FFTs or Blend?

[muntahunta]

I have been reading up around the internet and so far the 2 types of prime95 testing that overclockers seem to favour are Small FFTs and Blend.

I have ran both on my computer for short amounts of time to test temps and I am getting:
71 degrees (socket) and 63 degrees (package) with Small FFTs
59 degrees (socket) and 53 degrees (package) with blend mode.

Which should I be paying attention to? some people say one, others say the other...

I am using an fx6300 with a m5a97 r2.0 motherboard.

I have it running at 1.325v @ 4.6GHz using a Hyper212 Evo cooler.

 

[Automata]

Small FFT will stress the processor more than anything else, while Blend does CPU/Memory. I don't know AMD that well, but I believe you are supposed to keep it under 55c.

 

[muntahunta]

Thanks for the reply.
I've read that before, but which should I use for testing my OC, obviously max temps are completely different so I dont know which to go by.
Besides this, I dont plan on stressing my PC out to this extent very often so it will be very rare that it will go that high.

The max temps for AMD are 72 socket and 70 package.

 

[RGone]

Because you have a m5a97 r2.0 motherboard, I would not be looking for the most max heat at all. If you can get 2 hours of P95 Blend without error and the CPU/Socket temp not exceeding 72c and the Package/Core temp not to exceed 62c using HWMonitor Free with no errors during the P95 Blend run; you will have done what about 95% of us do for stability testing. About it. Otherwise it is mostly just a discussion about personal favorites that mostly goes on and on.

RGone...

 

[muntahunta]

Cheers RGone, i've knocked my overclock down to 4.5GHz (dont wanna get greedy and blow it up). i'd say a 1GHz overclock is good enough for me.
temps didnt get anywhere near 60 degrees now and i've been running this test for half hour. we'll see what happens in the next hour.

 

[RGone]

I don't really ever judge my own overclocks except by the speed with which my video renders are completed. I do a lot of video rendering and the time to finish is important to me. Such though is not the same for everyone. Gamers likely actually need the most cpu speed they can obtain within sanity and reason of course.

When I finally call my rigs complete, to satisfy myself, I need to see 4.8GHz stable hour after hour under video editting load. No sense in losing what I have modified to make up my own home video. So stability is what I need with my target speed of 4.8GHz on my FX-8350s.

I did some testing a few years ago now and will put them inside a spoiler below, so you can glance at what I found. Cpu Speed seemed like a sawtooth pattern as to how much 'extra' increase I got for each 100MHz in cpu speed. When I realized that there definitely seemed a up and down pattern, I got in the habit of testing my time to complete my video edits and I use the cpu speed with the least heat and fastest time to complete render. On the FX-8350 that I built my video edit rig around...that is 4.8GHz.

Spoiler

Testing does FX-83xx seem to flat line after 4.3GHz

With no two rigs dead on just alike...I may spend upwards of 2 weeks just juggling this and that to get to what is the best overall for the particular rig I am using at the time. One other thing I need to mention is that I seldom use my computers with their MAX speed. When I had my system built with the FX-8350 I use now but on another motherboard and with different cooling it was stable at 5.2GHz for over 2 hours of P95 Blend. Todays configuration is stable to about 4.953GHz and I drop back to 4.8GHz for stability insurance. No reason to struggle for that extra little bit of speed when I get the bestest and mostest for the energy expended at 4.8GHz and more likeliehood of remaining stable day in and day out.

Luck man.

RGone...