F5F Stay Refreshed Power Users Overclocking Pentium g3258 OC

Pentium g3258 OC

Pentium g3258 OC

X
xImFizzy
Member
213
01-01-2016, 11:28 AM
#1
Sorry for the English issues. I'm attempting to push my Intel Pentium g3258 to 4.3GHz using a stock cooler. This is my first overclocking attempt. I configured the settings as follows: Clock ratio 43, Core ratio 43, CPU Vcore 1.2v, everything else default. As you can see, it doesn't work. How should I properly overclock this CPU? I'm using a Gigabyte B85M-DS3H-A. Thanks for your help and please don't laugh at me.
X
xImFizzy
01-01-2016, 11:28 AM #1

Sorry for the English issues. I'm attempting to push my Intel Pentium g3258 to 4.3GHz using a stock cooler. This is my first overclocking attempt. I configured the settings as follows: Clock ratio 43, Core ratio 43, CPU Vcore 1.2v, everything else default. As you can see, it doesn't work. How should I properly overclock this CPU? I'm using a Gigabyte B85M-DS3H-A. Thanks for your help and please don't laugh at me.

A
alex_oo3
Member
249
01-01-2016, 08:03 PM
#2
Unlikely because the power consumption gap between Small FFT and Large FFT is minimal. The Large FFT puts more strain on the L3 cache and RAM, which explains the slight increase in power use. Since it remains stable at 3.8, memory usage is probably the main issue. With a dual-core CPU on a capable motherboard, the VRM isn't likely overwhelmed, leaving only the L3 cache as the probable cause. Try reducing the uncore and you should see an improvement in core ratio.
A
alex_oo3
01-01-2016, 08:03 PM #2

Unlikely because the power consumption gap between Small FFT and Large FFT is minimal. The Large FFT puts more strain on the L3 cache and RAM, which explains the slight increase in power use. Since it remains stable at 3.8, memory usage is probably the main issue. With a dual-core CPU on a capable motherboard, the VRM isn't likely overwhelmed, leaving only the L3 cache as the probable cause. Try reducing the uncore and you should see an improvement in core ratio.

S
sonic3003
Member
210
01-03-2016, 12:27 AM
#3
Initially, the uncore shouldn't be raised until a stable core ratio is confirmed. The 4.3GHz cache isn't expected to be fully stable. Begin at 3.8GHz core with stock voltage, adjusting the core in small increments of 0.1GHz until you hit the voltage limit. For higher speeds, raise the voltage in 0.05V steps until stable, ensuring it doesn't exceed around 1.2 on the stock cooler to maintain reasonable temperatures. My G3258 performed well, reaching up to 4.2/3.6/1.2V on stock cooling and 4.5/3.6/1.35 with aftermarket upgrades, though results may differ. Good luck—it's a great chip!
S
sonic3003
01-03-2016, 12:27 AM #3

Initially, the uncore shouldn't be raised until a stable core ratio is confirmed. The 4.3GHz cache isn't expected to be fully stable. Begin at 3.8GHz core with stock voltage, adjusting the core in small increments of 0.1GHz until you hit the voltage limit. For higher speeds, raise the voltage in 0.05V steps until stable, ensuring it doesn't exceed around 1.2 on the stock cooler to maintain reasonable temperatures. My G3258 performed well, reaching up to 4.2/3.6/1.2V on stock cooling and 4.5/3.6/1.35 with aftermarket upgrades, though results may differ. Good luck—it's a great chip!

S
sc0rp1a
Junior Member
32
01-03-2016, 12:46 AM
#4
Initially, the uncore shouldn't be increased until a consistent core ratio is achieved, and a 4.3GHz cache isn't expected to be fully stable. Begin at 3.8GHz core with stock voltage, adjusting in small increments of .1GHz until you hit the maximum voltage limit. For higher speeds, raise the voltage in 0.05V steps until stable, but keep it below about 1.2 on the stock cooler to maintain reasonable temperatures. My G3258 performed well at around 4.2/3.6/1.2V on stock cooling and 4.5/3.6/1.35 with aftermarket upgrades, though results may vary. Good luck—it's a fun little chip! Thanks for the help. Currently, I'm using clock ratio 4.3, uncore 4, 1,175V, and it seems stable. However, I'm unsure about acceptable temperatures since I only experienced around 77°C during stress tests. If I run that test longer, how long should I do it to confirm stability?
S
sc0rp1a
01-03-2016, 12:46 AM #4

Initially, the uncore shouldn't be increased until a consistent core ratio is achieved, and a 4.3GHz cache isn't expected to be fully stable. Begin at 3.8GHz core with stock voltage, adjusting in small increments of .1GHz until you hit the maximum voltage limit. For higher speeds, raise the voltage in 0.05V steps until stable, but keep it below about 1.2 on the stock cooler to maintain reasonable temperatures. My G3258 performed well at around 4.2/3.6/1.2V on stock cooling and 4.5/3.6/1.35 with aftermarket upgrades, though results may vary. Good luck—it's a fun little chip! Thanks for the help. Currently, I'm using clock ratio 4.3, uncore 4, 1,175V, and it seems stable. However, I'm unsure about acceptable temperatures since I only experienced around 77°C during stress tests. If I run that test longer, how long should I do it to confirm stability?

B
61
01-03-2016, 09:35 AM
#5
77 in Prime works well. I plan to test it for 24 hours using blend mode—if it succeeds, you're in good form.
B
BlueBerryDylan
01-03-2016, 09:35 AM #5

77 in Prime works well. I plan to test it for 24 hours using blend mode—if it succeeds, you're in good form.

S
skierboy14
Member
61
01-05-2016, 01:24 PM
#6
77 in Prime is acceptable. I plan to run it for 24 hours using blend mode—if it succeeds, you're in good shape.
I encountered an error: when starting the CPU stress test everything appears normal, but the PSU + RAM (in-place large FFT) test fails after a few seconds. Adjusting CPU settings didn't help consistently. The MemTest with OC reports one error but doesn’t specify what. Could be a PSU issue? I have a 500W PSU and 4GB RAM.
UPDATE: I'm testing at 3,8ghz on 1.1V and it passed the large FFT test for 15 minutes. Is this related to the PSU?
S
skierboy14
01-05-2016, 01:24 PM #6

77 in Prime is acceptable. I plan to run it for 24 hours using blend mode—if it succeeds, you're in good shape.
I encountered an error: when starting the CPU stress test everything appears normal, but the PSU + RAM (in-place large FFT) test fails after a few seconds. Adjusting CPU settings didn't help consistently. The MemTest with OC reports one error but doesn’t specify what. Could be a PSU issue? I have a 500W PSU and 4GB RAM.
UPDATE: I'm testing at 3,8ghz on 1.1V and it passed the large FFT test for 15 minutes. Is this related to the PSU?

J
jaroke25
Member
68
01-06-2016, 06:45 AM
#7
Unlikely because the power consumption gap between Small FFT and Large FFT is minimal. The Large FFT puts more strain on the L3 cache and RAM, which explains the slight increase in power use. Since it remains stable at 3.8, memory usage is probably the main issue. With a dual-core CPU on a motherboard supporting quads, the VRM isn't likely overwhelmed, leaving only the L3 cache as the probable cause. Try reducing the uncore and you might improve the core ratio again.
J
jaroke25
01-06-2016, 06:45 AM #7

Unlikely because the power consumption gap between Small FFT and Large FFT is minimal. The Large FFT puts more strain on the L3 cache and RAM, which explains the slight increase in power use. Since it remains stable at 3.8, memory usage is probably the main issue. With a dual-core CPU on a motherboard supporting quads, the VRM isn't likely overwhelmed, leaving only the L3 cache as the probable cause. Try reducing the uncore and you might improve the core ratio again.

J
julkakuk
Junior Member
14
01-06-2016, 07:59 AM
#8
Yes that helped, thanks a lot.
J
julkakuk
01-06-2016, 07:59 AM #8

Yes that helped, thanks a lot.