FX-4300 : Golden or merely Good ?

[Kenrou]

So i got tired of torturing my 8370 and went back to my old 4300 now in my wife's PC (i made a thread asking for opinions about 8 months ago if memory serves), and i would like to know how reliable is IBT. I only used the standard test + Prime95 small FFT 30m + Crysis 3 after with no problems because my MB is giving off signs of old age and is no longer 100% stable, so don't know what the effects of high or very high would be on it. Replace my sig with this CPU and the tweaking stands as :

5ghz (25 x 200) - CPU 1.404v - CPU/NB 1.250v - CPU/VDDA 2.675V - NB 2600mhz / HT Link 2600mhz









I prostrate humbly before you and ask for the boon of knowledge oh wise ones

 

[caddi daddi]

I only use IBT to test cooling systems, just max out the ram and let it rip.
1.404 at that clock is truly golden.

 

[Silver_Pharaoh]

5Ghz at 1.404Vcore?
That's what ​I would consider golden

Run Prime95 blend for 2 hours, if you pass and there's no WHEA errors in the Event Viewer then you're set!

 

[caddi daddi]

if it does I want THAT chip!!!!!

 

[Mr.Scott]

It IS above average, but IMO not golden.
There are many 5 gig @ 1.40-1.44v FX-4300's at HWB.
Will it complete Cinebench R15 or GPUPI for CPU 1B at that setting?

 

[Silver_Pharaoh]

It IS above average, but IMO not golden.
There are many 5 gig @ 1.40-1.44v FX-4300's at HWB.
Will it complete Cinebench R15 or GPUPI for CPU 1B at that setting?

I still want it even if it can't run Cinebench at those speeds

if it does I want THAT chip!!!!!

I'LL PAY DOUBLE!!!

 

[Kenrou]

Would that I could but like I said my Crosshair isn't stable anymore and the past couple days I'm pretty sure I'm getting the FF CPU error. Decided to stick to AMD for now (until Zen becomes the norm if ever) so getting a Saberkitty maybe next week with a bit of luck, then I'll resume testing to see just how "badass" a chip I really got

 

[Johan45]

I see you're still not using the AVX version. The standard version on standard setting is way too light for the FX

 

[Kenrou]

The AVX version u linked reported higher GFlops but had lower temps then the regular/Prime95 and required much higher voltage to report a stable OC then the one i got from the official website, is that normal ?

 

[trents]

Your stress testing is perhaps rigorous enough to ensue stability in non intensive tasks like word processing or surfing the net. Prime95 blend for two hours would be better. Mostly, I use IBT for a quick check on temps and cooling capacity and I find that the "Very High" setting is fairly equivalent to the Prime95 blend test as far as stability testing.

 

[Johan45]

Yes that's normal kenrou, the AVX instuctions can process alot of info much faster but that requires more power/heat to accomplish. We have noticed in benching what a difference in scores AVX can make. So if you happen to run a software that takes advantage of those instructions and you haven't tested them could easily cause a crash/lockup
Go to this page http://www.overclockers.com/forums/...Marathon-Season-IV-HWBOT-x265-Benchmark/page9
You'll see jiccman posted two shots in that thread one with AVX and one without. His score went up 25% that's huge

 

[Kenrou]

Aight will do as soon as the Saber arrives, and i will test the 4300 before it goes back into wife's comp (unless u guys were serious in buying it), methinks it will give the OC a nice boost. Danke for the info

 

[caddi daddi]

don't think I'm not.........
if that will run p95 blend for 2 hours at that clock and voltage you can have a nib 6350.

 

[Kenrou]

So, still with the Crosshair, seems to have cleared up whatever was ailing. More testing with the 4300, 5ghz - 1.392v - 4.29 Cinebench but still failing Prime95 blend or IBT higher then standard. Its not voltage that even at 1.5v it failed. Ran memtest and it came out ok, so either the mobo is still not 100% or ? maybe chipset driver ? using XMP profile for my corsair 1866 cl9. Again no crashes on anything except stresstesting, SuperPI and gaming are working as it should afaik.

 

[Mr.Scott]

SPI only uses 1 core. Of course it would pass.
I see 1.225 vcore in your screenie. More voltage man. Temps are still ok.

 

[Kenrou]

SPI only uses 1 core. Of course it would pass.
I see 1.225 vcore in your screenie. More voltage man. Temps are still ok.

What ? you lost me, what only uses 1 core and vcore reads 1.392v, go further down on the mobo section

 

[Johan45]

It's likely the 2600 NB, what kind of volts are you giving the CPU_NB. 1866 + 2600 will likely need 1.3 ish maybe more

 

[Kenrou]

1.25v, will raise to 1.3v and try again.

Edit : near instant freeze even at 1.4v. What makes this strange is that like i said it my system never crashes except for stress testing. Played the most demanding games like Crysis 3, BF4, GTA V for hours straight and nothing bad to note. Run SETI@Home when not doing anything else, do video coding... Wouldn't that be considered as system stable ? i got used to only using Prime95 small FFT and IBT standard to pass the voltage requirements for each step since i always had this problem (2nd hand motherboard, maybe 1yo when i got it).

Edit 2 : cant find that IBT AVX, anyone got a link ?

 

[Mr.Scott]

My bad.
You still need more than 1.392v for 5 gig. You'll be too hot before you get it stable there....if at all.

 

[Kenrou]

My bad.
You still need more than 1.392v for 5 gig. You'll be too hot before you get it stable there....if at all.

So far my "stable points" have been :

4.7ghz - 1.308v
4.8ghz - 1.344v
4.9ghz - 1.368v
5.0ghz - 1.392v

Saber is taking too long to get here