Personally I first run Intel Burn Test on high, then Prime95 for a more lengthy period. I also run cinebench and encode a couple videos in handbrake to help test out stability for real world use.

 

I did also run some tests in AIDA and Intel XTU.

 

Is that your CPU multi on 42x you're talking about or your'e uncore/cache multi?

 

They are two different things and you want to push CPU multi higher first before increasing uncore/cache.

 

 

Sorry Noel but I'll have to disagree with you on that one. I don't think a tweak to the sim is going to net the same results as a 300Mhz increase in clock speed.

300Mhz over the 4200 baseline is what it is: exactly 7% 'sweeter' than baseline. In practical terms, this means when you land the NGX at KSEA in heavy weather in PNW scenery w/ slider hard right, all sliders, you're going to go from 15 frames to 16 frames per sec, neither of which is acceptable to most users. My real point was that one should think very carefully about whether or not the added system stress, energy cost, and earlier demise of your CPU is worth that final 7%. The other part of my point was that it is easy to see improvements in frame rate of 7% by adjusting a few sliders to the left, and that in the double-blind test neither you nor I would notice the change in visuals because once again, we're talking 7%. No magic there I'm afraid. 90C? Ouch! I also oppose running synthetic benchmarks to insure you've got stability. FSX does not stress your CPU and memory subsystem nearly as heavily as Prime95 or other does. This practice, using synthetics, ends up causing users to use voltages that may not be needed to achieve the same outcome as using FSX itself since temps are considerably lower in FSX than in synthetic benchmarks.  My approach is to use the synthetic just for 15 minutes or so, then migrate over to FSX and lower voltage a notch and test.   I first heard this argument from JJ at ASUS and it makes perfect sense when you think about it.

I'd prefer running a higher clockspeed, and running those synthetic stress tests allows me ensure a high level of system stability. Hell if I didn't run these synthetics and care about my overall system stability I'd be running this CPU at 4.8Ghz right now.

 

But just remember, not everyone uses their PC's for just FSX. Sure I could probably run FSX at a lower voltage, or at a higher clock speed on the same voltage, but I may not be stable in other games/programs for example.

 

90C during stress testing on a non-delidded Haswell is no big deal and is considered pretty normal. Now mine didn't quite run that hot (about 85C) but in everyday use (including FSX) I idle around 27C and max out about 60C, no doubt the OP would have a similar experience during every day usage.

 

The increase of FPS in FSX may be small, but it's the overall smoothness and fluidity that truly makes the difference.

 

Each to their own.

 

Noel, let's just agree to disagree on this one I think. I hope the OP enjoys his new PC!

The increase of FPS in FSX may be small, but it's the overall smoothness and fluidity that truly makes the difference.

It's ~7% better 'smoothness' and ~7% better frame rate, no matter how you frame the value of it to you personally.  In the realm of statistical significance, it meets the mark.  However, and this was my main point, this kind of improvement is dwarfed by, for example, moving the mouse pointer off the VC during the 5 sec period it remains visible and parking it on the top menu bar.  In the PMDG NGX & T7 this is turning out to be good for somewhere around a 15-20% improvement during the period the pointer is visible.  Unchecking the High Def VC cockpit checkbox in FSX aircraft graphics setting is good for around 10-15% improvement and is arguably barely perceptible visually, surprisingly.  This is my point to people who are embarking on a new overclock--to think about that last few % since the yield is measurably poor, and the increase in power consumption and most assured heightened risk for an early CPU demise is also measurable.  Instead, don't forget to look deep into minor configuration changes that generate relatively big improvements in total performance.

 

It's OK ant, you're free to express how you see it, and so am I.

Oh Noel :huh:

 

Lets throw those percentages out the window for a moment and forget about them.

 

For me personally, I can see the difference in FSX between 4.2Ghz and 4.5Ghz, perhaps you don't but I do.

 

So should I say to myself or anyone else "Oh increasing your clockspeed by 7% will give you better FSX performance, but it's only 7% so don't worry about it". No, I don't think so, because I want the best performance I can get.

 

If you can run a higher clock speed, great, do it! If you're not comfortable with it, don't. End of story.

 

Let's leave it at that.

 

 

Lets throw those percentages out the window for a moment and forget about them.
 
For me personally, I can see the difference in FSX between 4.2Ghz and 4.5Ghz, perhaps you don't but I do.

 

I'm afraid the percentages matter and can't be thrown out Anthony.  Further, in a double-blinded test, w/ my making a few discreet changes in your fsx.cfg that you would not perceive were made, I don't think you could see the difference.  In fact, I'll go one step further:  I'm guessing we could fool you into choosing the 4.2 clock over the 4.5 clock more than 50% of the time ;o)

 

It is what it is my friend, ~7%. 

It might be that way in your opinion, not mine. I can see the difference, and a 7% improvement is still a 7% improvement. I'll take that 7% improvement any day!

 

Just leave it at that Noel and agree to disagree.

 

Let's move on to something else like enjoying FSX for example. I'm currently debating getting the new DX10 Fixer so am heading over to that thread to do some research.

 

I'll bid you adieu at that. :wink:

We're in agreement--a 7% improvement is a significant improvement.  That, however, was only a small part of the discussion.

 

Cheers  :smile:

We're in agreement--a 7% improvement is a significant improvement.  That, however, was only a small part of the discussion.  Most of my temptation to jump in to these discussions comes when I see others being coached by someone who proposes the sort of comment that implies the Holy Grail starts at 4.5 or whatever, which is complete nonsense and IMO stems primarily from the fact this is the highest overclock the particular party was capable of.  We see it all the time.  I'm very familiar w/ how FSX scales w/ clock speed and I'm afraid there's no magic involved.  Indeed, 7% is 'better', 7% 'better', and this is the other reason I enjoy jumping in especially when folks newer to the hobby may be steered into subjecting their sometimes pricey CPUs to volts/temps that will undermine the chip's longevity.  By now most folks understand this so I'm not accusing you of this Anthony, but the mindset of 'sweet spots', as I clearly stated in the very first reply, is 'your' sweet spot, not 'the' sweet spot.

 

Cheers Anthony--we also agree to enjoy FSX for what it is.  Adieu!

You seem to be being like a dog with a bone on this one Noel. lol 

 

If your talking about not being able to detect the difference between 4.2Ghz with a tweaked fsx.cfg and 4.5Ghz with lesser tweaked config, I'm really sure where you're going there.

 

All things equal (which is what I'm talking about), including fsx.cfg being tweaked in the exact same way, same settings/fixes etc, 4.5Ghz is always going to work out faster/smoother than 4.2Ghz on that given chip.

 

In your words, "it is what is is". And it's faster!

 

How much a user will notice it will obviously vary, but I can tell you 100%, I do notice it, and if I am tweaking my system I always run the same uniform tests so I can measure the improvement.

 

In some situations will FPS still drop dramatically? Sure, but I'd rather know I've got my system running to it's full potential to minimise that as much as possible.

 

Whether a user is comfortable pushing a higher overclock is completely up to them and it is their responsibility to make sure their voltage & temps are acceptable and not excessive. I could push my chip further, and have done so, but decided 4.5Ghz was where I was happy to settle given we're expecting a very hot summer here in Australia.

 

I was merely trying to point out that 4.5Ghz on a 4770K is very achievable for most chips well within the thermal and voltage guidelines, hence being the "sweet spot" for a 4770K.

 

Adieu  :Peace: