©ASUSTEK COMPUTER INC. ALL RIGHTS RESERVED.

bass_junkie_xl · ‎12-01-2017

Hello fellow over clockers I have a
8700k @ 5 ghz delided @1.33v adaptive mode
Board Asus x hero Wi-Fi
Ram 16gb g.skill rgb 3000mhz
Timmings 14 14 14 34 1.35v

I ran prime 95 26.6 blend with 15.5 gb ram coverage
For 12 hrs and Asus real bench for 12 hrs all pass with
Typhoons reader says there samsung b die so I delided let's see what they can do .

I left vccio and vcssa on auto
And put 4000mhz speed timmings
18-19-19-39 1.4v and ran hci men test over night for 12 hrs with 12 Windows of it using all my ram for 2000 % coverage
Woke up no errors.
What's my next step ?
I have used prime and Asus real bench and games to test my
5 ghz with 3000 xmp but haven't used
Prime 95 or Asus real bench with the 4000mhz oc yet .

Do I need to re test Asus and prime again even though hci men test ran for 12 hrs ?

Silent_Scone · ‎12-01-2017

Use the system for what you built it for. That's more than enough coverage.

13900KS / 8000 CAS36 / ROG APEX Z790 / ROG TUF RTX 4090

JustinThyme · ‎12-01-2017

Prime95=how to make a CPU a space heater. It does nothing real world. If you want to run it for a cooling solution test that's about the extent of its usefulness.
When checking OC I just run real bench (not called real because it sounds cool) in a loop for 15 mins or AIDA 64 and call it a day and enjoy the OC.
If you get stability issues later then address it. Thing with over clocking you can beat the crap out of it for a week with P95 and have no issues, then launch one app that twists it a different direction and it locks up. The highest clock is not always the best either. Yes I like running it up just to see how high it will go but then I always dial it back so I'm not listening to fans screaming. Ive had my 7900X to 5.1 made it though several benches, got hot as hell. 5.0, still hot. 4.9 a bit better. 4.8GHz is where Im sitting and its much cooler with less Vcore and I can jack my cache to 3.2 from 2.4 and this actually beats the 5.1 real bench score. Keep that in mind as you tweak that the highest clock may be < best clock for performance.

“Two things are infinite: the universe and human stupidity, I'm not sure about the former” ~ Albert Einstein

bass_junkie_xl · ‎12-01-2017

Yeah I ran hci men test for 2000% coverage and today I been playing battlefield 1 64 player 1080p no fps lock on 1080ti for 5 hrs straight and going no exe crashes or stutters so
Going from 3000 cl 14 14 14 34
To
4000 cl 18 19 19 39
Is pretty sweet most 4000 kits need 19 19 19
Or 19 21 21 41
I don't think trying 19 19 19 39 would be Werth it ? To re run all these tests again or is 18 19 19 39 pretty good for 4000mhz

meankeys · ‎12-12-2017

JustinThyme wrote:
Prime95=how to make a CPU a space heater. It does nothing real world. If you want to run it for a cooling solution test that's about the extent of its usefulness.
When checking OC I just run real bench (not called real because it sounds cool) in a loop for 15 mins or AIDA 64 and call it a day and enjoy the OC.
If you get stability issues later then address it. Thing with over clocking you can beat the crap out of it for a week with P95 and have no issues, then launch one app that twists it a different direction and it locks up. The highest clock is not always the best either. Yes I like running it up just to see how high it will go but then I always dial it back so I'm not listening to fans screaming. Ive had my 7900X to 5.1 made it though several benches, got hot as hell. 5.0, still hot. 4.9 a bit better. 4.8GHz is where Im sitting and its much cooler with less Vcore and I can jack my cache to 3.2 from 2.4 and this actually beats the 5.1 real bench score. Keep that in mind as you tweak that the highest clock may be < best clock for performance.

Good advice Justin - I like the way you think 😉

Silent_Scone · ‎12-01-2017

If only adjusting the memory timings and no other settings then you can use google stress app test for 1 to 2 hours.

I'm assuming the timings written are a typo, however, as the above post suggests you want to relax timings. This increases latency and reduces performance. As you are stable there is no need to do this.

13900KS / 8000 CAS36 / ROG APEX Z790 / ROG TUF RTX 4090

bass_junkie_xl · ‎12-02-2017

Silent Scone wrote:
If only adjusting the memory timings and no other settings then you can use google stress app test for 1 to 2 hours.

I'm assuming the timings written are a typo, however, as the above post suggests you want to relax timings. This increases latency and reduces performance. As you are stable there is no need to do this.

to be clear i changed my ram from xmp 3000 mhz speed with timmings 14-14-14-34 2t
to
4000 mhz speed timmings 18-19-19-39 2t and ran hci mem test for 2000 % coverage for 6 hrs and ran prime 95 26.6 blend for 6 hrs and did a short 2 hr asus real bench no errors .

ran battlefield 1 64 player for 5 hrs with 1080p no fps lock to stress system and no crashes .

my question was what other tests should i run with ram @ 4000 mhz with good 18-19-19-39 timmings or am i stable now ?

my 8700k @ 5 ghz 1.33v adaptive with power saving , cache / ring = 47 , ram @ 4000 mhz

one thing i did notice is games and over all system does feel faster from 3000 vs 4000 my fps in some games went up and minamum fps went up in a few games even though i have a 1080ti . thanks for replys

8700k @ 5 ghz delided 1.33v stable with ram oc of 4000mhz help