Arcmanov's rig evolves...

[Arcmanov]

Just finished a 1-hour test with LinX....




Interestingly...the max recorded temp on core0 is actually 10 C higher than what I achieved last year with the Silver Arrow.

It struck me as odd, because I expected the max temp to be the same or lower.  After comparing settings between then and now, I think I've discovered why...


Last year, I needed 1.440 Vcore to maintain stability at 4 GHz, and now, that same stability requires 1.512 Vcore.
Of course, more Vcore means that the CPU is generating significantly more heat at full load now.  Very interesting.

I can only speculate as to why this is happening.

Perhaps it is due to a steady 'degradation' of the motherboard circuitry supplying current to the CPU.
Maybe even a slight degradation of the CPU itself.  Who knows?  One thing is certain though,
this motherboard and CPU are real troopers. 

I'm not worried though.  I have confirmed stability, and 'normal' load temps are within acceptable parameters.


Enough testing...time to return to the Battlefield.

'Viper two-two is inbound to the AO...' 

[phoenix31tt]

Is the extra cost worth it for similar performance? Seeing as you have a haf-x space shouldn't be an issue right?

Or is it? I've never worked with air cooling in the haf-x so I dunno

[Arcmanov]

Is the extra cost worth it for similar performance? Seeing as you have a haf-x space shouldn't be an issue right?

Or is it? I've never worked with air cooling in the haf-x so I dunno

I think it's worth it.

Even with the size of the HAF-X, the TSA was so big that it was difficult to do anything around the CPU area without having to remove the thing entirely.

Also, one of the main reasons for me getting this is it's forward compatibility with LGA-2011.
I'm not sure if Thermalright would release anything to upgrade their present coolers to that socket.

Plus, I think it was way past time for me to move on from those big-azz air-coolers anyway.



@ Hard O/C...if you're referring to the LinX load temps, they're supposed to be high.
The purpose of LinX is to stress the cores way beyond normal, real-world usage to absolutely guarantee stability.

How many real-world apps would really have your CPU operating like that?  Zero.

...and yes, I removed whatever TIM that came with the block and applied AS5.

...and keeping HT on at 4GHz with a C0/C1-stepping Core i7 920 requires more voltage than I would like,
   so that comes on for benchmarking only. 

If I had a D0-stepping CPU, well that's a different story.

[captainchris]

The stock paste that comes with the corsair h coolers is known to be Shin Etsu, a VERY good TIM.

[shivanandrs]

Just finished a 1-hour test with LinX....

You have any idea what ambient room temp. was while running this test?

[phoenix31tt]

Yea I can are your point with trying to work around it, coming to think of it even if your case is big with an oversized cooler working with memory or even the 8pin power might be troublesome

Guess u should get slightly better ambient case temps ad well since no heat is being dispelled directly in the case?

[Arcmanov]

I started the test around 8:20 or so, but the sun had already made the room a little warm.  Can't say what the room temp was at that point though.

I'll run it again tonight and see.

Next month (or maybe December) I'll order four of the most murderous hi-rpm 120 mm fans for some serious push-pull.

Someone already mentioned the Silverstone Penetrator, so I'll have a look at those.

Swyped from another Galaxy with Tapatalk

[MessiaaH]

In my own liquid builds i've noticed with an efficient radiator, having powerful push/pull configuration makes no difference what so ever.
i juss using 3 low rpm high airflow scythe and thats it dey, i tried 6 fan push\pull, and faster fans, temps din budge.

So doh expect much from that fan upgrade. Mite be a waste of money juss adding more noise to ur rig.
But if d efficiency of d corsair rad is low, u mite scrape like 2 degrees.

[Arcmanov]

Well, you never know until you try.  Plus, the information I'm providing here will help other prospective buyers to make a decision one way or the other, no?

I just remembered that I may have some Antec Tri-cool fans laying about.  I'll sting out a few of the eight long screws from the existing fans and try a lil something later.

The saga continues..... 


Swyped from another Galaxy with Tapatalk

[MessiaaH]

Scene scene, let us know if they made a diff.
Sent from my Windowa Phone using Board Express

[MessiaaH]

Arc, wah clock is that u using in d top right dey.

[Arcmanov]

The clock is part of the HTC Home APIs and you can download it here...

...and you might have been right about the fan vs rad issue, but the ambient temp
was so much hotter this time (11am) that I can't be sure.

I did some more tests after I installed two Antec fans as pull...





Temps seemed the same or higher but my methods have been less than scientific. 

I'll have to re-run all these tests between 11 pm and 1 am to offer a definitive verdict.

What I will also have to toy with is mounting the fans as intake, and not exhaust.
Corsair's recommendation actually says that it should be mounted as intake.

We'll see how that goes. 

The saga continues...

[Arcmanov]

So I flipped all my top fans, and the ones on the rad to intake, and immediately noticed a 4 degree drop in max load temps... 






...and I erroneously stated that Corsair recommended using the fans as exhaust, but they actually recommended the opposite.

So there's your answer if you want this thing to perform at it's best.  Four fans in push-pull, as intake.


I think I'l leave it there for now.  My M249 is getting rusty. 

[captainchris]

hold up!!

so wat u saying is..... you pushing the hot cpu radiated air into the case..... de fudda munt?

[Arcmanov]

Yep, but I have nice exhaust fans, so the temp penalty within the case is negligible.

[Arcmanov]

I was browsing some search results regarding H100 installations and came upon this German website.  (You will need Google Chrome to see it in English)
That article is a great source of info with respect to case airflow.  They show you how to get the best airflow using 2, 3, 4, or multiple fans.

Using their recommendations, I changed my front fan, and side panel fan to exhaust, and wouldn't you know,
those two simple actions actually netted a 3-degree drop in overall CPU idle temp.  No longer does the temp
fluctuate between 36-38 C.  It's now at a rather respectable 35 C.

(I'll experiment with changing the rear 140 mm exhaust fan to intake as well to see if it will have any effect.)

I was also able to lop off an extra 2 C in average load temp as well...



My case may no longer collect dust as quickly as before now that there are no intakes close to the floor.  Undeniable win.

[captainchris]

nice nice!!! good find!!

will have this done soon!

+1

[W1nTry]

^ men eh see the +1 emoticon we have now?   

@ Arc, I read the article you posted... maybe i'm interpreting it incorrectly but blue is good and red is bad.... isn't 3 intake 2 exhaust at top best according to their pics?  Not following the side exhaust...

[captainchris]

i know you asked arc, but from what i was reading yes.. and i already knew that.. thing is....the fan i see doing a lot there would be a 120/140mm pulling from outside/underneath and pushing UP ( positioned at the bottom of the chassis )

[captainchris]

adding to that, the silverstone raven02 i had, followed that and push 3 180mm fans from the buttom to the top and the temps i got were EPIC, as you guys saw in my benchies thread.