• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

Just Putzen Round

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.
Did anyone else pick up some goodies on Black Friday?
I'm most likely heading over to Microcenter to pick up a chip for a fellow forum member and send it to him, you never know what I'll walk out of there with.
 
Pushing the clocks on this chip a little higher now... Still exploring with vcore set in BIOS to 1.3875v. The combo has been "crunching" 8-threads of Rosetta for ~6 hrs. now and it will continue overnight. I haven't stress tested with Prime yet but still for me long runs of Rosetta or WCG are a real everyday type of load usage.

FX-8370 4.723GHz Rosetta load.PNG
 
Pushing the clocks on this chip a little higher now... Still exploring with vcore set in BIOS to 1.3875v. The combo has been "crunching" 8-threads of Rosetta for ~6 hrs. now and it will continue overnight. I haven't stress tested with Prime yet but still for me long runs of Rosetta or WCG are a real everyday type of load usage.

Seems like a nice low voltage chip PolRoger, I need 1.46875 Cpu V at 4.7 to pass 2 hours Prime Blend with mine. Mine was purchased when they were first available for retail. If I can pass 2 hours prime, it's good enough for me to F@H with and not crash. I ran the chimp challenge Feb of 2013 for then entire time at my daily 4.7 Oc without a problem. Assuming Rosetta is basically the same stress as F@H, 6 hours is stable in my book.

One thing that seems to hold true with these Fx chips is the old adage low low leakage chips are good on ambient cooling and high leakage chips are good for sub ambient. I've seen a few of these low leakage/voltage chips come through here, especially of late and they're really good to a point on ambient cooling. Though the term good can be taken with a grain of salt, I've seen some of these low leakage ones still get into the mid 5's when really pushed, which is not too shabby. When taken cold though they seem to top out a lot sooner then higher leakage ones. It seems as if they are binning the high leakage chips for the Fx 9xxx chips and leaving the low leakage chips for the Fx 8xxx series.
 
Well it "crunched" all night as well (~15 hrs. total)... I'm now testing ~4.8.

What do you find is the "sweet-spot" range for NB/HT/MEM speeds on the FX series? I've also got a 8370E coming later in the week from Newegg to compare with this one. It does seem better on ambient cooling than my initial launch 8350 which I later sold off.


FX-8370 4.831GHz Rosetta load.PNG
 
Last edited:
Well you have CPU_NB and HT Link Speed both getting close to 2500Mhz and that is not shabby. Also the ram is at roughly or close to DDR3-2100.

Here lies the rub. You are not testing with a known stressor. I have seen P95 fail with faster ram than you have showing with low CPU_NB speed. BUT you are neither too high nor too low on ram speed and CPU_NB and HT Link Speed. So if your stressor is what you are always going to use and it is fine for you then so are the buss speeds surrrounding your overclock. Probably about all you need overall and anywhere for cpu speed and ancillary speeds.
RGone...ster.
 
I agree Prime 95 is a very tough stressor for both the FX and Haswell. The current settings probably wouldn't pass. I'm running now under an AIO (H110). I'll probably wait until I can get the setup under custom water with a big rad before bringing on the heat of a Prime test. I backed off some on NB/HT and went to C9 when I bumped up the bus from 219 to 224.
 
I run my Fx 8350 @ 205*23 which gives me HT and Nb Freq at 2665 for daily clocks and my ram at 1914 Mhz 24/7. For benching I've been able to run my Nb Freq between 2800-2900 and it helps "depending on the bench". Though, I need to add a heaping of voltage to the Cpu Nb v to get above 2800.
I've also got a 8370E coming later in the week
^^^^Because you you don't have enough chips? :p he he

It does seem better on ambient cooling than my initial launch 8350 which I later sold off.
I remember when you had that chip.
 
Thanks everyone for the kind good wishes and comments will look in here from time to time, and try to stay in touch!

Merry Xmas to all.

Ajay.
 
^^^^Because you you don't have enough chips? :p he he

I always kind of thought that when it came to overclocking and computer parts that "Moar" is Better. :)


I ran Rosetta for another ~6 1/2 hrs. with my previous settings... Bumped vcore up a little and I'm going to try and see if I can dial in ~5.0.
 
I always kind of thought that when it came to overclocking and computer parts that "Moar" is Better.
Funny there are few of us here that think the same way. :D

I ran Rosetta for another ~6 1/2 hrs. with my previous settings... Bumped vcore up a little and I'm going to try and see if I can dial in ~5.0.
Sounds like a plan!
 
This new round of FX is definitely different. Takes much less voltage than they did a couple years ago. That 8320 I have, which is suppose to be low end, Runs at 5.0 with 1.45v but the darn thing won't scale under cold at all. You're on the right track Rog. Looks good to me and I wouldn't be surprised if that CPU will run 5.0 with the cooling you have right now.
 
So it crunched all night at 5.0 with around 1.42 Cpu V.... That's pretty darn good in my book.

Just noticed the VID in core temp 1.2875 :shock:
 
Just noticed the VID in core temp 1.2875 :shock:

Yeah, I remember back a year or so ago when C_D bought those two FX-8350s and the one with the lowest x20 P-State was 1.296 and it would n0t do a nickle over 4.8Ghz but do that on AIR with not quite 1.4Vcore.

All FAB processes seem to mature and maybe part of the reason we are seeing funny stuff the last 9 mos or so out of the 3 years that FX has been around. I think the FX-9xxx series changed the binning process or as the first year went bye, AMD decided to bring out other models based on HOW the binning was turning out. I seriously doubt GF is still making any FX processors but more likely only the APUs and what we are seeing now in retail has been FAB'd sometime ago. But are just now reaching REtail market under new part numbers and having different default TDP's and Speeds.

It will be very interesting to see if there are a great number of these seemingly low-leakage processors that will actually run the big numbers we saw out of the FX-8 core cpus some 20mos ago. So far the average run of the mill user has not put enough cold water on them to know. Plus most of us with 2 year old PileDriver cores have no need of another cpu. We already have one that will run 4.7ish all day and then buzz to +5.5Ghz for benching. This gives me doubts we will ever fully know what the latest 6 month old or newer batches of released cpus will ever do.

There is something to be said though for a cpu with a low 20x P-State that will do 4.8Ghz all day on air. Even if it is not the Big Boy +5.5Ghz benching clocker some of us wish we had to play with. Hehehe.
RGone...ster.
 
So far this sample seems like very nice ambient clocker for daily type usage. I haven't really pushed hard on it yet but I have been able to boot to windows and light bench with it at ~5.466GHz and I think there is still more there.

http://www.overclockers.com/forums/...-HWBOT-Prime?p=7762240&viewfull=1#post7762240

I need to get it under one of my custom water setups but I need to switch out a cpu block as I only have one AMD mount (XSPC Raystorm block) and my others are Intel.

On a side note... After more than 12+hrs crunching Rosetta 100% load with vcore set in BIOS to 1.4125v it BSOD with a 101 code.
 
Hopefully it'll keep going for you Rog. My 8320 would bench Cinebench11 at 5.4 with 1.45v on my loop but the thing pretty much quit there. Over 5.5 it wasn't stable enough even with LN2 to run R15 and topped out at 6.6
You likely need a bit more V_Core. What's your CPU_NB volts set to?
 
I've now dropped it back from 229 to 228 and I was running/testing with CPU/NB set to 1.1875v in BIOS.
 
For the ram and CPU_NB speed it'll likely need to be in the 1.25v area maybe a bit more. I have found IBT AVX very useful in setting my CPU_NB voltage. Run the bench on high or very high and you'll notice as you raise the CPU_NB volts the Gflops will increase to a point and stop. That's typically where it's running the most efficient.
View attachment IBT AVX.zip
 
Back