F5F Stay Refreshed Power Users Overclocking Safe OC for i5 4690K?

Safe OC for i5 4690K?

Safe OC for i5 4690K?

L
LissieBear
Member
163
12-11-2016, 09:38 PM
#1
I’ve begun adjusting my 4690K in the past, but I noticed it would frequently crash in games, especially GTAV. Back then, it was running at about 4.5 to 4.6Ghz with AUTO voltage settings, since I’m still learning.
How could I achieve more stability?
After that, I mostly kept the processor in turbo mode.
L
LissieBear
12-11-2016, 09:38 PM #1

I’ve begun adjusting my 4690K in the past, but I noticed it would frequently crash in games, especially GTAV. Back then, it was running at about 4.5 to 4.6Ghz with AUTO voltage settings, since I’m still learning.
How could I achieve more stability?
After that, I mostly kept the processor in turbo mode.

A
AGLOS6
Member
184
12-18-2016, 07:34 PM
#2
It's more reliable to adjust the vcore manually. Auto tends to raise core voltage unnecessarily to ensure stability, since each CPU is unique. I keep the speed step feature on. For continuous operation, aim to keep the core voltage under 1.3v whenever possible, and treat 1.35v as the upper limit. Higher voltages generate more heat. If you can sustain 4.5ghz at 1.3v without issues, sticking at 1.35v isn't necessary. The additional 0.05v represents wasted energy and heat. Begin around 1.2 or 1.25v and gradually increase the multiplier step by step. For example, try x42 (4.2ghz). Use tools like p95 version 26.6 with modest FFTs (steady load) and track temperatures with realtemp. If readings stay within safe limits—roughly 80-85°C—they should be fine for p95 (normal conditions).
A
AGLOS6
12-18-2016, 07:34 PM #2

It's more reliable to adjust the vcore manually. Auto tends to raise core voltage unnecessarily to ensure stability, since each CPU is unique. I keep the speed step feature on. For continuous operation, aim to keep the core voltage under 1.3v whenever possible, and treat 1.35v as the upper limit. Higher voltages generate more heat. If you can sustain 4.5ghz at 1.3v without issues, sticking at 1.35v isn't necessary. The additional 0.05v represents wasted energy and heat. Begin around 1.2 or 1.25v and gradually increase the multiplier step by step. For example, try x42 (4.2ghz). Use tools like p95 version 26.6 with modest FFTs (steady load) and track temperatures with realtemp. If readings stay within safe limits—roughly 80-85°C—they should be fine for p95 (normal conditions).

J
jayde0284
Junior Member
36
12-20-2016, 01:39 PM
#3
I understand you're asking about the cause of a crash, possibly due to overheating. Could you provide more details about the device or situation you're referring to? This will help in giving a more accurate explanation.
J
jayde0284
12-20-2016, 01:39 PM #3

I understand you're asking about the cause of a crash, possibly due to overheating. Could you provide more details about the device or situation you're referring to? This will help in giving a more accurate explanation.

_
_Dirty_
Member
163
12-20-2016, 08:15 PM
#4
It’s smarter to adjust the vcore manually. The auto setting often raises core voltage too much to ensure stability, since each CPU behaves differently. I keep the speed step feature turned on. For continuous use, aim to keep the core voltage under 1.3v whenever possible, and treat 1.35v as the upper limit. Higher voltages generate more heat. If you manage to run at 4.5ghz with 1.3v without issues, sticking at 1.35v isn’t necessary. The additional 0.05v simply adds unnecessary power and heat.

Begin around 1.2 or 1.25v and gradually increase the multiplier a bit at a time. Try x42 (4.2ghz). Use tools like p95 version 26.6 with small FFTs (steady load) and watch realtemp. If it stays safe—around 80-85°C for p95 under normal loads—proceed. If it reaches 85-90°C, you’ll likely need a better cooler, improved case cooling, or an air conditioner to lower room temperature. Test stability with software such as rog realbench. If everything holds, increase the multiplier slightly to x43 or x44, then recheck with p95 and realbench. Keep repeating.

If crashes occur, consider raising the core voltage slightly to 1.26 or 1.27, then retest. If that fails, go up a bit to 1.28 or 1.3v. If crashes persist, you may need to reduce the overclock. The goal is to maintain temperatures and voltage within safe limits while ensuring stability. Once you achieve stable performance at p95, realbench, or similar benchmarks, you can gradually lower the core voltage and reassess.

If you encounter overheating, try increasing the core voltage a touch to 1.26 or 1.27 and retest. If crashes continue, further adjustments might be necessary. The aim is to balance performance with thermal management.

P95 is extremely demanding—it forces the CPU to work harder than any other workload, whether it’s gaming or video editing. This test simulates the worst-case scenario, which is ideal for assessing limits. If your CPU and cooling system can handle such stress, you’re in a good position; otherwise, you might face performance drops.

For CPUs like the 4690k, reaching 4.7ghz at 1.27v is possible, but crashing above 4.4ghz is likely. My setup runs at 4.5ghz with 1.28v and I reduced it slightly due to high ambient temperatures (around 91°F). Typical cooler testing environments are usually around 22°C (71-72°F), while my current room temp is nearly 33°C (91°F).
_
_Dirty_
12-20-2016, 08:15 PM #4

It’s smarter to adjust the vcore manually. The auto setting often raises core voltage too much to ensure stability, since each CPU behaves differently. I keep the speed step feature turned on. For continuous use, aim to keep the core voltage under 1.3v whenever possible, and treat 1.35v as the upper limit. Higher voltages generate more heat. If you manage to run at 4.5ghz with 1.3v without issues, sticking at 1.35v isn’t necessary. The additional 0.05v simply adds unnecessary power and heat.

Begin around 1.2 or 1.25v and gradually increase the multiplier a bit at a time. Try x42 (4.2ghz). Use tools like p95 version 26.6 with small FFTs (steady load) and watch realtemp. If it stays safe—around 80-85°C for p95 under normal loads—proceed. If it reaches 85-90°C, you’ll likely need a better cooler, improved case cooling, or an air conditioner to lower room temperature. Test stability with software such as rog realbench. If everything holds, increase the multiplier slightly to x43 or x44, then recheck with p95 and realbench. Keep repeating.

If crashes occur, consider raising the core voltage slightly to 1.26 or 1.27, then retest. If that fails, go up a bit to 1.28 or 1.3v. If crashes persist, you may need to reduce the overclock. The goal is to maintain temperatures and voltage within safe limits while ensuring stability. Once you achieve stable performance at p95, realbench, or similar benchmarks, you can gradually lower the core voltage and reassess.

If you encounter overheating, try increasing the core voltage a touch to 1.26 or 1.27 and retest. If crashes continue, further adjustments might be necessary. The aim is to balance performance with thermal management.

P95 is extremely demanding—it forces the CPU to work harder than any other workload, whether it’s gaming or video editing. This test simulates the worst-case scenario, which is ideal for assessing limits. If your CPU and cooling system can handle such stress, you’re in a good position; otherwise, you might face performance drops.

For CPUs like the 4690k, reaching 4.7ghz at 1.27v is possible, but crashing above 4.4ghz is likely. My setup runs at 4.5ghz with 1.28v and I reduced it slightly due to high ambient temperatures (around 91°F). Typical cooler testing environments are usually around 22°C (71-72°F), while my current room temp is nearly 33°C (91°F).