FX 6300 Overclock

[shea99]

So im running the FX 6300 and ive tried an oc of 4.6ghz with a Hyper 212 Evo with 2 AF120's on it.

So when stress testing at a multiplier of 22 and a bus freq of 210.

Also the cpu manual voltage is 1.4v and cpu/nb is 1.25v.

So when stress testing with prime 95 after about an hour the 6th and 6th core seem to crash and fail, however the rest of the cores where fine with further testing of 12 plus hours. Im fairly new to oc'ing so some help would be appreciated.

Oh and the temp was 63c under fullload so have some head room to about 75 or would you recomend not going that hot on a 6300

Many thanks 

 

[CDHoward]

Don't let temps get to 70-75c. In any case, your FX will start throttling itself at that point.

 

I don't know what mobo you have, but try to get a stable 4.5 MHz overclock and be glad of that.

 

My 8350 is oc'ed to 4.5 MHz on an Asus M5A97 Evo R2.0 mobo, and cooler master T4. I can't get it higher than that without instability and too much heat.

 

For the best bios settings to oc your 6300 on your particular mobo, Google it. You should come across a forum or two with exact instructions and pics.

[LienusLateTips]

What mobo

[shea99]

6 minutes ago, CDHoward said:

Don't let temps get to 70-75c. In any case, your FX will start throttling itself at that point.

 

I don't know what mobo you have, but try to get a stable 4.5 MHz overclock and be glad of that.

 

My 8350 is oc'ed to 4.5 MHz on an Asus M5A97 Evo R2.0 mobo, and cooler master T4. I can't get it higher than that without instability and too much heat.

 

For the best bios settings to oc your 6300 on your particular mobo, Google it. You should come across a forum or two with exact instructions and pics.

managed to get a stable 4.6mhz oc does run a little hot under 100% load. After 3 hours testing it was around 72c consistantly 

I uped the cpu voltage downed the bus to 200 ( original speed) and then set 23 on the multiplier.

 

[shea99]

1 minute ago, JDE said:

What mobo

Asus m5a99x evo r 2

[shea99]

 

49 minutes ago, Strike105X said:

Is 72c the the core voltage or the socket temp? If its the later then it fine. Just as a heads up fx6300 cpu's are the worst overclockers in the fx lineup.

bit noobish but its the cpu reading at the bottom of core temp so im guessing socket temp ???

[shea99]

12 minutes ago, shea99 said:

 

bit noobish but its the cpu reading at the bottom of core temp so im guessing socket temp ???

I will post a pic later while under a stress test 

[miagisan]

2 hours ago, Strike105X said:

Is 72c the the core voltage or the socket temp? If its the later then it fine. Just as a heads up fx6300 cpu's are the worst overclockers in the fx lineup.

 

This is actually a missconception FX processors don't throttle unless:

 

a. The power phase can't keep up with the energy demand at higher frequencies.

 

b. You have activated some energy efficiency options in bios (like c-state or cool and quiet to give as an example).

 

The only protection the FX CPU has its to turn itself off when reaching 90c, because at 90c the glue that holds the metal lid is at risk to start and melt.

I always thought the 6300 was the best of the overclockers. Huh. 

 

 

Anyways got mine to 4.6 at 1.412v, package gets to 62c using a hyper 212 evo

[CDHoward]

3 hours ago, Strike105X said:

Is 72c the the core voltage or the socket temp? If its the later then it fine. Just as a heads up fx6300 cpu's are the worst overclockers in the fx lineup.

 

This is actually a missconception FX processors don't throttle unless:

 

a. The power phase can't keep up with the energy demand at higher frequencies.

 

b. You have activated some energy efficiency options in bios (like c-state or cool and quiet to give as an example).

 

The only protection the FX CPU has its to turn itself off when reaching 90c, because at 90c the glue that holds the metal lid is at risk to start and melt.

RE Your last paragraph: Interesting. I didn't know about that. The FX is hardier than I thought, then, maybe.

[miagisan]

16 minutes ago, CDHoward said:

RE Your last paragraph: Interesting. I didn't know about that. The FX is hardier than I thought, then, maybe.

The recommended temps of a 6300 is 62 core, 72 socket

[shea99]

[shea99]

4 minutes ago, miagisan said:

The recommended temps of a 6300 is 62 core, 72 socket

Which temp reading software you use to see the socket and core temps ?

 

[miagisan]

8 minutes ago, shea99 said:

Which temp reading software you use to see the socket and core temps ?

 

I use hwinfo64 for package temps. I can't seem to get core temps to show up so I just go by the package

[shea99]

4 minutes ago, Strike105X said:

It depends a bit on the software used to monitor temps, what software are you using ?

 

Edit: this was addressed...

 

The FX8350/8370 are the best, followed by FX8320, then FX8300, then come FX43XX and lastly FX6300. If its 62C under full load then that's excellent. 

 

Yes and no, let me elaborate after one more quote :p.

 

Generally yes this is true, but its also more complicated then that, an FX can generally take up temps higher then the recommended specs but the problem is that in many cases, somewhere around 75C and at high frequencies (~4500Mhz+) you can get into a phenomenon called electromigration which makes the CPU under perform or which can even lead to cache errors thus making it unstable, its not a case for every FX cpu however its a lottery, i've seen FX CPU's run at ~80c for years  without issues, though those are in the minority do take note. Socket temps wise, well that really does vary from mobo to mobo and the quality of the socket, some going above 75c will start having issues even melting while others are fine even at 85c. It depends on the quality build of the mobo, though in many cases high socket temps actually reflect high VRM temps.

 

 

I recommend hwinfo64, from my tests it was the best software for reading AMD FX temps. Please post again with that.

Downloaded and just running for 10 mins will post pic then thanks 

[CDHoward]

Small Note: Someone has already told you that the 8350 is a better overclocker. The reason I can only get a solid 4.5 is due completely to my motherboard - which is still petty decent, but it's the next model down from yours. Yours has more VRMS etc.

 

EDIT: Some people make the awful mistake of getting the LE version, which is awful for overclocking and not really designed for it.

[shea99]

4 minutes ago, Strike105X said:

 

A little on the hot side ?? 

 

[CDHoward]

9 minutes ago, shea99 said:

A little on the hot side ?? 

 

Hotter than I'd allow. I wouldn't let mine go past 65.

[shea99]

18 hours ago, Strike105X said:

I also agree its to risky, personally i think you got 3 options:

 

a. Dial it down to safe values.

b. put a fan on the VRM's and one on the back side of the mobo where the cpu is located if the case permits it, like this:

 

 

 

Now in those images i used a slim 120mm fan for the back but you can use the fan that came with the CPU and use a different fan for the VRM cooling, of course i can't guarantee it but in most cases this might actually give you a 10-15c headroom (cooling it).

 

c. Getting a better cooler, which to be honest i don't think its worth it...

 

 

will try 4.5mhz later and post thermals 

[miagisan]

On 1/30/2018 at 11:08 AM, Strike105X said:

If its according to your profile and its a Gigabyte then its one of the TMPIN readings.

Thanks! Knowing that I see my cores don't go over 53.might have some headroom to go a bit higher on my OC. 

[Spudbilly]

5 hours ago, Strike105X said:

Sorry i misread it, i thought you where talking about socket temps >.< I guess i was to tired... one of the TMPIN's is actually the socket temp value, the Package temps should normally be the core temps, on gigabyte boards it might be a good idea to use hardware monitor, AOD also gives accurate reading of core temps.

I used to have an fx 8320 and I was getting 5ghz benchmark stable with the exact same mobo. It was not 24/7 stable. A lot of the advise everyone so far has suggested is spot on. However, for a maximum overclock there are a couple of other things to consider. 

 

 The trick to maximizing stability with the pile driver chips is to make sure that you have a good power supply. I could not get more than 4.9 stable at all. I borrowed a buddies gold rated higher wattage psu and took out my bronze 500w psu and bang for 5ghz. Also I set my ram to the lowest multiplier and I had a mixture of bus and multi overclock and kept my ram frequency as low as possible. I found that faster ram did not really improve performance but higher bus frequency did improve performance marginally with single core performance. Also that asus board also has a lot of power delivery tuning options. Overclock.net had a really nice oc guide for asus boards with comprehensive details about how to adjust all those options in the bios. The other is Northbridge oc and higher voltages. There is another voltage that I had to increase to get that 5ghz stable for benchmarks. 

 

If if you are thinking that heat is your barrier for 4.6 ghz alone then I would suggest that you really need to worry more about your core temp than your socket temp. Remember that you are getting 72+ on your socket stress testing not gaming or video rendering. However if you are consistently exceeding 62 on your core temp gaming then I would be worried about temps. 

 

In summary. Dial back memory and lower your cpu multi to the lowest. Then slowly and methodically increase your base clock and increase voltage appropriately until you find your maximum base clock. Then after you can increase the multi and voltages to achieve your 4.6 as long as temps are acceptable. Then if that is stress test stable then you can increase the memory speed and play with voltage and timings. But don’t worry about what speed you accomplish as the performance gains are super minute. Then after that you can play with northbridge overclocking and nb voltages. 

 

Doing it it this way you will achieve the best performance overall both single core and multi core. 

 

Gl and and don’t forget to take it slow don’t skip steps or under test. Record every step so that you can know where an instability originated from. To do this properly could take a couple days to do successfully.