Introduction:

"Wohoow ... awesome". That's my first reaction when AMD published their first overclocking results of the 45nm Phenom II; not only because they showed frequencies way above 6GHz, but mainly because they eliminated the cold bug issue, which was the plague of every Phenom ánd Intel cpu in the recent past. In Las Vegas, during the CES event, the people of AMD teamed up with SF3D and Sampsa from Team Finland to perform the world's first overclock on Liquid Helium, which has a boiling temperature around -269°C, only four °C above the absolute zero point. Effectively, they were able to run the CPU at -232°C, resulting in a 6.3GHz 3D stable CPU.

When the first series of official AMD Phenom II reviews was released we could finally find out if the results obtained by the AMD team at CES was repeatable or not. We went through almost 30 reviews of the new Phenom II X4 940 but were quite disappointed by the overclocking tests which were almost solely focused on air-cooling. Time to complete the testing; I teamed up with Belgian LN2 OC guru Blind_Ripper, got 30L of LN2 and set up the following system to check the true overclocking limits of the Phenom II:

Madshrimps' Ph-II LN2 Test Setup
CPU	AMD Phenom II X4 940 (0850BPEW, batch 0323)
Cooling	Dragon F1 EE
Mainboard	DFI Lanparty DK 790GX-M2RS
Memory	2 * 1GB G.skill F2-8500CL5S
Other	Asus ENGTX295 Antec 1000W PSU Western Digital 320Gb SATA HDD Windows XP SP3 OCZ Freeze thermal paste

It’s always good to have a goal when you’re dealing with LN2 as you don’t have unlimited time or resources. Our aim:

Highest overclock on LN2, borderline stable

Highest overclock LN2, stable for 2D benchmarks

Highest overclock on LN2, stable for 3D benchmarks

We turned to HWbot to find out what competition we’re up against, the bar was set at <12 seconds in superpi 1m and off we went!

Results


(Clickable>


SuperPi 1M - Pifast - 3DMark05 (clickable)


(clickable)

In HWBot I managed to hit first place in the 1M benchmark, however, a Japanese top-overclocker named Kyosen, who does not participate in the HWbot league, has already been able to hit 11.156s. No doubt, I will re-try this CPU and see if I can get any closer to his score.

Conclusive thoughts

To be honest, I have mixed feelings regarding the obtained results of this overclocking session. On the one hand, Blind and me did hit some nice numbers such as the first place in SuperPi 1M in the Phenom II X4 940 category over at HWBot (third overall) and the cpu was in fact quite stable in 3D, 5.6GHz using only 1.76v, but ... those numbers were all done on a single core, not quad core. With all cores at the same frequency, the maximum stable will probably be around 5.4-5.5GHz, which is quite far under the 6.1GHz we reached in 1M. We hope that it's just because our sample is not the best there is and that with a new, handpicked, sample we'll be able to reach higher frequencies.

Regarding the motherboards, I have to say I'm impressed. Not only did it withstand temperatures below -180°C, it was in fact quite stable as well. In addition, we had not a single bios or boot issue during the entire 7h overclocking session, which is fairly impressive.

To end, I'd like to give my appreciation to:

Wendell and Erik from DFI for providing us the DFI LanParty DK 790GX-M2RS

Milan from Asus for sending us a sample of the Asus ENGTX295

Steeve from AMD for the Phenom II X4 940

Tobias from OCZ for the package of OCZ Freeze

Blind_ripper ... for the big help during the session

---

I hope you enjoyed reading this OC report, until next time, click the banner below to read up on our previous overclocking endeavors:

---

Extra's: LN2 in action


LN2 flowing over the container, no coldbug, not at all


Running 3DMark05 using one core: check the temperature in this load situation, it barely increases