Core i7/X58 Overclocking Thread

Quote:

---

Originally Posted by dejanh

I was thinking of picking up Swiftech H20-220 Ultima Plus WC kit to cool this better but since I am new to WC, I am not really sure this thing is worth the money...thoughts?

---

danamics new em cooler at price similiar to watercooling might be an option.
the review isnt out yet but the cooler seems as a choice for those going watercooling.

Quote:

---

Originally Posted by Ghigo

I got same problem on vista 64 and know of other 2 with same i guess is jsut buggy software, you got an ati card ?

P.S. I see you got a 4870-512 like me :)

If can, can give a try with the 280 gtx ?

---

Well that is good and bad news considering I traded my 4870 for a GTX 260 Well live and learn ATI Strikes again.. Shoot! Thanks for letting me know I could have banging my head against the wall for awhile.... :shrug:

I no think is ati problem but more crap software from gigabyte, if you look to older easy tune versions it was a bug festival and really few people able to use it.

Gotta say, I'm loving my Core i7 / ASUS P6T / Noctua NH-U12P 1366SE combination.
I'm running 3.6 Ghz at stock voltages, and I have all 6 case fans wired up to the motherboard to automatically adjust fanspeed (speedfan)
EASIEST OVERCLOCK EVER.
http://www.exoid.com/R1F/3600.jpg

Wish I could figure out why my pi1m time is fine but my 06 CPU score blows. :(

Holy crap. I figured it out. Setting msconfig to "4" cores under boot had my score down to 3k. Leaving it to default so windows handles it caused the score to jump up where it ought to be.

i have come across some interesting results concerning HT while OCing on my setup...

@ 4.2 with HT enabled
http://i37.photobucket.com/albums/e8...der/0642ht.jpg



@ 4.2 with HT DISABLED
http://i37.photobucket.com/albums/e8...r/0642noht.jpg



the jusp in overall score isn't that big, but if you just look at the 3D numbers, you'll see a pretty big diff. The overall diff being smaller is only accounted for the decrease in CPU score when HT is disabled.

As an aside, while watching the benchmark roll, there were times when I noticed a 30fps increase with HT disabled...

seems these chips have a great deal of untapped gaming potential to be had by simply disabling HT

its noted for gaming HT off is the best option.

Good find sonofander, let's see if the result is repeatable by a few people on different hardware to rule out any anomalies :)

Quote:

---

Originally Posted by T_Flight

hey folks,
I have the CAS8 1600 Dominators. If you have this Kit and the RIIE see my thread about rhe BIOS in the Xtreme Bandwith section. It won't run stable on 0403 original BIOS. I flashed mine to 0602 and was able to run 6400 QPI, DDR1600 on the DRAM Freq, 1.64xxx (keyed in 1.64) and mnaually adjusted timings to 8-8-8-24. I have not messed with any other OC settings yet becasue I was fighting a Video issue the last 3 days. It turns out my card was defective and needs to be RMA'd.

Those settings seem much more stable with the new BIOS. Before my RIIE would not boot after setting the above advertised specs into the BIOS. It would only run at BIOS deafaults of 1333. Now it runs smoothly, stable, and as advertised. I highly recommend flashing your BIOS for complete memory support if you are using the RIIE.

I also measured and verified DRAM voltage through ProbeIt and got a 1.639v actual reading on the DMM. These are safe advertised settings. The 6400 QPI is advertised for the Extreme i7 965.

Just passing on my findings on this particular RAM, and the BIOS issue I found. Thought it would help. The Dominators are really good modules. I can't wait to tweak them and start OC'ing when I get my Video Card RMA taken care of.

---

I have basically the same setup as you, just using Core i7 940 instead of 965. Same Dominator sticks. I was able to run them at 1800MHz 7-7-7-20 stable with 1.65V and 1.437V QPI/DRAM. However, I have also noticed that upping the core speed will affect your ability to run the ram so high. I am still investigating this issue...

I am using the newest 0705 BIOS btw.

Quote:

---

Originally Posted by flopper

danamics new em cooler at price similiar to watercooling might be an option.
the review isnt out yet but the cooler seems as a choice for those going watercooling.

---

Unfortunately, that one is not licensed for sale in North America, and who knows when it will become available. I wanted it as well, even contacted the manufacturer but they said no dice. Maybe I will pick it up in Europe when I go this coming summer depending on what the reviews say.

Oh, and in terms of the watercooling components, I looked at the reviews for the components that come as part of that kit I mentioned, and they are very decent. Most of those components are listed even on this site as some of the best to buy. The only two things I would like different than in the kit is a bigger rad (3 vs. 2) and maybe a voltage adjustable pump. Building the same kit myself is stupendously expensive considering I can buy the whole kit for $290.

Is Loadline Calibration Ok for Core i7??? Also the reason why HT enabled lowers scores is the APP is not multi threaded or optimized, compare with a benchmark that is great at multi threading, like Lost Planet.

well then it seems very important to have a good idea of what you want to use your machine for when OCing...

I can stabalize 4.2 @ 1.45v and really good temps with HT disabled. If I turn HT back on, I have to raise my vcore to 1.48 to keep prime stable and my temps shoot through the roof even on really good water.
Right now I'm working on getting 4.3 stable w/out HT and I'm already working with 1.475vcore. I'd hate to see what volts it would take to maintain that clock speed w/ HT enabled


anybody running the RIIE and have luck stabalizing 4.3 or higher? I'm feeling a bit stuck... it seems that if I raise my QPI voltage to 1.45v real, it stabalizes my CPU on small ffts, but the higher QPIv makes the rest of my system unstable on the large ffts. If I drop my QPIv my system will prime large ffts all day, but the small one BSOD my machine :shrug:

I have dialed in 4.25 though :D


http://i37.photobucket.com/albums/e8.../4250prime.jpg



I know it's only 45min, but acceptable for me to try getting to the next level...

Quote:

---

Originally Posted by unclewebb

All Core based CPUs are designed to start throttling just before they hit TJMax. Usually when Distance to TJMax is 2. RealTemp will show this in the Thermal Status area at the bottom. When it says OK that means there have been no thermal throttling episodes since you started up your computer. If it says LOG that means there has been at least one thermal throttling incident logged since you last booted up and if it says HOT that means your processor is presently throttling. Your CPU will automatically drop the multi and total MHz only long enough to get your temperatures under control and then will return the multi to its previous value. It will continue to cycle the multi up and down to control heat and will let the CPU basically sit at its maximum temperature.

Here's what an E8400 looks like after running Prime Small FFTs for 3 hours with the CPU fan turned off.
Only one core hit 100C and that only happened once even though the CPU fan was turned off for 3 hours.

http://img204.imageshack.us/img204/276/hote8400fw5.png

The Core i7 is designed to do the exact same thing. As long as your overclock isn't maxed out, they can both run quite reliably at some very high temperatures. So far the Core i7 looks like it has more temperature headroom compared to Core 2 Duo.

---

Will the multi still drop with hpet enabled and EIST disabled? I havent tried this yet with full cpu throttle?

Hi again guys

few more tests with Core i7 920 and X58-UD5




200 BCLK
DDR3-1600Mhz 9-9-9-24
Uncore x16
memory 8x


http://img376.imageshack.us/img376/2...hzprimeit8.jpg








----------------------------




205 BCLK
DDR3-1640Mhz 9-9-9-24
Uncore x16
memory 8x
Turbo ON




http://img520.imageshack.us/img520/156/4300dy7.jpg










superpi 1M


http://img520.imageshack.us/img520/1016/spinovodj3.jpg




:up:

as far I went today.
I might try for sub 9 tomorow.
the cpu and board are however picky at those speeds.
http://www.svensknlp.nu/databilder/maxed.jpg

Another quick question. How do RAM timings effect how you OC? For instance mine will have 7-7-7-20 timings, does that make it any different for me to OC than those of you with 8-8-8-XX timings?

I just wanted to say THANKS GUYS. I was able to use the info gathered
in this forum to get my new rig setup. Unfortunately i am still waiting on
a water block so am not doing anything spectacular here. Here is what
i have so far on air.

Asus P6T
I7 920 - batch=3836A866
Patriot 1600 6GB

Multi: x19
BCLK:200
Dram: 1600
Uclk: 3208
QPI Link: 9624
CPU: 1.23v
QPI/Dram: 1.28v
Vdim: 1.56v
Loadline: Enable

Quote:

---

Originally Posted by Jupiter

I just wanted to say THANKS GUYS. I was able to use the info gathered
in this forum to get my new rig setup. Unfortunately i am still waiting on
a water block so am not doing anything spectacular here. Here is what
i have so far on air.

Asus P6T
I7 920 - batch=3836A866
Patriot 1600 6GB

Multi: x19
BCLK:200
Dram: 1600
Uclk: 3208
QPI Link: 9624
CPU: 1.23v
QPI/Dram: 1.28v
Vdim: 1.56v
Loadline: Enable

---

Sweet, looks like you have a nice 3.8ghz OC on your hands there. Did it pass the 24/7 testing yet, or is it just gaming stable?

Quote:

---

Originally Posted by Hot'n'Chilly

Sweet, looks like you have a nice 3.8ghz OC on your hands there. Did it pass the 24/7 testing yet, or is it just gaming stable?

---

I tried running prime but it got over 70C pretty quick so i stopped it.
I will run some 24/7 testing once i get a water block. I did test at
3.8 and 4.0 while gaming 'NFS Undercover' and it ran perfectly fine.

Quote:

---

Originally Posted by Jupiter

I tried running prime but it got over 70C pretty quick so i stopped it.
I will run some 24/7 testing once i get a water block. I did test at
3.8 and 4.0 while gaming 'NFS Undercover' and it ran perfectly fine.

---

Your results are encouraging. I've been lowering the voltages trying to find the sweet spot today. My initial vCore was 1.31/1.35 (Windows/Bios). Now I am at 1.3/1.26. For stability testing I've been using OCCT v2.01b for 30"-60". In between testing I've been gettin' my game on making sure the settings are good.

I'm using the Gigabyte EX58-Ext. (cpu batch# 3837A728). I'm also using the Cooler Master V8 and I gotta say I like it so far. Keeps the temps manageable until I can put everything back under the wet stuff.

200 BCLK
DDR3-1600 Mhz 9-9-9-24
Uncore x16
Memory 8x
Turbo ON

Keep pushing it and posting your results! :up:

http://img.photobucket.com/albums/v3...920ss_126v.jpg

Looks like 1.25v lowest 1 hour OCCT stable vCore.

http://img.photobucket.com/albums/v3...0ss_12498v.jpg

Super Pi: 32M

i ran the same test HT on / HT off and got 13827 ht on 13542 ht off the difference is basicly the cpu score

Is everyone turning Loadline Calibration on?

12GB (6x2GB) GSkill PI 1333 7-7-7-19-1T
ASUS P6T Deluxe
i7 920 cooled by XSPC Delta V3, 750lph, 240 rad

With this system BCLK will not do 190MHz, it is currently priming at 185MHz (3.72GHz), relaxing the RAM to 9-9-9-27 doesn't help, setting tRFC to 110 doesn't help, using 2T or 3T doesn't help. But with only 6GB the board will do 190x20 or 200x19.

Has anyone else found that BLCK is limited by populating all RAM slots?

I will only use loadline enabled on mine. Loadline off = 1.39 bios, 1.36 idle, 1.31-2 load for prime stable, overshoot should be ~.075 v (1.39bios-1.315load) if GB follows intel specs. Loadline on = 1.35 bios, 1.328 idle, 1.31 load for prime stable, overshoot should be again ~.075v from load or ~1.39, only with loadline on I am not being babysat by bios setting = max overshoot, I have to realize overshoot is still going to ~1.39, ie whatever overshoot was from loading vcore with loadline off.

Between being babysat with 1.36 idle vs no babysitting and 1.328 idle, I will take the lower idle.