F5F Stay Refreshed Power Users Overclocking phenom II 965 operates at 4.0ghz and seeks to advance further.

phenom II 965 operates at 4.0ghz and seeks to advance further.

phenom II 965 operates at 4.0ghz and seeks to advance further.

Pages (2): Previous 1 2
C
CobbleWalker
Member
238
03-11-2024, 11:08 PM
#11
My CPU is a quad-core stock unit. I'm wondering if you're mixing up the models. But thank you for your responses.
I decided to begin with a stock configuration I found online, which was incorrect.
I adjusted the multiplier and increased the voltage during crashes. I also set the memory speed to 1333mhz.
Currently, I'm running at around 4.0ghz with a voltage of 1.4125, and I'll keep you posted.
It crashed recently, reaching up to 1.425.
Later I reported it crashed again at around 1.4375 volts.
C
CobbleWalker
03-11-2024, 11:08 PM #11

My CPU is a quad-core stock unit. I'm wondering if you're mixing up the models. But thank you for your responses.
I decided to begin with a stock configuration I found online, which was incorrect.
I adjusted the multiplier and increased the voltage during crashes. I also set the memory speed to 1333mhz.
Currently, I'm running at around 4.0ghz with a voltage of 1.4125, and I'll keep you posted.
It crashed recently, reaching up to 1.425.
Later I reported it crashed again at around 1.4375 volts.

Y
Yuukimaruchan
Junior Member
48
03-11-2024, 11:08 PM
#12
my specifications processor: phenom II 965 be cooler: zalman 9900 led ram: kingston hyperx blue 1600 (moved xmp profile, currently running at 1.65v) motherboard: asus m5a97 graphics card: gtx 760 power unit: corsair 650w case: zalman z9+ ventilation: 2 x 120mm fans at top blowing out, 1 x 120mm fan at rear blowing out, 2 x 120mm fans on sides blowing in, 2 x 120mm fans front blowing in overclock settings: cpu at 4ghz with multiplier applied northbridge frequency: 2600 mhz ht link: 2400 mhz cpu spread disabled cooling controls: disable nquit, disable llc cpu voltage: 1.468v nb voltage: 1.3v dram voltage: 1.65v nb voltage: 1.2v thermal readings: 47°C during prime 95 blend test aiming for 4.2°C, but if i set cpu voltage to 1.55v i might sustain prime 95 for about 30 minutes before freezing system remains stable without blue screen or crash, only freezes please advise on the best path forward. the main issue here is that, being about 7 years old, your AMD 965 BE overclock is constrained by heat. You can push further if you manage the temperature better! Do you have a budget for a new cooler? I noticed you lowered the memory speed from 1600mhz to 1333mhz, which helped reduce heat. Consider manually adjusting these settings and testing at 1333mhz with 9,9,9,24 at 2T or 2N speed, setting DDR3 voltage to 1.50v. If you prefer not to upgrade your cooler, you’re limited to fine-tuning the overclock until thermal limits are reached. A 4.2ghz overclock was common in 2010, one of the models I referenced for my guide. You still have that CPU working fine for your DJ setup. I once hit 4.3ghz on air cooling, but I couldn’t confirm it, as my next CPU used a 2500K Intel. At the time of the 4.2ghz run, I used Thermalright TRUE with 2 x 110cfm 120mm fans. Back then, CoreTemp reported temperature accurately, while CPU-Z matched the BIOS voltage settings (hope this is true).
Y
Yuukimaruchan
03-11-2024, 11:08 PM #12

my specifications processor: phenom II 965 be cooler: zalman 9900 led ram: kingston hyperx blue 1600 (moved xmp profile, currently running at 1.65v) motherboard: asus m5a97 graphics card: gtx 760 power unit: corsair 650w case: zalman z9+ ventilation: 2 x 120mm fans at top blowing out, 1 x 120mm fan at rear blowing out, 2 x 120mm fans on sides blowing in, 2 x 120mm fans front blowing in overclock settings: cpu at 4ghz with multiplier applied northbridge frequency: 2600 mhz ht link: 2400 mhz cpu spread disabled cooling controls: disable nquit, disable llc cpu voltage: 1.468v nb voltage: 1.3v dram voltage: 1.65v nb voltage: 1.2v thermal readings: 47°C during prime 95 blend test aiming for 4.2°C, but if i set cpu voltage to 1.55v i might sustain prime 95 for about 30 minutes before freezing system remains stable without blue screen or crash, only freezes please advise on the best path forward. the main issue here is that, being about 7 years old, your AMD 965 BE overclock is constrained by heat. You can push further if you manage the temperature better! Do you have a budget for a new cooler? I noticed you lowered the memory speed from 1600mhz to 1333mhz, which helped reduce heat. Consider manually adjusting these settings and testing at 1333mhz with 9,9,9,24 at 2T or 2N speed, setting DDR3 voltage to 1.50v. If you prefer not to upgrade your cooler, you’re limited to fine-tuning the overclock until thermal limits are reached. A 4.2ghz overclock was common in 2010, one of the models I referenced for my guide. You still have that CPU working fine for your DJ setup. I once hit 4.3ghz on air cooling, but I couldn’t confirm it, as my next CPU used a 2500K Intel. At the time of the 4.2ghz run, I used Thermalright TRUE with 2 x 110cfm 120mm fans. Back then, CoreTemp reported temperature accurately, while CPU-Z matched the BIOS voltage settings (hope this is true).

M
MrBrown12344
Member
124
03-11-2024, 11:08 PM
#13
Here are some quick details to keep you motivated.
My Phenom ii boards are both Deneb cores, identical to the ones in your system. During production, all Phenom ii chips were uniform. AMD grouped them by performance, creating lower-end 2-core models with slower clocks and top-tier quad-core fast CPUs. Yet, by the end of production, engineers significantly enhanced quality, ensuring nearly every CPU had four functional cores and high-speed features. The marketing team still aimed to offer 2-core options for budget users, so engineers deliberately underclocked and disabled two cores. That’s where my CPUs originated.

Both of my CPUs were initially sold as 3.2GHz dual-core units. I unlocked and boosted them, achieving speeds up to 4GHz on one and 3.45GHz on the other. The only difference between them was clock speed, which I managed to be slightly better on one.

AMD never released a Phenom II model exceeding 3.7GHz, and many online users were thrilled to reach 3.9GHz—so hitting 4GHz is definitely worth pursuing.

When overclocking an AMD Phenom, it’s essential to identify your peak CPU speed, northbridge frequency, and RAM speed. Each contributes to overall performance, but the CPU speed has the biggest impact. On my 4GHz system, lowering the northbridge frequency improved CPU responsiveness. Still, reducing RAM speed didn’t yield further gains.

On the other machine, running at three-core mode lets it operate smoothly at 3.7GHz, but core 4 struggles and causes crashes at 3.5GHz.

Experiment to discover your own performance limits.
M
MrBrown12344
03-11-2024, 11:08 PM #13

Here are some quick details to keep you motivated.
My Phenom ii boards are both Deneb cores, identical to the ones in your system. During production, all Phenom ii chips were uniform. AMD grouped them by performance, creating lower-end 2-core models with slower clocks and top-tier quad-core fast CPUs. Yet, by the end of production, engineers significantly enhanced quality, ensuring nearly every CPU had four functional cores and high-speed features. The marketing team still aimed to offer 2-core options for budget users, so engineers deliberately underclocked and disabled two cores. That’s where my CPUs originated.

Both of my CPUs were initially sold as 3.2GHz dual-core units. I unlocked and boosted them, achieving speeds up to 4GHz on one and 3.45GHz on the other. The only difference between them was clock speed, which I managed to be slightly better on one.

AMD never released a Phenom II model exceeding 3.7GHz, and many online users were thrilled to reach 3.9GHz—so hitting 4GHz is definitely worth pursuing.

When overclocking an AMD Phenom, it’s essential to identify your peak CPU speed, northbridge frequency, and RAM speed. Each contributes to overall performance, but the CPU speed has the biggest impact. On my 4GHz system, lowering the northbridge frequency improved CPU responsiveness. Still, reducing RAM speed didn’t yield further gains.

On the other machine, running at three-core mode lets it operate smoothly at 3.7GHz, but core 4 struggles and causes crashes at 3.5GHz.

Experiment to discover your own performance limits.

T
TheSimple
Member
229
03-11-2024, 11:08 PM
#14
Thanks! I'm planning to focus on overclocking just the CPU again. It's running smoothly on Prime 95 blend for two hours at 4GHz, with a max temperature of 46°C. I'll try pushing it to 4.1GHz next. If that doesn't work, I'll move on to adjusting the RAM settings.
T
TheSimple
03-11-2024, 11:08 PM #14

Thanks! I'm planning to focus on overclocking just the CPU again. It's running smoothly on Prime 95 blend for two hours at 4GHz, with a max temperature of 46°C. I'll try pushing it to 4.1GHz next. If that doesn't work, I'll move on to adjusting the RAM settings.

S
SoyDash
Posting Freak
859
03-11-2024, 11:08 PM
#15
You're currently running at 4.1 GHz with a voltage of 1.475V, but you're encountering unusual prime 95 errors during the blend test. You're considering adjusting your RAM settings to see if that resolves the issue.

Additionally, the first core experienced a hardware failure after about 15 minutes of the blend test. You increased the CPU voltage to 1.48V and retested.

The blend test later froze after about an hour, stopping completely without errors. You then increased the voltage slightly to 1.4875V and retested.

One hour and ten minutes into the blend test, core 1 started showing errors again. You raised the CPU voltage to 1.5V and are now running the blend test again. Any other suggestions would be appreciated.
S
SoyDash
03-11-2024, 11:08 PM #15

You're currently running at 4.1 GHz with a voltage of 1.475V, but you're encountering unusual prime 95 errors during the blend test. You're considering adjusting your RAM settings to see if that resolves the issue.

Additionally, the first core experienced a hardware failure after about 15 minutes of the blend test. You increased the CPU voltage to 1.48V and retested.

The blend test later froze after about an hour, stopping completely without errors. You then increased the voltage slightly to 1.4875V and retested.

One hour and ten minutes into the blend test, core 1 started showing errors again. You raised the CPU voltage to 1.5V and are now running the blend test again. Any other suggestions would be appreciated.

X
XxKripxDeMoNxX
Senior Member
536
03-11-2024, 11:08 PM
#16
Even at 1.5v the system stalled. It's unclear what to do next. Could adjusting the NB and HT make a difference?
X
XxKripxDeMoNxX
03-11-2024, 11:08 PM #16

Even at 1.5v the system stalled. It's unclear what to do next. Could adjusting the NB and HT make a difference?

C
COLIN20052012
Posting Freak
857
03-11-2024, 11:08 PM
#17
Test 1, assess how quickly your CPU performs with all other components and memory at 1333. Don't just rely on the multiplier, consider using the HT bus as well.
Avoid raising the voltage to 1.5V – it has minimal impact.
Test 2, evaluate the speed of your northbridge with all components and memory at 1333.
Test 3, determine the speed of your memory with all other components connected – set memory to 1600, then boost the HT bus for even greater performance.
Your goal is to discover the optimal balance by overclocking everything together for maximum efficiency.
C
COLIN20052012
03-11-2024, 11:08 PM #17

Test 1, assess how quickly your CPU performs with all other components and memory at 1333. Don't just rely on the multiplier, consider using the HT bus as well.
Avoid raising the voltage to 1.5V – it has minimal impact.
Test 2, evaluate the speed of your northbridge with all components and memory at 1333.
Test 3, determine the speed of your memory with all other components connected – set memory to 1600, then boost the HT bus for even greater performance.
Your goal is to discover the optimal balance by overclocking everything together for maximum efficiency.

R
Redcloud_Ninja
Junior Member
4
03-11-2024, 11:08 PM
#18
Lodders :
Test 1, observe the CPU speed when all components are stocked and memory is at 1333. Avoid only using the multiplier; also consider the HT bus.
Don't increase the voltage to 1.5V – it has minimal impact.
Test 2, check the northbridge speed with stock components and memory at 1333.
Test 3, see how fast the memory runs with stock components – set it to 1600, then boost the HT bus for higher speed.
Your goal is to discover the optimal balance by overclocking all parts together for maximum performance.
So far I’ve used the multiplier and achieved 4 GHz with 1.4678V and 1333, but trying 4.1 GHz at 1.5V causes the system to freeze during Prime95 for 10–30 minutes.
Then I set CPU at 4 GHz, voltage at 1.4678V, increased northbridge to 2800 MHz, and set CPU/voltage to 1.3V. I ran a Prime95 blend test for 8 hours without crashes or errors, with the highest temperature reaching 51°C.
I also increased RAM to its XMP profile (manually adjusted) to 1600 MHz and 1.65V.
That’s the current configuration. It seems adding more MHz to the CPU will require significantly higher voltage, possibly around 1.55V or more.
R
Redcloud_Ninja
03-11-2024, 11:08 PM #18

Lodders :
Test 1, observe the CPU speed when all components are stocked and memory is at 1333. Avoid only using the multiplier; also consider the HT bus.
Don't increase the voltage to 1.5V – it has minimal impact.
Test 2, check the northbridge speed with stock components and memory at 1333.
Test 3, see how fast the memory runs with stock components – set it to 1600, then boost the HT bus for higher speed.
Your goal is to discover the optimal balance by overclocking all parts together for maximum performance.
So far I’ve used the multiplier and achieved 4 GHz with 1.4678V and 1333, but trying 4.1 GHz at 1.5V causes the system to freeze during Prime95 for 10–30 minutes.
Then I set CPU at 4 GHz, voltage at 1.4678V, increased northbridge to 2800 MHz, and set CPU/voltage to 1.3V. I ran a Prime95 blend test for 8 hours without crashes or errors, with the highest temperature reaching 51°C.
I also increased RAM to its XMP profile (manually adjusted) to 1600 MHz and 1.65V.
That’s the current configuration. It seems adding more MHz to the CPU will require significantly higher voltage, possibly around 1.55V or more.

T
Toodaloo_246
Senior Member
439
03-11-2024, 11:08 PM
#19
I think around 4 Ghz is the practical upper limit for your CPU. If it were mine, I'd set it to 3960Mhz (multiplier 18 and HT bus at 220). This would provide a stability buffer and let you lower the CPU voltage slightly. It would also run your memory at 1760Mhz, offering better performance than 1600 during tests. I'd configure the northbridge at x12, or x13 if stability is confirmed. Good luck!
T
Toodaloo_246
03-11-2024, 11:08 PM #19

I think around 4 Ghz is the practical upper limit for your CPU. If it were mine, I'd set it to 3960Mhz (multiplier 18 and HT bus at 220). This would provide a stability buffer and let you lower the CPU voltage slightly. It would also run your memory at 1760Mhz, offering better performance than 1600 during tests. I'd configure the northbridge at x12, or x13 if stability is confirmed. Good luck!

Pages (2): Previous 1 2