AMD Overclocking Club

Page 23 - Seeking answers? Join the Tom's Hardware community: where nearly two million members share solutions and discuss the latest tech.
Toggle sidebar Toggle sidebar

LOL alright ill be sure to do that! give me a few hours for prime95 testing (also sense im a noob i have never used prime 95...wish me luck!)
 

yes the voltage was increased to just under 1.5 so its 1.4 something if remember correctly (yes i know thats a lot of power! but it keeps it stable! i think its stock 1.3 i also have the phenom ii 1100t stock cooler on it (i upgraded to the six core gave my bro the dualcore)
 
prime95 blend test is good for testing memory
while the large ffts is good for checking temp max
if it blue screens then usually it is something related to memory
if it just shuts down then it is usually cpu temps or not enough voltage
make sure if you are playing with voltage to google your cpu and see what safe voltages are
need to do some research
also might help to down clock your ram in the beginning
1333-1066
1066-800 etc
then when you are stable raise mem speed back up to stock speed and timings
would need to set your timings in the bios for the stock it came with

while running P95 constantly check temps
after about a half hour you should be close to max
about two hours with p95 running and you are pretty stable
during testing if one worker in P95 has an error and the others are okay
that is probably memory related especially if blend testing
I wouild leave NB and HT set to auto and not mess with their voltages
also disable CoolNQuiet and C1E in BIOS
later on can try putting CoolNQuiet back on and see if it stays stable

I am still learning so hopefully one of the Veteran OCers can help out
and double check my advice
 
Actually if a single worker stops, and it happens to that exact worker more than once related to a specific core, you may have a core that is... weaker than the others. Generally upping the voltage on the cpu will help that core individually.

Said core can also error out because it is more sensitive than the others to any memory, nb or other issues. To troubleshoot, back off the memory then test. Then back off the NB and re-test and so on. Usually it just comes down to a core that does not perform as well as the others.
 



I had that issue myself
worker 4 would keep crashing
but the other three were fine
that was on Prime95 64bit version

I had my memory at the stock 1333 but the timings were at 8-8-8-20 due to OC
once I manually set my timings to 9-9-9-24 stock
then it was fine

but also I use Patriot Signature series ram which is decent but not OCing ram
 
i am wondering that if we can get 5ghz+ at a single core (only for fun), i am not considering performance of pc, just want to know that if anyone can hit above 5+ on at least 1 core at normal cooling
(and thus ultimately want to know the oc limit of a thuban core)

@saint19
seems that the link is not working,
agree, those 2 cores can mean 7g in highly thread task

@garage1217
lol, 1100t at 7g@2v will use around 0.6kw only

(disabling (underclocking) 2 cores can improve oc on other cores (with stock heatsink) and thus can improve performance in less treaded task (by assigning oc'ed core to that task) and thus we can get more performance and oc within safe temp limit with stock heat sink.)
 
😱
6.5g, and that too with all core active, i think it can hit 8 with 4 cores, and 10 with 1 core active individually if that cpu can survive 2v
have you done any multithreaded benchmark at that clock (because i think that can beat daneb at 7g by 25percent more score)
seems like you are handling liquid air

but i am unable to understand to understand this *Submitted by saint19 | Sat, 09 Jul 2011 21:59:15 +0200 | Rejected by CPU-Z 1.58*
what does they mean by rejected by cpu-z 1.58?
 


Phenom II or generally, Thuban, Deneb core and other series can't hit the 8GHz even if you disable cores. Some can hit the 7GHz with luck, the only line that's hitting the 8GHz on AMD is Bulldozer but all we know that performance is bad, very bad...

Those 6.5GHz was only for validation, I did SuperPI and wPrime at 6.2GHz and 6.3GHz respectively. AMD isn't fully stable on frequency, I mean, you always see that your CPU frequency change some little megahertz while Intel doesn't. CPU-Z reject the frequency by that problem, but few weeks ago they announced that some of those rejected validation are already valid, including mine....

And finally, the max with air was 4.6GHz, those 6.5GHz was with liquid nitrogen at -190ºC.
 
yes, speed of amd cpu fluctuates

'those
6.5GHz was with liquid nitrogen at -190ºC.' [:truegenius:2] I was suspicious, thatr why i talked about liquid air.

And have you tried 3dmark to check that where it stands in ranking at 6.5g
(i think it will be/is at No.1)
 


Not yet, but soon I will do it....again...
 
CPU: B55
Name: Deneb
Stepping: RB-C3
Frequency: 3811.2mhz
Ref* Mulitiplier: 206.03 * 18.5
CPU Voltage: 1.432v
NB Frequency: 2060Mhz
NB-Volts- Auto
RAM: 8GB 1333 overclocked 1400Mhz DDR3 9-9-9-24 1T CL9
Motherboard: MSI 770 - G45
Cooling: Cossair H100 Watercooling 4*120MM Fans Push/Pull

I cannot get a temp reading since my cores are unlocked.

2213509.png
 
CPU: AMD Phenom II x4 960T Black 3.0GHz AM3+ Unlocked with 1/6 bad core (#3) which turns it to a Phenom II x5 1600T Thuban
Name: Heka? Not totally sure.
Stepping: 0
Frequency: 4026.8 @ 1.45V
Ref*Multiplier: 201.33*20
CPU-NB V: 1.225V
CPU-NB: 201.33*13 @ 2617.4
NB V: 1.110V
RAM: Mushkin DDR3-1600 - 7-8-7-24 @ 1T @ 1.65V (200x8)
HT Link/Ref: Auto x10 @ 2000MHz peaking at 2617.9
MoBo: Gigabyte 990FXA-UD3 ATX
Cooling: Front, Rear 120mm, Top 200mm, Push/Pull 120mm on Hyper 212+
OS: Windows 7 Ultimate 64-bit
GFLOPS: N/A
Case: TT V9

amdoverclock40265.jpg

show_oc.php


Any suggestions?
I haven't tried 4.0GHz with any higher NB. I have had the NB as high as 3k, but the CPU Freq. had always been at unstable speeds when testing. Will update further after a few weeks.
Any way to have P95 produce results and log files somewhere other than the desktop?
Also, CPU-Z recognizes my Memory Modules as being in Slot 3 and Slot 4. However, according to the Gigabyte manual, the RAM Mods are in DDR3_1 and DDR3_2. Should I give it any attention? I only ask because, according to the manual, DDR3_1 is the furthest from the CPU, and DDR3_2 is second closest to the CPU. Shouldn't the RAM be closest as possible to the CPU?

Thanks!
 


I merged all worker windows. You can see Worker #0-4 within the bottom window.
 
You must log in or register to reply here.

TRENDING THREADS

Latest posts

Moderators online

Share this page

COMPANY
RESOURCES
FOLLOW
Top Bottom