problem with adaptive voltage+ Gameing + Real Bench

[Legionarius13]

hi all, first time overclock done after a week
4770k @ 4.4ghz 1.30 vcore..... ringbuss 3.9 auto Voltage
CPU AGENT +0.200
CPU ANALOG +DIGITAL +1.50
INPUT VOLTAGE 1.850
2133 XMP
running fine 10 hours X264 10X Real Bench
AIDA 64 bench test + 2 hours stress test
had voltage set to manual
l
change it to adapitve after watching JJ from Asus saying its fine if you dont use synthetic stess test like ADIA 64 + Prima95

but now getting v-core of 1.360 and core VID 1.360

i run a bench mark on Rome 2 totalwar V-core 0- 1.312 .... VID CORE1 - 1.361
why is this?
First screen Real Bench
Second rome 2 totalwar

http://imgur.com/B14LRB5,NOB1Z68#0
http://imgur.com/B14LRB5,NOB1Z68#1

thinking of going back to manual so dont get big voltage spikes

 

[ihog]

Why? Because adaptive/auto voltages on motherboards are inaccurate and generally suck. I always use manual voltage, and I recommend you do so, too.

 

[Legionarius13]

Yeh I've set mine back to manual mode 1.30 with adaptive was getting 1.321 a lot over my setting any way to fix this? Is it bad for CPU to have that voltage all the time

 

[ihog]

Bad? Not necessarily, but your CPU's life will definitely be shortened. Will you ever notice or upgrade before the CPU kicks the bucket? Maybe not.

I personally have SpeedStep enabled so that the CPU clocks down when it isn't in use and then ramps up to the manual voltage and frequency I set when necessary.

 

[Legionarius13]

so whats the best option to fix this?

 

[ihog]

Fixing what?

 

[Legionarius13]

whats speed step?

 

[ihog]

http://www.intel.com/cd/channel/reseller/asmo-na/eng/203838.htm

Enhanced Intel SpeedStep Technology allows the system to dynamically adjust processor voltage and core frequency, which can result in decreased average power consumption and decreased average heat production.

The option to enable or disable it should be in your bios.

 

[Baralis]

hi all, first time overclock done after a week
4770k @ 4.4ghz 1.30 vcore..... ringbuss 3.9 auto Voltage
CPU AGENT +0.200
CPU ANALOG +DIGITAL +1.50
INPUT VOLTAGE 1.850
2133 XMP
running fine 10 hours X264 10X Real Bench
AIDA 64 bench test + 2 hours stress test
had voltage set to manual
l
change it to adapitve after watching JJ from Asus saying its fine if you dont use synthetic stess test like ADIA 64 + Prima95

but now getting v-core of 1.360 and core VID 1.360

i run a bench mark on Rome 2 totalwar V-core 0- 1.312 .... VID CORE1 - 1.361
why is this?
First screen Real Bench
Second rome 2 totalwar

http://imgur.com/B14LRB5,NOB1Z68#0
http://imgur.com/B14LRB5,NOB1Z68#1

thinking of going back to manual so dont get big voltage spikes

Any program that uses AVX instructions will override voltage settings by as much as 0.1V if using any voltage settings other than manual. This is by design and normal but one must be aware of it particularly when OCing.

 

[ihog]

Thanks for the insight, Baralis. I did not know that.

 

[Legionarius13]

so do games use AVX instructions?

 

[Baralis]

so do games use AVX instructions?

Yes some games do take advantage of AVX instructions.

 

[Legionarius13]

so stay away from real bench could 1.31 be bad to run my CPU all the time?

 

[commandosupremo]

Why would you run real bench all the time anyways? I highly doubt that most gaming or real world applications utilize AVX to the same extent that stress testing programs do. If you've used prime95 you will quickly note that it gets the cpu hotter than just about anything else out there. I wouldn't say avoid it unless you stay with adaptive and you don't have the little bit of temperature or voltage overhead to handle the extra bit of voltage and temperature the additional draw will give you.

If this is a concern to you, switching to manual solves the problem with few and minor disadvantages over adaptive.

I dialed in my overclock manually and switched to adaptive because I could tell an ever so slightly less power consumption and temperature output but I made sure to leave enough headroom on the voltage that if it decides to draw that extra .1V for AVX that I'm not in the danger zone. I also wasn't going for extreme OC either though, I set a frequency target, reached it and stopped. So our goals may be different.