-
Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!
AMD FX 8370E Noob build and overclock to 4.8ghz
- Thread starter Cwwc-Stalker
- Start date
- Joined
- Aug 14, 2014
As far as I've been told the only reason to push CPU/NB is if you have "high speed" ram. Higher values lets you use the ram to its full potential or something like it. Also helps with overclock stability, like my 4.8ghz isn't fully stable unless I raise it to 1.3v.
Someone more knowledgeable feel free to correct me though, never really found any concrete info on this.
Someone more knowledgeable feel free to correct me though, never really found any concrete info on this.
- Joined
- Sep 28, 2015
At least for Phenom II's the sweet spot depends on the CPU setting. Faster is generally better, but the fastest stable setting may not be the best for a given CPU clock speed (see Dolk's guide for details). The gains are in any case quite marginal and show up best in purely synthetic tests.
- Joined
- Sep 28, 2015
Based on the little benchmarking I've managed to do, OC'ing CPU-NB/HT can be an alternative to OC'ing RAM. I saw similar gains AMDOD's memory test for raising the CPU-NB/HT from 2200 MHz to 2400 MHz as I did for changing the timings from 9-9-9-25-2T to 8-8-9-25-1T - and the tighter timings weren't stable. MaxxMEM2 showed a lower latency when changin the timings, but slightly higher mem score (in GB/s) when OC'ing CPU-NB/HT. However, the MaxxMEM2 scores weren't very consistent.
- Joined
- Aug 14, 2014
Yep but unlike the Phenom II the FX don't get any real gains from anything above 2600mhz. From what I've read some boards won't even post above that.
The trick here is having the HT Link and CPU/NB at the same frequency and like any overclock try to pull it off with the lowest voltage/heat possible.
The trick here is having the HT Link and CPU/NB at the same frequency and like any overclock try to pull it off with the lowest voltage/heat possible.
- Joined
- Sep 28, 2015
Slightly OT, I just tried Aida64's stress test and temps stayed really low compared to Prime95. Package temps stayed below 45 C and socket temps maxed out at 55 C. Five minutes of Prime95 max heat test took me to 51 C and 68 C, respectively. Clearly Aida64 is pretty light when it comes to stress testing. If you want alternatives to Prime95, have you tried RealBench v2.41? Some people run the benchmark several times (10+) in a row to test stability, others just use the program's stress test, which is heavier than just running the benchmark. The "upside" is that it stresses the GPU as well, which may lead to hotter air inside the case and thus make a barely stable system unstable.
PS. My system produces a consistent result mismatch in the RealBench stress test even at stock settings, so I wouldn't pay much attention to that. If your system doesn't crash, your settings are likely okay.
PS. My system produces a consistent result mismatch in the RealBench stress test even at stock settings, so I wouldn't pay much attention to that. If your system doesn't crash, your settings are likely okay.
- Joined
- Jan 1, 2008
- Location
- Colorado Springs, CO
I have found that if I cannot prime95 stable (usually for well over 2 hrs)
, I most likely will fail work units in F@H. I've been on a prolongs folding hiatus, but that showed me that I needed to play less loose with the overclock. That being said, I usually do a 2 hr stability check (on blend) for end clocks (IE, a stopping point) and 10-20 mins on ceiling pushing (just trying to get to a goal). Then, when I goto bed, if I have done a 2 hr test, I just let her rip all night. The fire alarm will let me know if bad stuff happens...
, I most likely will fail work units in F@H. I've been on a prolongs folding hiatus, but that showed me that I needed to play less loose with the overclock. That being said, I usually do a 2 hr stability check (on blend) for end clocks (IE, a stopping point) and 10-20 mins on ceiling pushing (just trying to get to a goal). Then, when I goto bed, if I have done a 2 hr test, I just let her rip all night. The fire alarm will let me know if bad stuff happens...
OP
- Joined
- Nov 5, 2015
- Thread Starter
- #70
I totally agree aid64 isn't as much of a ball buster as p95 but you would never even come close to the stress p95 gives anyway. Plus The only warnings I get are illegal sumouts which are memory related. Nb kinda stuff
Edit: I pass R15 I pass aida 64 I pass overdrive bench and if p95 doesn't I can live with that. I get no BSOD no freezes or hiccups. Since my rig is just for fun and gaming now and then I haven't had one issue yet in 3 weeks of everyday use.
- Joined
- Sep 28, 2015
I agree, P95 is a bit of overkill for your use. I use it just to get some peace of mind (if it's P95 stable for 8 hours, temps will not be a problem with anything I might throw at it), but if I were looking for the fastest settings possible for gaming, I'd use RealBench.
Btw, if you're planning on changing your RAM timings, calculating pi (SuperPi 1M or y-cruncher) is one way of testing stability. I was trying different timings some time ago to find the limits of my current modules, and while I could pass MaxxMEM2 and AMD Overdrive memory benchmarks with some settings, calculating pi with those same settings produced occasional errors.
Btw, if you're planning on changing your RAM timings, calculating pi (SuperPi 1M or y-cruncher) is one way of testing stability. I was trying different timings some time ago to find the limits of my current modules, and while I could pass MaxxMEM2 and AMD Overdrive memory benchmarks with some settings, calculating pi with those same settings produced occasional errors.
- Joined
- Sep 15, 2015
- Location
- Pensacola, Fl
I think the op was about why this core or that was failing during p95 blends, seems like you need more voltage and way better cooling. Not sure why you're getting those temps at that vCore though.
Glad to hear it! despite what intel fanboys say these AMD chips well at lest mine anyways are sold chips for the price! I know for a fact mine oced beats the crap out of i5's in everything except gaming by what 10fps? does 75 vs 85 matter that much lol? hell I can beat out more I7's in lots of benches. Is it as poweful compared to the top end intel chips? NO but it's not far off and at 300-400 cheaper it's a no brainer.
Now gaming wise you have to pair it with an r285 or 380 at stock speeds but once you overclock the 8370e you can easily use an r390x without a bottleneck.
Sure the ipc is not as good as intel but to be honest this thing in 99% of tasks will crush i5's. Those sites that are intel funded will put amd into the ground everytime.
I do not hate intel nore do I hate amd, They both make good chips and I just happen to like the color red more then blue
Agreed, also take into account the fact that a lot of software use Intel compilers which are designed by default to give the lowest optimisation to anything but 'GenuineIntelProcessor' flag... I have a link to a place that goes in depth to discuss this and it was written up some 5 yrs ago... Source
If they were doing it back in 2010, how do we know they are not doing it today? Imo, unless your a professionally qualified programmer that knows their stuff and has access to the code, you wound never know.
This part is particularly worrying for Intel's competitors from that article link I posted..
Last edited:
- Joined
- Jan 10, 2012
have you guys noticed that in his first picture that the rad and pretty exhaust fan cover his vrm and prevent air from flowing over that heatsink?
this was hashed out here quite a while ago (2 years?) In a large thread. I'm mobile or I'd link you. I recall hearing that there were still differences, but not as much? The problemishes isn't there, it's with the cpu. Hopefully zen tales care of it.
EDIT: @ caddi/all - something tells me the op won't be back.
- Joined
- Jan 10, 2012
yea, 11/06 was his last.
