• Competitor rules

    Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.

Fx-8350 Golden chip?

Soldato
Joined
10 Oct 2012
Posts
4,842
Location
Denmark
Im wondering if i have been very lucky with my current fx-8350. Currently testing it using a corsair h100i(quite mode) and right now its sitting at 4.8ghz using 1.42v @46° core, 53° cpu package under AIDA64 stability test.

What are you guys/gals results/thoughts. I might actually have to take my statement back that i made in an older post to teppic regarding the 8350 being a crap OCer for people on air(i know im not on air but if the H100i can cool it this good, im sure there is a few super air coolers out there able to achive close to the same result). Either that or i got extremely lucky with the silicon lotteri.

29fatxi.png

Im well aware that the stability test have only run for just under 10 minutes in this pic but as i already have tested it for 2 hours i didnt see a reason to do it again.
 
That's pretty good, 1.5v is often needed to get around 4.8GHz. I expect you could get it to 5GHz stable with a bit of tweaking as long as you weren't running artificial stress tests - those would probably push the temps too high. Under water you can take it above 1.5v but I'd be reluctant to myself (I think 9590 overclocks have used 1.6v). I'd say that whatever you can get at 1.5v is a good aim, as for general loads the temps should hopefully be all right. The closer you get to 5GHz the more you overcome single threaded bottlenecks in performance, allowing the other cores to be much more efficient.

(It's often the board's phase design that gives up on higher clocked 8320/50s -- far less likely on a Sabertooth or UD5/UD7/Crosshair Z board)
 
Last edited:
Im wondering if i have been very lucky with my current fx-8350. Currently testing it using a corsair h100i(quite mode) and right now its sitting at 4.8ghz using 1.42v @46° core, 53° cpu package under AIDA64 stability test.

What are you guys/gals results/thoughts. I might actually have to take my statement back that i made in an older post to teppic regarding the 8350 being a crap OCer for people on air(i know im not on air but if the H100i can cool it this good, im sure there is a few super air coolers out there able to achive close to the same result). Either that or i got extremely lucky with the silicon lotteri.

29fatxi.png

Im well aware that the stability test have only run for just under 10 minutes in this pic but as i already have tested it for 2 hours i didnt see a reason to do it again.

Just got an FX-8350 for a server test bed Build, sadly i'm not going to have time to play with it as it needs to be shipped out, but a quick 20 Minute Prime95 test at stock on a £35 Air cooler brought the cores up to 40c in HWMonitor and no more....

Stock volts were about 1.32v, I don't know but that seems pretty bloody good to me, certainly much better than what has been reported in the past.

the sticker is the latest AMD logo (only out a few months) Better yields? some tweaking they didn't tell us about?

The chip seems very nice, it might even win me over.
 
Last edited:
Overclocking potential has improved a huge amount in Richland over Trinity APUs (both being Piledriver), so it's quite possible they've improved things or tweaked it a little in the last few months. The 9590 most likely took advantage of some improvements.
 
I think there is a very big difference between multiplier OC and BCLK or whatever its called for AMD cpus. BCLK seems to have a much bigger impact on performance. Using the 1.42v i would be limited to around 4.4ish instead of the 4.8ghz i am now using multiplier only. Im very interested in what you guys can get your chips to run at.

I really think the fx-8320-8350 is a great and underappriciated chip. Considering its still stuck on the 32nm i personally think the performance we get from one is pretty darn good.
 
It's definitely underrated. The problem is that Bulldozer was a massive let down, and Piledriver is still tarnished with that. I think you do need to get Piledriver cores to 4.5GHz to get the best out of them - a lot of the time in multi-threaded apps/games one or two cores will be hammered and it can hold back other 'helper' threads running on the other cores.

Increasing the bus speed will have slightly more performance benefits, but you're going to hit a wall far sooner than the CPU is actually able to go.
 
It's definitely underrated. The problem is that Bulldozer was a massive let down, and Piledriver is still tarnished with that. I think you do need to get Piledriver cores to 4.5GHz to get the best out of them - a lot of the time in multi-threaded apps/games one or two cores will be hammered and it can hold back other 'helper' threads running on the other cores.

Increasing the bus speed will have slightly more performance benefits, but you're going to hit a wall far sooner than the CPU is actually able to go.

I don't think it will, far from it.
Some of the better boards can do like 350 bus, you could even underclock the multiplier and go all bus speed.
 
I don't think it will, far from it.
Some of the better boards can do like 350 bus, you could even underclock the multiplier and go all bus speed.

I've never tried bus speeds that high, it'd stress the entire system for sure though. From what I've seen Phixsator's experience is common on Piledriver, higher clock speeds via the multiplier than via the bus.
 
You've got to be careful with the bus speed, can cause all sorts of issues including corrupt HDD/SSD. It does offer better performance but usually you can clock higher with the multi so it balances out anyway.
 
I've never tried bus speeds that high, it'd stress the entire system for sure though. From what I've seen Phixsator's experience is common on Piledriver, higher clock speeds via the multiplier than via the bus.

I've tried bus speeds that high, you can get higher clocks with the bus and the multi for sure, but using a lower multi with high buses allowed me the furthest overclocks, at least with Deneb/Thuban.
In fact I think MSI brought out a freak of a mid range board that did 400.

You could for example lower the stock multiplier and then offset that with a really high bus.

You've got to be careful with the bus speed, can cause all sorts of issues including corrupt HDD/SSD. It does offer better performance but usually you can clock higher with the multi so it balances out anyway.

That's true for Intel and BLCK, but AMD's no where near as linked yet for the HT Ref.
 
Back
Top Bottom