CPU Overclock/System Properties Question

Ex_Brit

Moderator
Staff member
I know there are loads of O/C forums out there but I would like sane answers, not the advice of some whiz-kid of 12 in Borneo or wherever.

My Alienware machine came overclocked from the factory 3.33 to 3.86ghz and I've found that it seems comfortable if I up the BIOS CPU Multiplier by 1, to 30, which gives me 4.00ghz exactly, so I figure why not have the extra speed? No overheating issues thus far and I have water cooling anyway.

Purely for vanity reasons I would like the system properties to reflect that, however it just shows the default factory CPU speed of 3.33ghz.

I read online that disabling C1E and Speedstep support in the BIOS would solve that. I tried that but system properties merely said "3.33ghz 3.34ghz", weird - meanwhile all other tests CPU-Z, Intel Processor ID Utility and ASUS Turbo-V are all showing 4.00 ghz.
So I enabled those settings again.

Question, why does Vista and Windows 7 sysprop not reflect the true speed of the processor?

Question 2, with my CPU core temperatures hovering between 65 - 85°C under load (8 grid computing applications running constantly) am I OK or am I playing with fire, literally?
 
Last edited:
Hi Peter,

System Properties doesn't show the actual speed, it just shows the speed of the particular model of processor that you have installed.

85* is a bit hot - under load it shouldn't go above 80. However, since you have an Intel CPU it *isn't* dangerous - they all have built-in shutdown sensors. If it gets to hot, the PC will always turn off instead of burning. BUT, they also have intelligent clocking - if it gets too hot, it'll also slow down ("frequency scaling").

You really should look into why it's at 85* C. Going from 3.33 to 4.00 (esp. since you're actually going from 3.86 to 4.00) with water cooling probably shouldn't reach that high.

Perhaps you're looking at something other than the core temp?
 
My CPU is under 30°C most of the time (the case fan only switches on at 30°), unless it's doing very intensive graphics (or MSE is stuck in a loop !), in which case it reaches towards 40°.
That's air-cooled and sans O/C, and with "cool and quiet" installed.
 
Well I was using Everest's CPU average rather than each core. If I do each core then it's hovering between 75 and 100 right now.

The thing is if I return it to 3.86 I don't see much difference. With the Speedstep on in the BIOS the CPU's automatically throttle back in any case depending on load.
Not sure what C1E does actually.

Feeling the water pump assembly and hose it doesn't feel that hot at all so I'm wondering if Everest errs on the plus side.

Maybe I'll use another utility. I'll look on the ASUS site for their own utility, probably more reliable as it's their motherboard.
 
Definitely Celsius Terry. I've now closed Everest and am using ASUS PC Probe II which says my CPU is a steady 83°C, Motherboard 54, voltages, Vcore 1.4v, +3.3 = 3.22v, +5 = 5.09v and +12 = 12.02v, CPU fan 1360RPM and Chassis 2 fan 1430RPM.

It seems to depend on whose opinion one listens to as I've read of people running CPU's at 100+ and they don't seem to worry, although I would have thought that would shorten the life of same somewhat.

I also have found some wildly varying figures for what the normal operating tempratures are for these components.

In other news I see that you guys are getting a minority government.

If I turn off my grid-computing which normally uses all the cores, all of the time, the temperature only drops to around 65°C anyway

Talking to myself here...LOL. Out of curiosity I decided to check for dust and vacuumed out the vents, the fans that I could reach easily and the water-cooler radiator, again what I could reach and now the average temperature for the CPU is hovering around 65-75°C - that's with all those apps going. An improvement.

I also just moved the case a couple of inches further away from the wall behind it and that seems to have lowered things even more by giving the various fans more room to vent. A further improvement.
 

Attachments

  • Capture3.JPG
    Capture3.JPG
    23.5 KB · Views: 2
  • Capture4.JPG
    Capture4.JPG
    71.6 KB · Views: 5
Last edited:
75 is decent.... Terry and I have both had issues with cat hair clogging up fans and causing problems, so we know both know what you mean :grinning:
 
Cat hair.....LOL That reminds me of a friend of mine. We had the side of his machine open and the cat disappeared. Only when I went to replace the side panel did we realise where the cat was...right inside the computer, thank God she didn't bite, scratch, pee on or otherwise meddle with anything. (Small cat by the way.....they say curiosity killed the cat, well it would have been true if noone had noticed...!!)

They love warm spots I guess.

In my case dust is something I'm ignoring more and more as I get older, I hate to admit. I still clean, but not daily like I used to. I also have my tower up on top of a mini PC desk next to my regular one so it's way off the floor which helps.

Thanks guys.
 
Last edited:
Did a bit more research and have lowered my CPU voltage to 1.35 and the temps immediately dropped 15°C average right through the scale from no load to max load with the CPU multiplier still at 30 for an O/C of 4.0ghz.
That's strange in that it was Alienware Support themselves that told me that in order to overclock I had to raise the CPU voltage from the default of 1.2v (for 3.33ghz) to 1.4v (for their original overclock of 3.86ghz).
From further reading I could easily raise this to 4.2 without further voltage adjustment, but I don't want to push my luck.
For them to make that kind of mistake on something we can do and then insist that flashing ones BIOS without them overseeing said action invalidates the warranty is bizarre to say the least.
 
Last edited:
Hey I was a bit busy a few minutes ago, didn't get to elaborate (had friends over).

A lot of problems may occur when OCing at lower voltages that don't appear at first. Make sure to run memtest extensively to ensure system stability - it'll often crash under load ("crash" doesn't mean system hang, it means incorrect read/write results which *may* under certain circumstances lead to a system crash).

Give me a minute and I'll post the name of a good (as in, quick to determine if your system is sensitive/fragile due to OC) tool as soon as I get on my other PC.
 
Here you are. It's called OCCT and I can't remember where I got it from.
 

Attachments

  • OCCTPT.rar
    1.7 MB · Views: 3
OK, I'll give it a go and see what happens. Have already stopped all unnecessary processes.
Thanks Mahmoud.
 
Well, I started it and it squawked after about 5 minutes and stopped. It seems that one has to manually set the maximum temperature. That, being an arbitrary amount, makes this test like all the others, subject to exactly how much I dare to let the computer fry before it burns to death.

I set it to 100°C and even that wasn't enough. I'm not setting it any higher than that. The CPU is bound to overheat if it is subjected to 100% load permanently, which I was monitoring and it was, despite the top of the UI saying CPU load 49.2%, or similar. I would assume such software overrides the Intel Speedstep ("On" in the BIOS) which automatically slows the CPU down when it detects danger, as if it was on that couldn't happen.

Under the heavy usage I give it, even with 8 BOINC projects running simultaneously in the background 100% of the time, it never reaches anywhere near that temperature now.
I'm a little dubious about letting something like this run if it's sole aim is to overload and see what a processor can stand.

I remember running MEMTEST on my old P4 HT which wasn't O/C'd in any way and it started beeping so I had to shut it down to cool off.

I don't understand these applications perhaps, but the point I'm trying to make is these sorts of tests obviously are aimed at overloading the CPU etc. and are bound at some point to cause failure unless the machine is kept in a refrigerated environment.

Maybe I'm not understanding the settings. I tried it with no other monitoring software and then I tried it linked to Everest Pro which I have (the ASUS utility isn't listed). Both same results.

If I'm way off the mark here please let me know!!!! Meanwhile I appreciate your help very much.
 
As an experiment I just set the BIOS to default. That should be 3.33ghz, however for some unknown reason it has set it at 3.48ghz (always does) & CPU voltage at 1.225 instead of 1.2. Alienware is really stranger than fiction sometimes.

Temps are running in the 60's max. with BOINC running full tilt.

I'll leave it at that for a while and see if the speed difference is noticeable.

By the way, BIOS default means my second DVD burner becomes my main one...can you believe Alienware wired them in reverse? So I always have to reverse their order in the BIOS - just to have things behave the way I like them to behave. With the way they use minimal wiring and hide it all away out if sight it's impossible to rewire without ripping things apart. You can't even swap the plugs around as there is no free cable to play with.
 
Last edited:
That's weird - maybe I gave you the wrong tool.

It's not supposed to monitor temp (i don't think it even shows it). It's supposed to run long and difficult math equations and compare the results against saved answers. If your PC overheats/overloads and starts finding wrong results, it'll squeak and squeal.

Addendum:

I'm on my Mac right now, as soon as I get on my PC I'll double-check.
 
Last edited:
Back
Top