AMD "Thuban" Core (Phenom II X6) XS Overclocking Charts

[poke349]

The download site for y-cruncher went down today. Hostmonster terminated the contract after some crap over TOS.

The temporary download link is (removed).

EDIT:
The site is back online on a new host, but it's still awaiting DNS propagation before it will activate.
(removed)

EDIT 2:
Looks like it's back online. (At least for me.)


I don't know how long it will take for me to get the site back up on a new host. So for now, this will have to do.

[DedEmbryonicCe1]

I am using a 1090T with DDR2 but I am not near a final overclock yet so I didn't want to post before that point. As far as stability testing I did 3.8GHz last night with a two hour Prime95 25.11 - Blend test:



Tonight I tried the same test at 4.0GHz and slightly higher voltage (~1.425v? core, 1.185v CPU-NB) and the system shutdown within a minute. The same motherboard and PSU were able to Prime my 940BE at 1.52v core and 1.47v CPU-NB for comparison.

If you're just looking for numbers with DDR2 I've benched 4.0GHz with 3DMark06, WPrime, 7-Zip, Cinebench 11.5, and Everest 5.50 so far.

[felix_w]

I am using a 1090T with DDR2 but I am not near a final overclock yet so I didn't want to post before that point. As far as stability testing I did 3.8GHz last night with a two hour Prime95 25.11 - Blend test:



Tonight I tried the same test at 4.0GHz and slightly higher voltage (~1.425v? core, 1.185v CPU-NB) and the system shutdown within a minute. The same motherboard and PSU were able to Prime my 940BE at 1.52v core and 1.47v CPU-NB for comparison.

If you're just looking for numbers with DDR2 I've benched 4.0GHz with 3DMark06, WPrime, 7-Zip, Cinebench 11.5, and Everest 5.50 so far.

Very nice DEC..thank you for your post...

2600NB with 1.185 could be your stability problem ? Maybe 1.2 could do the trick for you. What ram is that u r using?

Is it ok for you to post the rest benches you mentioned?

[zaraza]

whats the upper limit of voltage you want to give the CPUNB? 1.4?1.5?

[wez]

whats the upper limit of voltage you want to give the CPUNB? 1.4?1.5?

1.5'ish on air, but usually the scaling is gone long before that.

[DedEmbryonicCe1]

CPU 4.1GHz - 1.424v
CPU-NB 2.6GHz - 1.185v

3DMark06 - CPU Score 7142


7-ZIP - Compressing 19967 MIPS / Decompressing 24744 MIPS / Overall 22356 MIPS


Cinebench 11.5 64bit - 7.29pts

CPU-Z has the correct speed. Cinebench displays my boot speed even after I've changed it with AOD. The earlier 6.91 score was at 4.0GHz.

WPrime 2.03 - 32M: 7.083s / 1024M: 206.545s

This one was a little odd because going from 3.8GHz - 4.1GHz I didn't notice a change in the 32M time here, only the 1024M score improved noticeably (222.925s @3.8GHz / 2.6GHz NB).

Everest 5.50 - See Below


As for Everest I can run the full suite at 4.0GHz and below but 4.1GHz isn't working out. I tried bumping vCore to 1.456v and CPU-NB to 1.26v using the smallest increments in AOD, and testing each along the way. When that didn't work I skipped ahead to a vCore of 1.504v and still the system shuts down during the same test - FPU Julia. During this test my Kill-A-Watt reports a jump to 414-415 Watts right before the crash. It looks like I'll need a stronger PSU or motherboard to eliminate variables. My mosfets are barely warm as they have lapped Enzotech/Swiftech forged copper heatsinks epoxied on with Arctic Alumina thermal adhesive plus a 1400rpm 120mm fan blowing on them. I want to believe they would at least get hot before they cause any problems so I'm leaning towards it being the PSU. I went back to the BIOS and disabled one core to see if that would work but unfortunately Everest just skips a lot of the CPU tests with my X5.

[Chri$ch]

new CPU for the list

• CPU Model: AMD Phenom II X6 1090T BE
• CPU Stepping: CCBBE CB 1011MPMW
• CPU Frequency: 4015 MHz
• CPU vCore: 1.38 - 1.392V
• CPU Multiplier: 20x
• CPU Turbo: Disabled
• CPU NB Speed: 3011 MHz
• HT Ref Speed: 200 MHz
• RAM Speed: DDR3-1606
• RAM Timings: 8-9-8-24-1T
• RAM Configuration: 2 x 2GB
• RAM vDIMM: 1.65V
• Motherboard: ASUS Crosshair IV Formula
• Chipset/Socket: 890FX + SB850, AM3
• Cooling: Water (Heatkiller 3.0)
• Temps: 26C Idle / 41C Load
• Operating System: Windows 7
• 32/64-Bit: 64
• Stable/Suicide/Untested: Stable

[felix_w]

Excellent DedEmbryonicCel, thank you for all info!!!!

[Particle]

I think I'm going to change how the chart is done and just have one chart with color-coded rows based on "stability quotient". Basically, it'll be how many hours or minutes or something of some stability test have been performed. Very few people (nobody) seem interested in doing what I would consider a meaningful amount of stability testing. Any thoughts?

[Titan7171]

I think I'm going to change how the chart is done and just have one chart with color-coded rows based on "stability quotient". Basically, it'll be how many hours or minutes or something of some stability test have been performed. Very few people (nobody) seem interested in doing what I would consider a meaningful amount of stability testing. Any thoughts?

Yes I have a thought.........for one, how about removing the people who havent shown the stability requirement until they do.

Also, lower the stability requirement to 4 hours instead of 8. 8 hours is to long especially if you want to play a game during the day. which brings up another point, when I run a stability program,( Ive done a few 6 hour runs of prime) and I personally like to be there monitoring now and then so if something happens I can go into the bios and check on settings. Running 8 hours of stability while you sleep only to wake up to a rebooted PC or one that might not even turn on is not a fun thing people want to do to get on a list.

These are just my thoughts and I was actually going to do the 8 hours of Y-cruncher.....but Im still testing settings and havent really gotten to the stability part yet.

people are still trying to figure out their stable settings, plus alot of people still dont have the cpu yet.

[I.nfraR.ed]

Maybe it's a good idea, but 20 - 30 min of LinX or Prime doesn't mean anything. It's very very basic, so you know that it's possible to make it full stable at some point with more tweaking.
Well, it could be full stable at that settings, but you didn't check, right? I ran 50 min Prime Blend stable at 4.3GHz with 1.45V, but then one of the cores gave an error, so...it's not stable.
2-3 hours is much better than nothing, though. As you saw, I ran y-cruncher for almost 8 hours when it failed... I'm willing to run it overnight, but I have no time atm to do tweaking.
Will try again next couple of days.

PS: totally agree with Titan...I ran y-cruncher overnight and when I got up I saw that it was still running, but it failed after some time I also want to be near the pc to make corrections in the bios if needed.

[demonkevy666]

anyone care run coredamage for me I want just want to see the max cpu temperature ?

[mngdew]

new CPU for the list

• CPU Model: AMD Phenom II X6 1090T BE
• CPU Stepping: CCBBE CB 1011MPMW
• CPU Frequency: 4015 MHz
• CPU vCore: 1.38 - 1.392V
• CPU Multiplier: 20x
• CPU Turbo: Disabled
• CPU NB Speed: 3011 MHz
• HT Ref Speed: 200 MHz
• RAM Speed: DDR3-1606
• RAM Timings: 8-9-8-24-1T
• RAM Configuration: 2 x 2GB
• RAM vDIMM: 1.65V
• Motherboard: ASUS Crosshair IV Formula
• Chipset/Socket: 890FX + SB850, AM3
• Cooling: Water (Heatkiller 3.0)
• Temps: 26C Idle / 41C Load
• Operating System: Windows 7
• 32/64-Bit: 64
• Stable/Suicide/Untested: Stable

Hmm... those MPMW's look might good. Requiring lower voltage to go beyond 4.2.

[FlanK3r]

I think, as said Chew, 3h prime its OK for most users.

[zaraza]

Uploaded with ImageShack.us

there we go 8 hours stabl

[Particle]

Looking good there, zaraza! I've updated you on the list. What are your new RAM timings and frequency so that I may update that too?

[zaraza]

Ram timings havn't changed they should be

8-8-8-20. Frequency is 1524.

Im going to try for 4ghz

my NB is 1.38v my CPUNB is 1.325 ( i feel that my CPUNB is far too high but who knows)

[BETA.]

I dunno if I got a bad chip or it's some bug with the mobo. But, i'm running the 1055t on a gigabyte 890gpa-ud3h realtemp says my stock VID is 1.4250 and in BIOS it's the same. So far I passed 20 passes of LinX with 14x multi, 286 Mhz bus speed, 1.55 vcore, 1.36 NB Vcore, 2.20v CPU PLL, 2288 NB Freq and 2002 HT Link. Ram is 9-9-9-24 1.65 @ 1525 Mhz. Everything else on auto. Currently working on Prime though. Should I head back to Frys looking for another chip?

[Particle]

If it can do 4GHz at 1.55V, I'd say you're fine. If you're asking if I think you should buy a second one, that's up to you. Personally, with a result like yours I'd probably just run with it. If you're asking if I think you should exchange it for a different one, absolutely not. That would be unethical to exchange it because it won't overclock enough when it is functional at stock.

[zaraza]

I dunno if I got a bad chip or it's some bug with the mobo. But, i'm running the 1055t on a gigabyte 890gpa-ud3h realtemp says my stock VID is 1.4250 and in BIOS it's the same. So far I passed 20 passes of LinX with 14x multi, 286 Mhz bus speed, 1.55 vcore, 1.36 NB Vcore, 2.20v CPU PLL, 2288 NB Freq and 2002 HT Link. Ram is 9-9-9-24 1.65 @ 1525 Mhz. Everything else on auto. Currently working on Prime though. Should I head back to Frys looking for another chip?

I believe my CPU PLL is set at 2.5v stock. Maybe its your PLL voltage thats holding you back.

[zaraza]

I am thinking it that it is my board holding me back and causing random crashes at 4ghz. I really do. But theres nothing i can do, im not gonna return the board.

[BETA.]

If it can do 4GHz at 1.55V, I'd say you're fine. If you're asking if I think you should buy a second one, that's up to you. Personally, with a result like yours I'd probably just run with it. If you're asking if I think you should exchange it for a different one, absolutely not. That would be unethical to exchange it because it won't overclock enough when it is functional at stock.

Well, it just passed it's 7th hour of prime, small fft's, and once it passes through the 10th hour i'm gonna call it "stable enough". After that, i'll start to tinker with other settings to see if I can lower the Vcore

I believe my CPU PLL is set at 2.5v stock. Maybe its your PLL voltage thats holding you back.

I have read many OC guides across many enthusiast forums around the net via google search, lol and many of these guides actually say to lower CPU PLL in some cases. It worked for me with my Intel i7 rig, so I just tried pulling the PLL with this one. I'm a noob at AMD OCing, so I could be totally wrong on doing this.

[zaraza]

Well, it just passed it's 7th hour of prime, small fft's, and once it passes through the 10th hour i'm gonna call it "stable enough". After that, i'll start to tinker with other settings to see if I can lower the Vcore



I have read many OC guides across many enthusiast forums around the net via google search, lol and many of these guides actually say to lower CPU PLL in some cases. It worked for me with my Intel i7 rig, so I just tried pulling the PLL with this one. I'm a noob at AMD OCing, so I could be totally wrong on doing this.

Go ahead and set it to 2.5 as well as your CPUNB to 1.25-1.3. Then boot into windows and check stability.

[Zeus]

According to the numbers this one must be very good:

[FlanK3r]

yes, u got this today?