• 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.

I9 9900k

Have run Cinebench several times now without hitch, and then a wPrime run, so far so good, but Realbench is next, different kettle of fish altogether, lol.

e4e49501-386b-4ada-860d-78ecf98a1dc9-original.jpg
 
Distracted said:
Thank you.

If i set 1.33 in bios and then hwinfo reports 1.27 - 1.323 vcore should i still be going by the setting in bios?

I saw that elmor said the boards were more accurate with the XI series but I don't know how that translates to bios settings vs hwinfo sensor readings.
Click to expand...
Perfectly normal droop, the only truly accurate measurement would be from a multimeter. It's effected by several factors and settings. As the cpu load increases the supply voltage drops, the cpu requests more power from the vrm. It communicates its request using the core VID over the SVID system. Set LLC to reduce droop but you get an initial spike of high voltage. A good option is adaptive voltage, which was designed to help. So the VID can move and up and down with load and the board attempts to hold a set point for the turbo speed. Means the Vcore isn't always set at the high point using power.
 
Last edited:
Well the voltage gap between 5.1ghz and 5.2ghz seems quite big, my bios is telling me I need 1.356v for a stable 5.2ghz,(Asus boards only) but I am hoping to be able to use a bit less. ;)

Realbench giving me a hard time, up to 1.3v and yet to get a full run even with the benchmark, but going to carry on regardless.

Asus turbo vcore actually gives a correct reading for vcore, it reads the same as what is set in the bios.
 
moorhen2 said:
Well the voltage gap between 5.1ghz and 5.2ghz seems quite big, my bios is telling me I need 1.356v for a stable 5.2ghz,(Asus boards only) but I am hoping to be able to use a bit less. ;)

Realbench giving me a hard time, up to 1.3v and yet to get a full run even with the benchmark, but going to carry on regardless.

Asus turbo vcore actually gives a correct reading for vcore, it reads the same as what is set in the bios.
Click to expand...
Are you using an avx offset?

Sorry if i have already asked.
 
Distracted said:
I found that as well. Small voltage increases up to 5.1 then 5.2 wanted a much bigger jump and 5.3 wanted even MOARPOWER.
Click to expand...

Yes it comes to a point that voltages needed get out of hand and the obvious temp rises that go hand in hand, without any gains performance wise.

I am at 1.310v in bios now, at 1.296 under load, but this does get the Realbench Benchmark to complete, needed to go to LLC8 due to vdroop though. But temps are getting into the 80's along with it, and I am fully water cooled with 2x 360's, a 240 and 120 rad's, and my trusty XSPC Raystorm Pro.
 
moorhen2 said:
Yes it comes to a point that voltages needed get out of hand and the obvious temp rises that go hand in hand, without any gains performance wise.

I am at 1.310v in bios now, at 1.296 under load, but this does get the Realbench Benchmark to complete, needed to go to LLC8 due to vdroop though. But temps are getting into the 80's along with it, and I am fully water cooled with 2x 360's, a 240 and 120 rad's, and my trusty XSPC Raystorm Pro.
Click to expand...

That's a lot of rad space!

I'm using 1 360 right now, but I'll be adding surface area asap.
 
moorhen2 said:
What are your temps like under full load may I ask ?
Click to expand...
With the single ek pe 360 and the gpu, vrms and cpu in the loop I was hitting low 80s in the realbench stress test.

This was a 15 minutes run set to use 4gb ram (32 installed) at 5.2ghz with -1 avx offset.

It could have been lower but I've still got my fans on some kind of adaptive profile that only responds to the cpu temp (asus mobo wont read my gigabyte card temps). This means they don't seem to ramp up until the cpu hits 60ish and don't respond to the gpu doing work at all.

I'm going to have to fix that before I can try anything else, otherwise the gpu is just heating the loop up too much. A job for tomorrow.
 
fornowagain said:
Perfectly normal droop, the only truly accurate measurement would be from a multimeter. It's effected by several factors and settings. As the cpu load increases the supply voltage drops, the cpu requests more power from the vrm. It communicates its request using the core VID over the SVID system. Set LLC to reduce droop but you get an initial spike of high voltage. A good option is adaptive voltage, which was designed to help. So the VID can move and up and down with load and the board attempts to hold a set point for the turbo speed. Means the Vcore isn't always set at the high point using power.
Click to expand...
Ty. Looks like I've got some more reading to do.
 
lowest i can run cinebench r15 on full win 10x64, i say this because stripping services back to there bare minimum i can run with less voltage. coretemp and hwmonitor were zero'd before running. LLC7



****EDIT just testing the above for 24/7 stabillity, needing 1.4v to pass 15min stress test in realbench2 with 8gb set and hitting 82c.
Also running as we speak 10 mins in, prime 95 ver 26.6 torture test, max heat etc, max temp 76c
 
Last edited:
Had a little play last night.
2x360 + 240 900rpm, 2080ti shares.
5.1ghz @ 1.309v
Hottest core
1hr OCCT large 65
Realbench 30 mins 74
~1.5hr Prime AVX 88

One thing annoying my ocd is the difference between hottest and coolest core. 88 prime but coolest 80. Reseated but nothing changed. But all in all, I am very happy with my OC.
 
Last edited:
You must log in or register to reply here.
Back
Top Bottom