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

[ah_khoo]

my attempt with a rented chip... (yes, i'm too poor to own an X6 )

CPU Model: AMD Phenom II X6 1055T
CPU Stepping: CCBBE CB 1012FPMW
CPU Frequency: 4100 MHz
CPU vCore: 1.488V
CPU Multiplier: 14x
CPU Turbo: Disabled
CPU NB Speed: 2637 MHz
HT Ref Speed: 293 MHz
RAM Speed: DDR3-1562
RAM Timings: 8-8-8-24-1T
RAM Configuration: 4 GB (2 x 2 GB)
RAM vDIMM: 1.65V
Motherboard: ASUS M4A88TD-V EVO /USB3
Chipset/Socket: AM3, 880G + SB850
Cooling: Air (Cogage TRUE Spirit)
Temps: 47C Idle / 65C Load
Operating System: Windows 7
32/64-Bit: 64
Stable/Suicide/Untested: Untested since only 1 hour prime stable

[lzxhenry]

CPU Model: AMD Phenom II X6 1055T
CPU Stepping: 1018 BPMW
CPU Frequency: 4004 MHz
CPU vCore: 1.5V BIOS
CPU Multiplier: 14x
CPU Turbo: Disabled
CPU NB Speed: 2860 MHz
HT Ref Speed: 286 MHz
RAM Speed: 1525 Mhz
RAM Timings: 7-8-7-24-1T
RAM Configuration: Team Elite 1600 4GB (2x2GB)
RAM vDIMM: 1.6V
Motherboard: Gigabyte GA-890GPA-UD3H
Chipset/Socket: 890GX, SB850/AM3
Cooling: Xigmatek Dark Knight (Stock Fan)
Temps: 38C Idle, 65C Load
Operating System: Windows 7 Ultimate 64bit
Stable/Suicide/Untested: Stable Prime95 (4 hrs+)

[Gameman]

Well I got it..i ain't reporting anything because i haven't tested it yet...I just wanted to do it...

3.6 will do for now until i can get a better motherboard with crossfire.... had to settle for an 880G for the next 2 month's

it feels a lot like my former quad at 4.0-4.2....encodes an xvid video just as fast as my q9550 @ 4.0...maybe a bit faster

I'll look into going to 4.0ghz later, but for now this is a significant boost in speed





I'll report later, I gotta first figure out what the hell i'm doing in the bios...

Thanks


First update

definitely a joy to encode with...2 min faster than my Q9550 and Q9400 at 4.0 for an Xvid 25 min sitcom video
(this is pass 2)...and COOL - believe it or not this 3.6ghz on the stock cooler...Thanks



I wanna thank you AMD guys the idea of the x6 -6 core...this is a hell of an upgrade

not to mention the best multi-tasking experience I have ever had...beautiful

[demonkevy666]

mine under volts really well 3.3ghz @ can't go low then 1.1375. :/

[crazydiamond]

mine does 3.2ghz @ 1.116v, 3.5ghz @ 1.2v

[Gameman]

I was only able to get the 880GM UD2H for now but it's definitely an all around TEMPORARY solution for if your graphics card suddenly dies like mine did...

I can actually play Crysis (original) with this x6 Thuban at 3.6 ghz
and the onboard 4250 non oc'd



settings:
1024x768 on 40'' monitor
avaerage 30 fps
fully playable if you don't care about eye candy
on MINIMUM

now i didn't expect for Crysis to even start with this onboard crap..


but 30 fps at 40'' isn't bad for an onboard graphics card...
definitely more than i expected...they certainly have come a long way in the onboard department compared to last generation

I imagine continued use of the onboard for a monster like this would blow the gpu though...

just waiting on a 4670 and then a 5830 later

Thanks

edit: while i'm asking anyone with the 880g/890g/gx have luck with that Hybrid Crossfire thingy?????
or would you have to use a garbage 5450/4350 for it to work????

[w0mbat]

CPU Model: AMD Phenom II X6 1055T
CPU Stepping: N/A
CPU Frequency: 3200 MHz
CPU vCore: 1.092V
CPU Multiplier: 14x
CPU Turbo: Disabled
CPU NB Speed: 2058 MHz
HT Ref Speed: 228 MHz
RAM Speed: DDR3-1220
RAM Timings: 8-8-8-24-1T
RAM Configuration: 4 GB (2 x 2 GB)
RAM vDIMM: 1.5V
Motherboard: ASUS M4A87TD EVO
Chipset/Socket: AM3, 870 + SB850
Cooling: Air (Scythe Big Shuriken)
Temps: 39C Idle / 57C Load
Operating System: Windows 7
32/64-Bit: 64
Stable/Suicide/Untested: Stable

[Gameman]

Finally did it..and on 880G too... 4.0 ghz seems to be stable (for me anyways)

anyways i'l edit this and add the figures later, but believe it ornot this was DONE ON THE ULTRA-CHEAP 880GM CHIPSET (880GM-UD2H) AND DID NOT REQUIRE THE FANCIER 890 AS OF YET....but we'll see



just happy that i don't have to go for the 890 GPA just as of yet....

had to raise the nb volts a little and the nb vid number...

but it seems to be good

sure a lot faster than at 3.6...

OK. Here, if for nothing else than posterity...

CPU Model: AMD Phenom II X6 1055T
CPU Stepping:
CPU Frequency: 4004 Mhz
CPU vCore: 1.45 in bios (cpuz 1.30??)
CPU Multiplier: 14x
CPU Turbo: Disabled - C and Q disabled - no turbo option
CPU NB Speed: 2573 MHZ
HT Ref Speed: (i asume this like the fsb or something) - 286
RAM Speed: DDR3 Kingston Valueram @ 1524 (1333)
RAM Timings: 7-7-7-20-CR1
RAM Configuration: 2 gb Valueram
RAM vDIMM: 1.70
Motherboard: GA-880GM UD2H
Chipset/Socket: 880GM + SB 750
Cooling: Stock
Temps: 40-45 (crap stock cooler only)
Operating System: XP
32/64-Bit: 32
Stable/Suicide/Untested: Untested as of yet - but does what I need it to do

I honestly gotta admit, I didn't think AMD had it in them...but there you go...

our friend Linus failed to mention that you gotta adjust the Northbridge...(at least on this one anyway)

thanks

updated: added proof of motherboard (ud2h)

[FL00D]

Lo guys,

Help me please. I've been having a hard time trying to stabilize my 1090T.

* CPU Model: AMD Phenom II X6 1090T BE
* CPU Stepping: CCBBE CB 1020MPMW
* CPU Frequency: 4000 MHz
* CPU vCore: 1.41V
* CPU Multiplier: 20x
* CPU Turbo: Disabled
* CPU NB Speed: 2600 MHz
* HT Ref Speed: default
* RAM Speed: DDR3-1600
* RAM Timings: 7-7-7-27-33-1T
* RAM Configuration: 4 GB (2 x 2 GB)
* RAM vDIMM: 1.7V
* Motherboard: msi 790FX-GD70
* Chipset/Socket: AM3, 790FX + SB750
* Cooling: Scythe Mugen 2 (Ultra Kaze 2000RPM push&pull)
* Temps: 41C Idle / 56C max Load
* Operating System: Windows 7
* 32/64-Bit: 64
* Stable/Suicide/Untested:

Prime95 blend: rounding errors within 4 hours. I raised the CPUNB voltage to 1.25V. Prime95 blend has been running for approx 4 hours. I wonder whether higher cpunb voltage will sort out rounding errors... I don't really think my CPU would need higher voltage since I got rounding errors at 1.4375V (1.2V CPUNB) as well. But I don't know, isn't 1.25V a bit too high for Thuban? It's ok for Deneb but I read that Thuban didn't need that high CPUNB voltage, not even at 2600MHz.

All answers are greatly appreciated.

PS: RAM tested 24/7 stable

[Particle]

Charts updated. Sorry for taking so long this time. The thread isn't as active as it once was so it didn't grab my attention as frequently to remind me.

[m0n4g3]

http://i529.photobucket.com/albums/d...LinxStable.jpg

* CPU Model: AMD Phenom II X6 1090T BE
* CPU Stepping: CCBBE CB 1015EPMW
* CPU Frequency: 4000 MHz
* CPU vCore: 1.39V (Although VID is set to 1.3375
* CPU Multiplier: 16
* CPU Turbo: Disabled
* CPU NB Speed: 3000 MHz
* HT Ref Speed: 250
* RAM Speed: DDR3-1667
* RAM Timings: 7-7-7-20-27-1T
* RAM Configuration: 4 GB (2 x 2 GB)
* RAM vDIMM: 1.7V
* Motherboard: MSI 890FXA-GD70
* Chipset/Socket: AM3, 890FX + SB850
* Cooling: Water (EK Supreme HF+ plate 1 with MCP650 pump and 360 RAD with scythe 1900rpm 110cfm fans)
* Temps: 14C Idle / 28C max Load core temp values so 24C idle / 38C Load
* Operating System: Windows 7
* 32/64-Bit: 64
* Stable/Suicide/Untested:

2 hours linx Stable. I'm classifying as stable. Don't wanna run it anymore than that tbh.

[FL00D]

Charts updated. Sorry for taking so long this time. The thread isn't as active as it once was so it didn't grab my attention as frequently to remind me.

Hiya,

I need a bit of help with testing. My 1090T seems to be stable @ 4.1GHz 1.425V VDD, 2.6GHz CPUNB 1.25V. Memory: Kingston HyperX 1600MHz 7 7 7 27 33 1T 1.7V (msi 790FX-GD70) I've run 13hrs of Prime95 Blend, 12hrs of Prime95 Large FFTs. No crash. I wish to run one more 8+hrs of something but I don't really know which test to run/wot app to use. I really want to make sure that my cfg is rock solid. [Actually another reason is that I've just updated the mobo BIOS so some more stress testing won't hurt] I was told that 8hrs of Linx with Memory (MiB) set to ALL would be the best. Then I read here that y-crucnher is the preferred app to stress test the CPU. Now wot? Isn't Linx just as good as y-cruncer?

Thanks a lot!

[FL00D]

Hiya,

I need a bit of help with testing. My 1090T seems to be stable @ 4.1GHz 1.425V VDD, 2.6GHz CPUNB 1.25V. Memory: Kingston HyperX 1600MHz 7 7 7 27 33 1T 1.7V (msi 790FX-GD70) I've run 13hrs of Prime95 Blend, 12hrs of Prime95 Large FFTs. No crash. I wish to run one more 8+hrs of something but I don't really know which test to run/wot app to use. I really want to make sure that my cfg is rock solid. [Actually another reason is that I've just updated the mobo BIOS so some more stress testing won't hurt] I was told that 8hrs of Linx with Memory (MiB) set to ALL would be the best. Then I read here that y-crucnher is the preferred app to stress test the CPU. Now wot? Isn't Linx just as good as y-cruncer?

Thanks a lot!

Guys, I wouldn't really recommend using Prime95. Use LinX instead. I tried to run LinX (memory size set to all) but it gave me error within the first two runs. Before that, I ran 13 hrs of Prime95 Blend, 12hrs of P95 LargeFFT. Yesterday I upgraded the mobo BIOS so there's a slight chance that the new BIOS has weaker overclocking potential but I don't really think so. I don't know whether it is possible to run almost 30hrs of prime95 with absolutely no crashes and then LinX crashes within 10 mins. This is quite peculiar. At 1.45V LinX seems to be working fine. Will see after another 12 hrs... That's a shame AMD messed up the bloody temp sensor. Adding 10C plus to my current temps is really disappointing. That being the case my temps may reach 60C @ 1.45V.

[Particle]

They all behave a little differently. If you can do small FFTs in Prime or OCCT as well as a run of blend, you're generally good to go. Small FFTs stresses the core logic really hard, and blend stresses the memory and cache systems more. If you've already done that y-cruncher is probably a good bet. It will hammer the memory subsystem to a high degree and will test that angle a bit more.

If you are really interested in a high degree of stability, I'd suggest you run these tests for 24 hours instead of 8.

---

@ Thread: Your guys are lucky with your chips. Both my 1055 and 1090 struggle with 4GHz at 1.55-1.575V.

[RhoXeD]

Hey there, finally got my 1090T heres some quick results.

* CPU Model: AMD Phenom II X6 1090T
* CPU Stepping: CCBBE CB 1023GPMW
* CPU Frequency: 4118 MHz
* CPU vCore: 1.474V
* CPU Multiplier: 20.5x
* CPU Turbo: Disabled
* CPU NB Speed: 3000 MHz
* HT Bus Speed: 2000 MHz
* HT Link Speed: 200 MHz
* RAM Brand and Model: G.Skill RipJaws 7-8-7-24 @ 1600Mhz
* RAM Speed: DDR3-1600
* RAM Timings: 7-8-7-24-2t
* RAM Configuration: 2x2gb
* RAM vDIMM: 1.65V
* Motherboard: Gigabyte 890FXA-UD5
* Chipset/Socket: 890FX / AM3
* Cooling: Xigmatek Thors Hammer
* Temps: 29C Idle / 48C Load
* Operating System: Windows 7
* 32/64-Bit: 64
* Stable/Suicide/Untested: Stable - In process



Uploaded with ImageShack.us

also - http://valid.canardpc.com/show_oc.php?id=1307791

[FL00D]

They all behave a little differently. If you can do small FFTs in Prime or OCCT as well as a run of blend, you're generally good to go. Small FFTs stresses the core logic really hard, and blend stresses the memory and cache systems more. If you've already done that y-cruncher is probably a good bet. It will hammer the memory subsystem to a high degree and will test that angle a bit more.

If you are really interested in a high degree of stability, I'd suggest you run these tests for 24 hours instead of 8.

---

@ Thread: Your guys are lucky with your chips. Both my 1055 and 1090 struggle with 4GHz at 1.55-1.575V.

I ditched Prime95. LinX stops within 10 mins with an error message while Prime95 would run for 12hrs or so without the slightest sign of instability.

Hmm, these chips should easily handle 4GHz. :S My 1090T can do 4GHz @ 1.408V (knock on wood). However, I can't get it go any further, at least not with my Mugen 2. 2-3 more hrs of LinX and then I'm gonna try my luck at the CPUNB.

Cheers for the info anyway!

PS: Wot do you think about LinX?

[Particle]

Hard to say. I've not used LinX yet. I generally don't use Linpack based testers since they don't produce a very consistent load on AMD CPUs.

[FlanK3r]

RHODEX: whau, nice air clock man!

[RhoXeD]

RHOXED: whau, nice air clock man!

thanks Flank3r, I have taken it to 4.3 already, but i didn't have but an hour, so not stable yet. Shooting to bench at 4.4-4.5 for air.

[FL00D]

Hiya again,

I've been struggling for 5 bloody days to get my 1090T stable @ 4.02GHz. I'm 100% sure that at least half of the entries on the list would be declared as unstable if ppl here tried to run LinX for 24 hrs. It is really very hard to pass LinX for 24+hrs. My 1090T probably needs more voltage. 1.416V was the highest I tried. I'm installing my new NH-D14 this afternoon. Hopefully it will enable me to raise the voltage a bit more. Anyway, in my particular case, I reckon, the memory was the culprit. I shouldn't have tried testing stability over and over again with the memory already oced. I do hope that tight timings (7 7 7 27 33 1T @ 1.7V) is the very reason why I could not pass LinX so far. LinX gave me errors randomly. In 6hrs, 11hrs, 21hrs. Quite peculiar, init?

Will report back in a few days.

Cheers,

[demonkevy666]

Take the ram out of the equations by letting the run 9-9-9-27-33.

first core speed, then Nb speed, last would be ram speed, for how I do it.

My boards bios is odd I've got massive over volt on load form 1.375 all the way up to 1.456 on load.
it seems to set up progressively too 1.35-1.392-1.424-1.456 from idle to load (it was set 1.365 in bios)

[FL00D]

Take the ram out of the equations by letting the run 9-9-9-27-33.

first core speed, then Nb speed, last would be ram speed, for how I do it.

My boards bios is odd I've got massive over volt on load form 1.375 all the way up to 1.456 on load.
it seems to set up progressively too 1.35-1.392-1.424-1.456 from idle to load (it was set 1.365 in bios)

Yh that's exactly wot I'm gonna do. I will leave the memory settings at default. My big question is whether I should/could test the core speed and CPUNB speed at the same time. You know I don't rly want to run 2-3 24-hour tests nevertheless testing more components at the same time is always risky since in case of instability you don't know which component needs more voltage. So first off, how many volts do you guys think the CPUNB needs @ 2.6GHz? I'd say 1.25V but I'm not sure.

[FL00D]

Take the ram out of the equations by letting the run 9-9-9-27-33.

first core speed, then Nb speed, last would be ram speed, for how I do it.

My boards bios is odd I've got massive over volt on load form 1.375 all the way up to 1.456 on load.
it seems to set up progressively too 1.35-1.392-1.424-1.456 from idle to load (it was set 1.365 in bios)

Yh that's exactly wot I'm gonna do. I will leave the memory settings at default. My big question is whether I should/could test the core speed and CPUNB speed at the same time. You know I don't rly want to run 2-3 24-hour tests nevertheless testing more components at the same time is always risky since in case of instability you don't know which component needs more voltage. So first off, how many volts do you guys think the CPUNB needs @ 2.6GHz? I'd say 1.25V but I'm not sure.

[Celestial Bliss]

CPU Model: AMD Phenom II X6 1055T
CPU Stepping: N/A
CPU Frequency: 4030 MHz
CPU vCore: 1.456V
CPU Multiplier: 13x
CPU Turbo: Disabled
CPU NB Speed: 2170 MHz
HT Ref Speed: 310 MHz
RAM Speed: DDR3- 1652
RAM Timings: 8-8-8-24-2T
RAM Configuration: 4 GB (2 x 2 GB)
RAM vDIMM: 1.65V
Motherboard: GIGABYTE GA-890FXA-UD5
Chipset/Socket: AM3, 890FX + SB850
Cooling: COOLER MASTER Hyper 212 Plus (Push and Pull)
Temps: 23C Idle / 48C Load
Operating System: Windows 7
32/64-Bit: 64
Stable/Suicide/Untested: Stable Prime95 Blend for 2 Days



Also this is the first time I have overclocked anything from AMD in a while. So chime in if anything looks wrong. I plan on running this rig 24/7 for crunching. Also I am not sure if my vcore is too high.

- Bliss

[mecho_gruh]

Here is my result, I left the voltages in BIOS on optimized mode, lowered NB and HTT multiplyers, lowered the IGP frequenses adn left the memory on auto, beacause it does'nt work very well with this mobo.
It seems that Gigabyte 890GPA-UD3H is very capable motherboard, it has rock solid voltages and high HTT.

CPU Model: AMD Phenom II X6 1055T
CPU Stepping: 1008CPMW
CPU Frequency: 4200 MHz
CPU vCore: 1.488V
CPU Multiplier: 14x
CPU Turbo: Disabled
CPU NB Speed: 2400 MHz
HT Ref Speed: 300 MHz
RAM Speed: DDR3- 1200RAM Timings: 6-6-6-15-1T
RAM Configuration: 4 GB (2 x 2 GB)
RAM vDIMM: 1.5V
Motherboard: GIGABYTE GA-890GPA-UD3H
Chipset/Socket: AM3, 890GX + SB850
Cooling: Noctua NH-D14
Temps: 34C Idle / 54C Load
Operating System: Windows 7
32/64-Bit: 64
Stable/Suicide/Untested: Stable Prime95 Blend for 7 hours