Thread Rating:
  • 0 Vote(s) - 0 Average
  • 1
  • 2
  • 3
  • 4
  • 5
Overclocking , overvolting & CPU Temps !
:hello: H Ello

I have Few Questions about Overclocking .

1.WIll Undervolting Reduce Performance .
2.WHy sometimes we Undervolt ? Whats the Gain in Undervolting ?
3.Overclocking Comes with Overvolting ?
4.How much will be the Performance Difference Between Stock Speed of 3.2Ghz &(VS) Clocked to 4.0GHz ? for example Phenom 955 BEStock vs clocked to 4.0Ghz!

5.IF we Have Room with Low Airflow & rather High Temp! How Well Will Cooler Master Hyper 212+ Cool Down Phenom 2 4x Cpus ? you know regular Temps while gaming are 60-70 ( in room with Bad Condition ) now i want to know How things will be with CM Hyper 212+ ? because i have EP-Z660 But Temps are Almost same but with good Airflow & Cooling of Room i can Reduce temps to 52-58 While Gaming ! If i get CM Hyper 212 WIll i be able to Cool down CPu to Lower Temps & will i be able to OC it ?

6.Imagin Phenom 955 BE , will i5 2400 Bring Too much Difference in games ?

Detailed & Simpler :

Whats the Difference between these 2 Platforms ?

Specialsystem+5870+i5 2400 vs Same rig+5870+phenom 955 ? Will i notice any difference ? how much fps ?

Sponsored links

1. As long as the CPU is stable (running prime95 on all threads for 8 hours or so without crashing) there is no performance issues.
2. Undervolting reduces heat, sometimes dramatically. Voltage is the primary source of heat, but voltage is necessary to push a CPU to the limits of it's overclocking.
3. Not necessarily. Default voltages usually can lead to a good overclock, but you'll only very rarely push a CPU to it's limits on stock.
4. I'd think that's pretty self evident, 4ghz is 25% faster than 3.2ghz so you should see about 25% performance increase in things that are generally CPU limited (IE PCSX2 in most cases) but much less in general gaming performance which is usually GPU limited.
5. Case airflow is more important than room airflow AS LONG AS the room isn't overly hot. If the room is usually 35c high airflow will be a lot less effective. A good CPU cooler is worthless though too if there isn't any outside air getting into the case.
6. In PCSX2? Since that i5 can't be overclocked hardly at all, you may end up with higher performance with the AMD.

At default rates that i5 will likely beat the AMD. Your best bet would be to try to spend a little more and get the i5 2500k as it's a GREAT overclocker and will give you the best performance by far of any CPU currently out (at a reasonable price range) for PCSX2 and almost anything else.
[Image: 2748844.png]
I'm not an expert at this but I have done it a couple of times so I'll share what I've seen.

1. Undervolting will not decrease performance so long as your undervolt is stable.
2. Not so sure on this, my guess is mostly to achieve better cooling.
3. Most of the time yes, especially with "big" overclocks. It doesn't have to be true though, for example with my previous motherboard(P5N-E SLI, now broken) I was able to achieve an overclock of 3,6ghz(over3ghz) without raising voltage at all and it was perfectly stable. With my current mobo it raises voltage slightly for the same overclock.
4. Quite a noticeable performance increase. While not perfect, the performance should be close to the amount overclocked. So 3,2ghz to 4ghz should be near 25% performance increase. The Phenom I is a bad overclocker though so 4Ghz might not be achievable.
5. I'm not familiar with either of those coolers so I wont comment.
6. PC games? Yes, although a 3.2ghz phenom is more than any pc game "needs" so minor noticeable difference. PCSX2 games? Yes.

(05-26-2011, 06:32 PM)Koji Wrote: At default rates that i5 will likely beat the AMD. Your best bet would be to try to spend a little more and get the i5 2500k as it's a GREAT overclocker and will give you the best performance by far of any CPU currently out (at a reasonable price range) for PCSX2 and almost anything else.

QFT, if you are considering the i5-2400, you should do yourself a favor and get the i5-2500K even with a mild overclock it will outperform any other cpu on PCSX2(with the exception of the 2600K ofcource Tongue). And the price difference is very small.
CPU: C2D E8400 @ 3.6
GPU: GTX 560Ti 2Gb
MOB: Asus P5QL
RAM: Crucial 4Gb
OS: Windows 7 64bit/XP 32bit
Forgot to mention the other benefit of undervolting besides keeping it cooler (and thus increases longevity of the parts) which is power saving. You may not think too much about how much electricity your computer uses, but it can save a pretty penny over the long term.
[Image: 2748844.png]

  1. Well answered
  2. Answered, as well.
  3. Again, answered, but I'd like to add a couple examples (results may vary, of course). I took an AMD Athlon II X3 435 from 2.9Ghz to 3.6Ghz on stock Vcore. I took my 2500K from 3.3Ghz to 4.5Ghz without manually adding to the Vcore (it does a little on it's own, though).
  4. Simple answer: Look at the benchmarks. I submited a 955 BE @ 4.0Ghz. Wink
  5. It should be pretty good. I know it works great for me, but I'm sure a lot of that is thanks to the processor. I can't say directly about use with that CPU (X4 955 BE), but judging by what it does with the stock cooler (and good reviews of the CM Hyper 212+), I'm sure it would do decent, at the very least.
  6. With PCSX2, I would again refer to the benchmarks. Looks like a 2400 @ stock would likely still best a 955 BE @ 4.0Ghz.
Detailed & Simpler :
One rig would be noticably better with PCSX2. That 955 BE and my 2500K are both in similar setups, so I can tell you the difference is pretty clear there. How many FPS better? To be most specific, the benchmark would show about a 15fps (~25%) difference between a AMD X4 955 @ stock and my 2500K @ about the 2400's stock clockrate.
What defines the stability is the signal getting always in the well defined area where it can be granted to be 0 or 1, the manufacturer already use the practical minimal voltage where this can be granted. So, undervoltage is risky, albeit possible. Normally associated with clock speed decreasing, just never to the point the signal can't reach the expected high level threshold.

On the other hand, under limitation, overvoltage increases the chance the signal reach the high voltage level in time to be reliable when the clock speed increases. The most obvious consequence is the temperature gets much more a concern.

Less known is the impedance effects which are dependent on both the voltage And frequency (the capacitive effect is the main figure here) which can compromise stability on higher frequencies do not matter the temperature and actually worsen with voltage increasing. Still, probably the other problems will manifest early enough to prevent this last to be a real concern.

Another long time ago known issue is called "electron migration" and is told can undermine the chip structure, but although a real issue, research found it is a very slow phenomenon and should only trigger actual issue long after the chip is already obsolete and probably out of circulation long ago. Not surprising I don't see being mentioned very much nowadays.
Imagination is where we are truly real
Some basic physique Smile

consumption = k * f * V *V
k = constant
f = frequency
V = voltage.
So Voltage have a bigger impact on consumption.

More frequency => less time to move "electron" between state.
more voltage => increase speed of "electron"

So the bigger your voltage, bigger is your max frequency. Now lots of manufacturer sell overvolted chip to give some margin and potentially increase stability. Sometimes margin can be very big.
So the lesser your frequency, smaller is your mininum voltage. It explain why idle chip, change first frequency then voltage.

Now a chip manufacturer is capable to generate a nice 2 dimensions graphics (array actually). We call that a shmoo
Here an example

So for each piece you are able to tell the max frequency at every voltage.
Well resumed in physical terms, gregory. To complete both posts, just think a 3 layers pie, the bottom layer is the 0 zone, the central is the "gray area" and the topper is the 1 zone... if the signal is given not enough time to reach the "active" zones it falls in that gray area and... kaboom stability.
Imagination is where we are truly real
Forgot one things the temperature. When it is cold (but not too cold [0]) the electron can go faster -> imagine you need go through a horde of people. If people does not move (cold) you can avoid them easily. If people move in every direction as an incredible mess (hot), you will shock them often and slow down your progress.

[0] Actually if it is too cold (do not remember the figure, but let's say less than 50K) the semiconductor material loose is semi-conductor property... I do not know how they can use liquid He (which is ~4K) to cool down CPU. I think they need to first warm the cpu then apply helium so the internal temperature is not too low.
I noticed your using an AMD processor like myself ...if you didn't know AMD makes a software to help you with overclocking your CPU and run benchmark and stability tests called Overdrive it's really easy to use.

Also a good benchmark program I like to use that is completely free with no 30 day trial or ads is Maxon Cinebench it has a really good test for your CPU and your OpenGL ( graphics ) that I find superb and will compare your results with other CPU/Video cards stock performance.

And if your wondering exactly what you've overclocked your computer to alot of people on here use CPU-Z Validator to see what they've changed their performance too

I myself am still somewhat new on the overclocking experience so one thing I took to heart for suggestions when reading up on it is to small changes at a time and then run a good stability test for like an hour so you can see how your computer manages not just the new changes but also the heat changes as well. Overdrive allready comes with a stability test that you can have run on your computer for as little as a minute to more than a full day if you want.
AMD Phenom™ II X4 955 Processor (4 CPUs), ~3.2GHz, 8192MB RAM, ATI Radeon HD 5700 Series

Users browsing this thread: 1 Guest(s)