Need a little Core i7 980X overclocking advice

[miahallen]

A whole 10 minutes of P95 huh? You might need a bit more than that. Even 10 minutes of LinX isn't enough.

I've had 24hr prime stable CPUs that still were not stable for my daily usage. "Stress testing" is highly overrated IMHO.

Anyhow, my 980X is only used for benching anyhow, so the only stability I need is enough to pass whatever bench I'm working on.....for example, how about 5.9GHz in 3DMark 05:

I was only testing Prime95 for a heatsink comparison I was working on.

 

[Alex@Nosound]

So you want 5.9Ghz for daily usage?

 

[miahallen]

So you want 5.9Ghz for daily usage?

my 980X is only used for benching anyhow

My i7 920 is fine for my daily use

 

[AdamK47]

Welp, I decided to replace the motherboard with an Asus P6X58D, get another 5870, and get the PCI-E version of the Xonar Essence (replaced parts soon to be sold on this forum).

The P6X58D reminds me of how much I've missed having an Asus motherboard. The layout is much better than the EVGA X58 SLI. For some reason EVGA put the 1x PCI-E slot right below the primary PCI-E 16x slot which would make it impossible for me to use an add in sound card. Had there been a way for me to use a Xonar Essence on the EVGA board with 3x CrossFireX, I would still be using it. The Asus board does have USB 3.0 and SATA 6Gb. Don't have anything that would use it, but it is nice to have it there just in case. One very nice thing the Asus board does better than the EVGA board is VDroop elimination. When I set the Asus board to "Load Line Calibration" there is absolutly zero fluctuation in CPU voltage. With the EVGA board, it would actually overvolt by .025V when load is applied using the "Without VDroop" option.

Oh yeah, pictures...

 

[AdamK47]

I've had 24hr prime stable CPUs that still were not stable for my daily usage. "Stress testing" is highly overrated IMHO.

Anyhow, my 980X is only used for benching anyhow, so the only stability I need is enough to pass whatever bench I'm working on.....for example, how about 5.9GHz in 3DMark 05:

I was only testing Prime95 for a heatsink comparison I was working on.

All that hardware just to benchmark reminds me of SPL competitions. All that car audio equipment just to see who can get the loudest. If it's a boat floater, go with it.

 

[Alex@Nosound]

My i7 920 is fine for my daily use

Ahh i see whats that under?

SS,dice,ect ?

 

[AdamK47]

The Asus P6X58D Premium seems to play with the 980X nicer compared to the EVGA X58 SLI. Currently running at 4.2GHz with 1.35V which is 1.336V according to CPU-Z with load line calibration. Up to 1 and 1/2 hour into LinX so far.

 

[Shmee]

grats! Go for 4.4! should be possible with that chip. Looks like it is better than mine, or at least it plays nicer with your asus board.

 

[AdamK47]

I received an error in LinX after 3 and 1/2 hours. It was on the very edge of stable. A slight bump in voltage and I know it could be 100% stable. The temps were kinda high though. Most of the time in the low 80s / upper 70s and sometimes spiking at 90. I'm now trying an even 4GHz at only 1.275V. I'll see how that goes stability and temperature wise.

One thing that has me worried is the amount of heat coming off of my three 5870s in CrossFireX. I use to run two of them at 950/1250 with 1.225 on the core. With three of them this is pretty much impossible since the primary card will start thottling after 15 minutes running something intensive. At 900/1250 with stock voltage everything is fine with zero throttling. The three cards in CrossFireX make up what seems like a giant brick of heat when running at 100% in my system. It's crazy how much heat is coming off these things. I have since put a 120mm fan along the side of all three of them to push air into the slim spaces between the cards and undervolted them slightly to 1.15V (1.175 is the normal voltage). This dropped temps further by 5 to 10 degrees and the fans on the three cards havent spun up past 50% when all are at 100% load. This has me wondering though... are there better CrossFireX friendly aftermarket coolers I can slap on these things?

 

[Spike99]

When I ran LinX, I did not get any errors with vcore under 1.35v. When running Prime95 that's when I started seeing errors and had to increase vcore. I would run Prime95 as a final check.

I only have 1 ATI 5870 and I've been waiting on Accelero EXTREME 5870 vga cooler from Arctic Cooling. I'm looking into this cooler only to reduce fan noise.... but my room does get warm when gaming. In any case, you probably going to have same heat even getting a gpu cooler with 3 5870's. The heat gotta go somewhere

 

[Shmee]

for three 5870's with those OC's you kinda want some water blocks

 

[AdamK47]

Just an FYI... the Megahalems performs better in a bottom to top mounting configuration in an Antec 1200. It allows a 120mm fan to be used on the side of the case. This with a single 120mm fan sucking the hot air off of the heatsink and into the 200mm top fan has lowered temps by 5 degrees.

 

[SHAQ]

Just an FYI... the Megahalems performs better in a bottom to top mounting configuration in an Antec 1200. It allows a 120mm fan to be used on the side of the case. This with a single 120mm fan sucking the hot air off of the heatsink and into the 200mm top fan has lowered temps by 5 degrees.

Why don't you have a 120mm pulling also? And for a 980X you should have sprung for the Mega Shadow. I just got one and it is sweet with the blue CM R4's. My idle is down to 38C from 43C before with a Ultima 90. I'll run some load temps in a bit.

 

[miahallen]

All that hardware just to benchmark reminds me of SPL competitions. All that car audio equipment just to see who can get the loudest. If it's a boat floater, go with it.

LOL....I used to do that too

Ahh i see whats that under?

SS,dice,ect ?

air

 

[Spike99]

Originally Posted by AdamK47
The Cooler Master V10 was getting into the 80s at full load with the 980X at 4133MHz with 1.375V. The stock Intel cooler was running it in the mid 70s. The Magahalems is now running it in the mid 60s.
.

That's pretty impressive temp with the Megahalem. I'm itching to get that... but i really can't complain about with the H50 even if I get 82C during load.
.

Recently, I checked a review about CoolIT ECO ALC at http://www.legitreviews.com/article/1292/6/. The ECO ALC is very similar to H50. In this review, they included the H50 and other air coolers... unfortunately they did not include the Megahalems. In a push/pull configuration the H50 basically beat most of them and since I have a push/pull configuration I wondered why I was getting load tepm of 82C with my setup. I did some additional tests and basically, due to noise I have low noise adapter for my Noctua fans that basically lowers fan RPM. After removing adapters and have fans (NF-P14) run their normal RPM (1300RPM) I got much better temps. Idle temp is now in the low 30's & load temps high 60's but some spikes to 72C was noticed. Although I still believe the Megahalem is better than the H50, the H50 is a very good cpu cooler. After testing, the noise was a bit too much so re-attach the low noise adapters. During normal use, the highest temp that I have noticed is in the 50's so it's okay...