Is this a safe overclock? 1st Time OCing

[Jonathan Yang]

Finally decided to overclock my rig (Specs are in signature) and I used the FSB method to overclock my rig. At the moment my FSB speed is sitting at 225Mhz.
This overclocks my AMD FX-8350 to 4.5GHz, Corsair Vengeance RAM to 1800MHz (They were rated 1600MHz with XMP Profile), and HT Link to 2925 MHz.

My idle/temp that it is at the time of writing this is 36C on the CPU and 27C on the mobo.
I've run Prime95 for about 10 minutes. It has been successful. Max Temp on there is 51C on the CPU, 30C on the Motherboard and 48C on the cores. (Aida64 software). There has been one where one core failed but that hasn't happened at all after I increased the VCore voltage by the next smallest increment. Which was +.025

My max VCore Voltage that CPU-Z is reporting is 1.366V.

I can't set it to a specific one because my Gigabyte Mobo, only lets me set +0.25, +0.30, etc from the Base VCore value of I believe 1.3V.

Everything seems to be stable but I know that since OCing the FSB overclocks literally all my components, is this still reasonable to say its safe to overclock my FSB to 225MHz without any near/long term risks, without downclocking my RAM.

Also what is a FSB speed that is tooo high and is having HT Link of 2925MHz, too high? Sorry I know more questions are just coming up but I'm new to this and I'm really just curious.

Cause if what I'm doing at the moment isn't a typical/safe overclocking method, I'll switch over to just using the CPU multiplier.



Thanks! Any advice/help/comments would be appreciated!

 

[quilciri]

In general you want to use the CPU multiplier first to overclock the CPU before touching the FSB, if at all. Then you can tweak the FSB/multiplier/memory timing ratio until you find the sweet spot that gives you the highest overclock. Most overclockers with an unlocked CPU never go beyond increasing the multiplier, as you can typically get within 80 or 90% on a modern CPU's max overclock on multiplier alone.

For stability test, 10 minutes is not enough. run prime95 Run the torture test, CPU and memory at least overnight (8-10 hours for each of CPU and memory), 24 hours recommended to be sure. This is important because each iteration of the Lucas-Lehmer depends on the previous one; if one iteration is incorrect, so will be the entire primality test. Errors may show up after 16 or more hours of testing (compared to, say, just four hours of testing)