Q6700 stress test

Status
Not open for further replies.

LinkedKube

Posts: 3,385   +55
Trying to run a stress test to see if my 3.5Ghz overclock is stable 100%. Prime95 is not using all 4 cores at 100% or even close, should I take this as a good thing as in my overclock is fine and prime couldnt even crash it or should I be worried.
 
Not good, you want it at 100%, cos you're looking at stability, try running 2 stress programs, its the same thing.
 
It hasn't exactly been determined how to stress test a quad-core processor.
Running four instances of Prime95 will work. Not sure about running two instances of Orthos though..

Here's what you do with Prime95:

I used these exact instructions with stress testing my dual-core, and it works. I just modified the steps a bit for four cores rather than two. Hope it works!



1) Install Prime95

2) Create four shortcuts of the installed Prime95

3) You need to change the shortcuts so they run as different instances. Right-click one of the shortcuts, go to "Properties", then to the "Shortcut" tab. At the end of the target line put a blank space, and add -A1 (so it’s “ –A1”)

4) Repeat step three for two other shortcuts you created (but change the extra characters you put at the end). You should now have one shortcut that is not modified, and three other ones modified differently (A1, A2, A3).

5) Open all four instances (the four shortcuts).

6) On the top toolbar, go to Advanced > Affinity, and uncheck "Let program run on any CPU". Leave the 0 in the box for the first one, and put a "1" in the box for the second instance.

7) Repeat step six for the other two instances. You should now have four instances of Prime95 open, each with a different affinity (0, 1, 2, 3).

8) Go to Options> Torture Test > Custom > Memory to use in (MB). This should be 1/4 of your AVAILABLE system memory for each instance.

As an example, I have 1GB of RAM, a dual-core CPU, and Windows XP that uses 256MB of RAM while idle. This means that I have 768MB of available RAM. Now, I divide 768MB up in half for my two cores, so I set each instance to use 384MB of RAM.

As for your situation, just take about 256MB of RAM from your total memory, then divide that into four. Input that amount into each of your instances of Prime95.

9) Click OK and let it run!

Stress tests should be left for a burn-in period to test stability, preferably 12-24 hours. You should also run memtest afterward to test the RAM (I assume it's also overclocked). Then, you can run a few benchmarks to see how much of a performance gain you have.


Just ask for clarification if something is not detailed enough.
I sincerely hope this helps! :D
 
Ok great, ran multiple Prime95's and got 100% on all cores, took 4 seperate programs though sheesh, I'm wondering how much power I'm draining this thing is beastly
 
Wow... I want a quad-core... I'm only going to finally have a dual core to keep for good at my birthday... -_-
 
my idle temp is 18C and full load is about 32C. Energy isnt that bad, I have 2 roomates, they are apple fan boys so of course they dont pull near as much as I do, but we split the bills equally.
 
Well, the Core2Duos don't pull that much power... per core :D

When you've got 4..... I think that the power efficiency of the Core2 just negated itself :D

And if you've sharing bills with 2 roommates.... Whats stopping you from plugging in one of those new Thermaltake 1.2kW PSU, and powering up 2x4core Xeons, and quad 8800gtx?





Money.
 
How do you manage to keep that quad core beast @ 32C fully loaded? Using watercooling or what?
 
He's using TEC watercooling, as in his sig (freezone). Given that cooler, 32C is pretty high, as its possible to go below freezing with that cooler (especially on idle).
 
believe it or not this cooler is somewhat effected by ambient temps, not much, probably only about 10% or so, but I'm only running it on low power :)
 
all coolers will be affected by ambient temps. This is because, no matter what kind of cooling you use (except LN2), air is still the way heat is transferred our of your computer. You'll find that the closer your computer is to ambient temp, the bigger effect ambient temp will have on your temps. Elementary Physics.

Also, for those on water cooling setups, it IS still affected by air.... what cools your water? Try running your watercooling setups WITHOUT your radiators.... Running your WCS into your outdoor swimming pool doesn't count, it still increases your total surface area....
 
all coolers will be affected by ambient temps. This is because, no matter what kind of cooling you use (except LN2), air is still the way heat is transferred our of your computer. You'll find that the closer your computer is to ambient temp, the bigger effect ambient temp will have on your temps. Elementary Physics. /QUOTE/

lol you again technically reworded my thoughts. I ran my quad on high the first day and got -5C with it at stock voltages. Since my vga's are on fans and I'm constantly playing...something, I feared condensation as the cold and hot temps met, maybe a low and high pressure area, creating a cataclismic reaction, like a tornado, just kidding, but you get what I'm saying.
 
You can get condensation, which was why even though I could have easily used TEC cooling, after months researching on the subject, I decided that its not worth it.

However, I was thinking of much more powerful TEC coolers, I didn't think that getting 32C was worth using TECs... If I'd have built it, I reckon it'll permanently stay below freezing.

Ever thought of cascaded TECs? :D
 
no, not at all, that's a bit much for me. on a zalmans my quad ran at 52C on low, that was at stock voltages. overclocked to 3.5 with this at 32C in quite formitable I believe.
 
I am having trouble getting your Freezone specs, but from what I see, its not impressive. It has a power draw of up to 56W, and that includes the fans, the pump and whatnot. In fact, it connects straight to your power supply :D

What I had in mind was at least a 320W TEC (this is at 15V though, which means it will be 256W at 12V). I would have needed a seperate power supply to power this mother, and it must be able to provide at least 21.33A on its 12V rail. Due to safety reasons I probably would have pushed for at least 24A on the 12V rail. My current PSU pushes 21A on each of the 12V rails max.

I don't have to tell you what kind of problems I ran into when trying to configure a cooling system with this thing, which was why after a few months, I gave up on the idea.

I'm sure others have done it, but most of the ones I've seen aren't safe, and didn't take into account the biggest drawback about TECs regarding temperature control. TECs cannot be controlled via normal thermostat, and if you do, the TEC will burn out. There are TEC control kits, but they cost a bomb, which is why most people don't bother with them.

Your retail kit should have everything considered, plus its such a low powered TEC, you might have been better off with a good WC system.
 
not many watercooling kits are internal, for the power management this thing cools great. I'm not going to spend 900 for one of the top end coolers. For the money this thing does the best job. Many benchmarks show that this is the best of what you can get at 300 usd and many competition products are way louder. If you want to see impressive go to the galory and check out my pics, i took pics of this thing first day, with speed fan at stock voltages. and that was running the thing on low. Watercooling is also a damn hasel, I dont have to go into that I'm sure you know. My set up is all internal and maintainence free. So I'll take 32C running full load over anything else. Do you know how hard it is to get a quad core to run at full load, at my clock rate, couldnt do it on fans and keep temps as low as I have.
 
I don't mean to say that its a crappy cooler, 32C is pretty good, and I've mentioned it before.

Its just that TECs have the potential to go much much lower than that.

I'm also saying that watercooling with the right components may bring temps close to that at load, due to how TECs work. Its a heat pump. If your CPU is pumping out more than 56W, which it definately should, then your TEC will not be very effective, which is why its cooling the water, instead of the processor itself, and that is also why it doesn't drop below ambient temp at load, not sure if it would at idle (depends on power output at idle).

It would be interesting to see the Swiftech H20-APEX Ultra Plus Liquid Cooling Kit and your freezone cooler go head to head, especially since the Swiftech comes with chipset and graphics card cooling as well.


BTW, top end watercooling kits don't come at 900 bux.

And your setup shouldn't be maintenance free. It would need the same maintenance as any other watercooling kit. Rubber/PVC tubes are porous, and you'll find your watercooling kit devoid of water sooner or later if you don't top it up.
 
I have the option to add quad ali coolers as well, not sure how well it will effect my overall cooling though. those are my worries. Soon I'm introducing sound dampening material, as a housing around the tec cooler in the case to lower dbs. Those are my plans, I'll post pics next week after i put it all together.
 
Just a short mention:

Here's some temp results for a TEC aircooler, which obviously has trouble cooling an overclocked E6400. Notice the load temps are just 1C cooler than a normal aircooler (its a "medium sized" aircooler as well, with a 90mm fan).

The TECs in your watercooling system should be of the same size (they push the same amount of Watts).

Also, you might find this REALLY interesting. FYI, the Ultra-X is an aircooler, with 4 heatpipes.

Edit: whoops, forgot to add the link.
 
Status
Not open for further replies.
Back