• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

*** Official AMD Phenom II X6 1055T & 1090T Overclocking Thread ***

somehow dont think I will try match your 4.4 and 4.51 clocks in the other thread ... though you dont prove its prime stable .. lmao

this chip boots but wont run IBT high on 4.2 with 1.49v so not sure i could get much above that on 1.5v and temps below the 62C mark for more than a suicide run

happy @ 4Ghz .. its fast enough for me and what I do .. for now .. :D
 
Cozman said:
happy @ 4Ghz .. its fast enough for me and what I do .. for now .. :D
Click to expand...

Agreed, 4GHz is more than enough for a regular overclock, I just wanted to show yhe potential the chip has in my other thread. Try running LinX 64bit for 10 passes using all your available RAM. Takes considerable less time than 8 hours of Prime and will stress your CPU a lot more. This was recommended to me by a colleague and works wonders!
 
2600 on CPU-NB - I primed for 8 hours yesterday - without rebooting played some bfbc2, then set prime off going again (all without rebooting) last night - It was still running ok until 16:00 today but then boogied after I started browsing the web on it lol.

Had enough of trying for more - guess my 4x2Gb sticks of RAM are limiting me getting higher on the CPU-NB too without volts that would affect temps more. Saying that I can't complain with memory running at DDR1600 9-9-9-24/2T (stock SPD settings) - These IMC's are allot better than the Denebs

Settling for 24/7 settings of, [email protected] with [email protected] - keeps it nice and cool (well below the threshold). Time to game sum, add me on bfbc2 as Zoomee if anyone heres playin it ;)
 
i had similar glitchs but was always after Internet Security auto updated ... (Avira) i think it was a core firewall update as each time i had to ok every outbound net connection

i do run IBT and LinX 64bit every once in a while after a game or a video encode to make sure all is honky dory .. lol .. maybe a little over cautious :D
 
Just finished sinking up the mosfets, time to try for 4GHz again!

EDIT - Bah! Still throttles down mid way through a heavy Linx test, curse you 4+1 PWM. Time to dig out a fan!
 
Last edited:
thats a real shame HeX ... I would guess the 4+1 PWM was always gonna be one of the hurdles that had to be overcome for a big overclock

still impressive you got what you have so far on the slightly lower spec PWM
 
After a bit of tinkering and tweaking I think I have got my 1090T to run 4Ghz on 1.425v (bios reading 1.445v cpu-z reading) a slight improvement of 0.025v lower

put my ram @ 1333Mhz but with tighter timings of 7-7-7-20/2T 1.65v and locked HT to 2000 with NB 2600 .... not made a huge difference but cpu temp and the TMPIN2 temp have lowered by 3C to 4C in general when stressing

passed LinX 64bit 10 runs and IBT 20 runs high without a glitch and prime is 2+ hours and rolling

system is much more responsive and snappier as well so maybe lower frequency and tighter timings works best for some or maybe my ram just isnt upto 1600Mhz 100% stable (hate it when something gets a rating label but isnt officially supported beyond a lower rating)
 
ive had to overvolt my ram. Its rated at 1.35v, on auto my board does 1.4v.

Found that forcing it back down to 1.35v made things unstable. Dont know if thats a mobo/cpu or ram issue.
 
Cozman. up the ram voltage alittle.

my ram is rated at 1600 cl7 1.7v but it doesn't work at that voltage...

it needs 1.75v to be 90% stable and 1.8v to be 95% stable
 
tried 1.75v for ram rated to run @1.65v ... still had a glitch now and then .. may or may not have been the ram of course .. seems the system freezes on occasions when something access the network/internet

gut feeling is its the ram but not tested it seperately to find out for sure .. not sure there is much difference between ram @ 1333 7-7-7-20 or 1600 9-9-9-24 so playing it safe on the 1333 setting ... specially if cpu passes at a slightly lower vcore
 
Martini1991

You don't really need to spend 80 quid on a cooler.
The Titan Fenrir and Tuniq Tower 2 work well and have very good mounting brackets with high clamping pressure.

Avoid heatsinks which use a spring clip with the standard plastic frame as they do not have enough clamping pressure and perform poorly on the Phenoms.

Something that has become apparent during my own testing with the 1090T
is that they don't really like running the memory at DDR 1600.

I am able to run much lower core votage with the RAM at 1333, CPU runs a lot cooler and is very stable at 4GHZ.

With the RAM at 1600 i would get errors running IBT and occasionally prime 95
Incresing amounts of core voltage to try and stabalise things wasn't really helping.
I did get what appeared to be good stability by upping the vcore over 1.5v shown in cpuz when under load.
This generates too much heat though and when the ambients incresaed during the day i noticed the CPU was throttling.
I spotted this in the logs on my prime crunching software. the time taken to do each K/N pair increased between midday and 9PM.

I dropped the RAM down to 1333MHZ and now can maintain stability with 1.404v vcore shown in cpuz under load.
Temps are now at least 10C lower.
Monitoring yesterday showed 37C early morning and 45C MAX during mid afternoon, these are very good load temps as i am running all 6 cores flat out continually.

With the higher vcore required to run with the RAM at 1600 i was seeing temps in the 50's..
I suspect that somewhere around 55C the throttling kicks in.
Thermal control is turned off in my bios but it would appear that either it is not really disabled or the chip itself has self preservation built in.

I see others are also experiencing issues running with the RAM at 1600.
It is fine for benching but if you run high work loads for long periods it does let go

With RAM at 1333MHZ, NB at 2600MHZ and the CPU running 200x20 VCORE at 1.4V under load, everthing runs cool and stable.

For short periods it is certainly possible to up the NB to 2.8GHZ+, up the RAM speed to 1600+ and get good Vantage benches, Fritz, Wprime or whatever but my experience is that they cant maintain these settings long term
 
1600 runs fine on these CPU's, there are several of us on here running 1600MHz mem. And a few of us running 8GB of it which is even more stressful on the IMC.

Personally mine does 3.9GHz @ 1.4V, CPU-NB @ 2600 - 1.175V, MEM @ 1600 - 1.65V and it's 9hrs Prime stable, 20 runs Linx, 20 runs IBT. No problems at all.

As for the throttling you are seeing, it won't happen at about 55C, mine has happily hit 59C without a problem.

What does cause throttling seems to bee the PWM/MOSFETS overheating on your mobo due to the power draw. Getting a fan to blow over them seems to fix this and lets you OC higher without throttling.
 
Hex
The throttling was very hard to spot.
It was not the normal big drop in speed.

Maybe only 5% or so.
I can see the times in seconds for each Work Unit.
Normal time 345 seconds, throttled 365 seconds so not a large amount.
The increase in processing time is related to CPU temp.

With a lower vcore i haven't observed any slowdowns.

If i didn't use my CPU's for number crunching then i would never have noticed it.


I've got some more RAM here i'll give that a try at 1600MHZ and see if i need to up the vcore to run it stable.

I am currently using OCZ but i do have GEIL running in some of my other PCs i'll do a swap round and see what happens.

Edit: Have to try that this evening, work calls
 
Last edited:
PCZ said:
Hex
The throttling was very hard to spot.
It was not the normal big drop in speed.

Maybe only 5% or so.
I can see the times in seconds for each Work Unit.
Normal time 345 seconds, throttled 365 seconds so not a large amount.
The increase in processing time is related to CPU temp.

With a lower vcore i haven't observed any slowdowns.

If i didn't use my CPU's for number crunching then i would never have noticed it.


I've got some more RAM here i'll give that a try at 1600MHZ ans see if i need to up the vcore to run it stable.

I am currently using OCZ but i do have GEIL running in some of my other PCs i'll do a swap round and see what happens.
Click to expand...


The throttling you are describing is exactly what few of us on here have had with PWM/MOSFET throttling. It isn't a big drop. What happens is you will find the CPU multi will drop down to say 1600MHz for 4-5 seconds, then go back to full speed again for a while, then drop again for a couple of seconds etc.

The system doesn't crash or error, but you're performance will drop ever so slightly.

I witnessed this in Linx, I was getting ~75GFlops a loop, for the first 3-4 loops, then as the pwm/'fets heated up and started throttling it'd drop to about 72GFlops for 3-4 loops, then as the heat increased it'd drop again to about 69GFlops, which it'd stay at for the remainder of the tests.

The reason it goes away when you lower the Vcore is because you aren't stressing the PWM/'Fets as much, therefore they don't heat up as much and don't cause the throttling.

You are right that it is a heat issue, but it isn't a CPU heat issue (unless you are getting some crazy high temps), it's a power regulator heat issue.

What board are you using? As it seems to be the ones with 4+1 PWM that are struggling.


As for the memory, you don't need to increase Vcore to make 1600MHz stable, if you need to increase anything it'll be the CPU-NB volts, as this is the memory controller.
 
Last edited:
You must log in or register to reply here.
Back
Top Bottom