaming in the battlefield and AI doing it's stuff, in the end the mission printed the avg.fps, I can't seem to find it atm :icon_redface:

2500K

It's over in the user missions section. Google for ARMA benchmark and it should pop up.

I didn't like it a lot because it doesn't give you decimals. If I'm going to bother trusting a benchmark scenario I want some damn decimals! :p

Also benchmarks don't always correlate well to the actual game. You're going to have to benchmark actual missions later anyways, because that's what matters.

As I mentioned earlier when I compared 4770K 3.5 vs 4.4 GHz I had a 40->50 fps increase on empty Stratis however in Showcase: Infantry only 53->56 fps... and it's the later one that really matters.