thlnk3r

FlanK3r, great job man. Wow! Is she Prime/Orthos stable?

Good luck

__________________
AMD Overclocking Guides · Terms of Service · My CM690
Socket 939 Appreciation Club and Knowledgebase

Drin

Hi everyone

Sorry I have not been posting much, I was basically working on my OC for well over four days straight (excluding sleep, needing to go to uni etc). Also had to take the time out to tweak my new set of ram, and test it out.

Bad news is, I do not think it is possible to hit 4.0GHz stable given my current motherboard, memory, temps and/or bios. I've exhausted pretty much every single possible avenue, and the best I can do is ~1hr 15 mins prime95 stable. I can pass IBT, even on max, but that does not mean my system is stable I at least learned a lot, and tweaked my system better for overall performance.

In fact, I used to think (and advocate) that OCCT/linpack/IBT was the best stress test there was, but I've found that prime95 is a better measure of stability (so I've found) once you start getting into the higher overclocks. This is because the IMC is usually the main culprit, and is very hard to get stable. Slappa sums it up perfectly here (taken from the 955 oc thread):

Chew* also agrees with this view point (or was the first to suggest it, iirc) and I've found both of them to be right in practice.

Another example, to pass IBT at 3.9 given the specs below, I only need 1.5 volts. However, it'll crash after 2 hours in prime 95 and I've found to have my IMC/OC at least 3 hours or more stable in prime 95, that I needed 1.52 vcore as well as an extra notch of NB-vid. I can get it stable with 1.5v, but with much looser timings, and as temps do not bother me much, I'd rather have it at 1.52.

In either case, once again this does not prove my system is completely stable. As thlnk3r, txtmstrjoe and many, many others have said a cpu is only ever completely stable when its turned off.

However, in all honesty if your system can survive a few hours of IBT/linpack/OCCT it is pretty stable, and definitely not flaky. You could probably use it for months without the oc'ing causing it to crash. It does seems MUCH harder to get a high OC stable in p95 however, so if you are uncomfortable with using a system that might not be completely stable definitely try to test 3 hrs minimum in prime 95.

Anyway this will be my 24/7 OC for now, unless I can tweak it a bit better. I've given up on 4.0GHz, and probably will not be trying to achieve it anytime soon. Sorry for all those people who repped me, I know it feels like a let down. Certainly, I feel like I failed myself considering I believed I was stable. I guess in the end however, that 3.9GHz is pretty good, especially considering I'm using a 'bad' board with 'only SB600' I feel like I've done an alright job in the end!


Anyway my new submission, if you could add it please Xeno!:

Core speed: 3904 MHz
Bus Speed: 244.1 MHz (244x16)
Core Voltage: 1.52v
HT Link: 1952.5 MHz
NB Frequency: 1952.1 MHz
CPU-NB Voltage: 1.27
Motherboard: ASUS M3A32-MVP Deluxe
Bios: 1705
Memory: 2x2GB DDR2 Team Dark 1066 @ 813 4-4-4-12
Memory Voltage: 2.2v
Cooling: Prolimatech Megahalems with 2 San Ace 1011's shrouded

I'll also post bios shots of my final OC, in case it helps anyone (doubt it would, but you never know).

Good luck to everyone and their OC's!

Attached Thumbnails

 

__________________

FlanK3r

not, its not stable, stable is max 3.9GHz. This is max validation for AIR cooling (its hard with AIR get validation more than 4200MHz for most Phenoms II) and total unstable. Superpi crash with this clock and my PC restarted then.

__________________
AMD Phenom X4 955 BE@3.2-3.89GHz-1.485V or 3.914 MHz-1.5V 24/7, max 4386MHz, Gigabyte GA-MA790X-UD4, 2x2GB Hyper X 1066MHz@1130MHz 5-5-5-15-2t CM Hyper 212, 2x320GB HDD, Asus HD5770 1024MB ddr5, saphire HD3870 512MB GDDR4, saphire x1950pro 256MB ddr3, Seasonic 650W and 500W, x4 965 BE C3 coming soon!


AMD AIR cooling challenge Cinebench R10

AIR 4386MHz, 1.55V 4386 MHz x4 955 BE C2 rev. with AIR

PC2:AMD X2 4600+ s AM2 (Windsor)
PC3:AMD Athlon 3000+ s754 (Venice)
in memories:AMD X3 8650@2760MHz[/U], ASUS M2N-E

gsk3rd

well i am sorry to say that I am glad that you can not get 4.0 ghz stable either. I have tried just abot everything under the sun. Though I am not giving in just as simple, I will probably be going water next week sometime. I may actually try your similar settings if you dont mind.

__________________

Drin

Very nice gerikoh, glad to see you are still pushing it! +rep for the effort, and nice OC!

That is very impressive indeed FlanK3r Very hard to get into an OS at those speeds, and very, very few have achieved it on air! +rep

That's a nice cpu-z validation, but may I ask what you tested for stability? 1.55v with that high a CPU-NB voltage (I'm reading it as 1.5) is a heck, and I mean a heck of a lot of voltage for a 940 . Even the most extreme air cooling setups using cold air fed directly onto the HSF/Mobo will see temps a bit high, and I just do not see it happening on a Zalman 9700. I used one way back in the day when I was on a dual core, and while it works ok for that it struggles with quad cores. In fact, the 940 stock cooler is great for a stock cooler and compares with the Zalman.

Of course if you could prove me wrong I would be very happy to see such a feat And I would +rep you for the effort. Unfortunately I just do not see it surviving any serious stress testing. Prove me wrong!

__________________

Drin

Do not mind at all Go for it, and if you do hit 4.0GHz fully stable I will be very glad I'll be the first to +rep!

If it helps, I can post screenshots of my bios. But as you would know, when you start to hit 4.0GHz on a 940 you have to tweak everything to your particular setup. Stilll, I'm glad to try and help out if I can.

And while I'm not going to try it anymore for now, I spent ages trying to hit 4.0GHz Did almost everything possible as well. I just do not think its possible given all my current variables, and unless something changes I just do not think it is worth the time for me to try and hit it on my system anymore.

TBH I have only ever seen two instances: One from Chew* (P95 stable, not sure on OS ver) and XxBenignxX who posted here on this thread (IBT, x64 vista stable). So it can be done, do not give up!

__________________

gerikoh

thanks. i'm getting a 2 week vacation from school and now i'll have more time to do my oc. i'll try to push for more. i'm aiming for 3.8ghz now.

__________________


Ultimate Quad Core Clock for Clock Benchmarks:

**The Prolimatech Club**

Thesis/HTPC Notebook: Puma's Cartwheel

thlnk3r

Drin, 3.9Ghz stable is an excellent overclock

FlanK3r, another great overclock on air! Man these Ph II's are pretty amazing.

Great job guys

__________________
AMD Overclocking Guides · Terms of Service · My CM690
Socket 939 Appreciation Club and Knowledgebase

nisean

Hello Xeno,
Here are my modest numbers. I found the "Attach Files" option and used it to upload the screen capture at 1600x1000.

- Phenom II X4 940BE
- ??? (Stepping=2). I've already mounted the chip and Sunbeam CCF!
- Rev RB-C2
- 3604 = 267 * 13.5
- Voltage: CPU = 1.45
- Voltage: CPU-NB = 1.425
- NB Clock = 267 * 9 = 2403
- Voltage: NB = 1.3
- HTT Link Speed = 267 * 7 = 1869
- Mobo = GA-MA790X-UD4P
- Mobo BIOS Ver = F4
- Cooling: Air
- OS = Win XP SP2

About the temps during Prime95 and the NZXT Beta case... I have not installed all the fans yet. Doing some experiments to find out the best airflow. So, when I started Prime95 for the final stability check, the system temps kept on creeping up and up and at about 63C core, suddenly the system got reset. I checked the DRAM heatsinks for those skyblue GSKILL sticks and they were hot enough to burn my fingers in seconds. The NB heatsink was just equally hot.

After letting them cool down for 5 minutes I opened one side of the case and put a Home Depot box fan at lowest setting right next to the open side at an angle. That did the trick. At the apartment temp of 78F (25.5C) without A/C (just swamp cooler) the box fan kept the temp all thru' 24 hours of Prime95 below 60C.

It is unlikely that I would ever use the PC that hard to do anything but man, the Prime95 can really heat up CPUs. I OCed the much easier Intel Q9550 to 4GHz in the middle of winter and did face the same Prime95 heat issue then. But I managed 4GHz in 48 hours. With the unlocked PhII 940 I spent over 5 days trying to get close to 4GHz and maintaining 1:2 FSB to DRAM ratio (with 1066 DDRII) but no luck whatsoever.

In the most intense real life use, I once did these at the same time on the Q9550 system on Linux:
(A) Three Transcodes using 3 ffmpeg GNOME shells, each transcode had clip duration of at least 15 minutes;
(B) 1 video capture from 60-minute VCR tape to PC using PVR150 with H/W MPEG encoder;
(C) Cinelerra render to YUVMPEG4 stream for video clip of duration of 35 minutes.

All these read/writes were going through Linux S/W RAID-0 on two SATA-3GB/s 7200RPM drives on Intel P45 Southbridge chip of Gigabyte UD3R motherboard. Even then the max CPU usage was about 70-80% counting all the cores. Max temps using sensors-detect was about 57C. Guess what? The Linux S/W RAID had a hard time keeping up with all these writes but never made any error.

Anyway, I am watching the temps on the Ph-II 940 system for last couple of hours at idle but with the case closed. The temps are about 45C on cores. With the case open but no box fan, the temp goes down to 38C. Definitely the case needs some mods.
Regards.

Attached Thumbnails

 

thlnk3r

Nisean, for a case that hardly has any airflow those are still good temps. Have you gotten any ideas on how you want the airflow set up?

Great job on the overclock

__________________
AMD Overclocking Guides · Terms of Service · My CM690
Socket 939 Appreciation Club and Knowledgebase