Overclock.net banner

1 - 7 of 7 Posts

·
Registered
9900k @ 5.0GHz 1.34v Maximus XI Code G.Skill RipJaws V 2x16GB 3200MHz @ 3400MHz 14-14-14-28 1.38v
Joined
·
5 Posts
Discussion Starter #1
Been lurkin for a long time. Thank you to the all the OC.net members who share so much knowledge, so readily!

I know my question has been asked/answered many times. But none of the standard answers are fixing my problem. My 9900k is locked at All-Core Multi: 50, Speedstep off, Speedshift off, C-States off, Manual Vcore locked, Windows 10 Power Plan set to Ultimate, AVX Offset -1... And yet recently I have noticed that HWINFO 64 records a Lowest Clock Speed of 4200Mhz. When I run Cinebench 15, 20, 23 or AIDA64, etc. the scores are right in line with my older results. And I have yet to notice any performance drop in games. But I don't want to ignore something this odd. I mean, where does 4.2GHz come from? I haven't found similar complaints on HWINFO 64's forum and this started before the latest update. I created a Sig, but can provide whatever anyone might neet for eval. Thanks!
 

·
AMD OC'ing Enthusiast
Joined
·
2,873 Posts
use hwinfo64's log feature to track WHEN and WHERE the clock drop occurs.

If you are using turbo in the BIOS to lock the 5ghz, its simple, turbo is not infinite, and occasionally resets the timer. In BIOS there would have been low and high current limits. something like 4096 and 1024 fopr values, and 56s or something for max power limit.

I do NOT own this platform, simply referring to my prior usage of an Asus rampage v extreme, and my current asrock killer x99m, both of which displayed this behavior with 2011-3 on my 5930k and 5820k
 

·
Optimal Pessimist
Joined
·
2,741 Posts
HWINFO64 shows all of the sources of throttling and whether or not the source is/has throttled. Look for a "yes" in the maximum column to see if it has throttled from that source (thermal, power limit, turbo limit, etc. - there are something on the order of 50 of them).
 

·
Registered
9900k @ 5.0GHz 1.34v Maximus XI Code G.Skill RipJaws V 2x16GB 3200MHz @ 3400MHz 14-14-14-28 1.38v
Joined
·
5 Posts
Discussion Starter #4
HWINFO64 shows all of the sources of throttling and whether or not the source is/has throttled. Look for a "yes" in the maximum column to see if it has throttled from that source (thermal, power limit, turbo limit, etc. - there are something on the order of 50 of them).
 

·
Registered
9900k @ 5.0GHz 1.34v Maximus XI Code G.Skill RipJaws V 2x16GB 3200MHz @ 3400MHz 14-14-14-28 1.38v
Joined
·
5 Posts
Discussion Starter #5
Heh... love this forum, thanks for the help!

I have all of the window timers etc. maxed in BIOS 127, 127, 4096 etc. but good call. I will check through HWINFO
 

·
Registered
9900k @ 5.0GHz 1.34v Maximus XI Code G.Skill RipJaws V 2x16GB 3200MHz @ 3400MHz 14-14-14-28 1.38v
Joined
·
5 Posts
Discussion Starter #6
I believe I figured it out. I tend to use TurboV from Asus to bump my OC a lil bit higher within Windows (BCLK: 100.5 or so). HWINFO log (thanks GeneO!) confirms that the recorded downclock to 4200MHz always occurs when I apply the BCLK adjustment. It doesn't always react this way, which is weird. But over 17 hours it only happened when I changed the clock. Maybe this will help explain things for somebody else. Thanks for the help!
 

·
Registered
9900k @ 5.0GHz 1.34v Maximus XI Code G.Skill RipJaws V 2x16GB 3200MHz @ 3400MHz 14-14-14-28 1.38v
Joined
·
5 Posts
Discussion Starter #7
Funny thing... it turns out that the downclock actually trips when I run Memory Read Test in AIDA64 Extreme. The reason I mistakenly associated it with the TurboV BCLK is simply because I tend to follow the same pattern when I boot Windows. First I bump BCLK, then I kill some ASUS programs: Freakin unkillable, totally useless "Armoury Crate" etc. Then I run 3 or 4 Benchmarks just to confirm that things are humming along as expected. Today I noticed the Minimum Clock reading in HWINFO 64 remain unchanged at my AVX Offset of 4920MHz after I closed TurboV. When I followed Up with AIDA64 benchmark tests, the Minimum Clock reading dropped to 4220MHz just before the Read Test finished. The Memory Write Test caused NO downclock. The Memory Copy test resulted in a downclock to 4320MHz. The Latency test didn't do anything weird, nor did the CPU tests. Ummm....

Now, all of my benchmark scores, including the sus Read and Copy test downclockers are the same as I have been scoring for months. My synthetic AIDA64 CPU Queen is between 106100 and 108500 and has been since I first overclocked my 9900k to 5GHz all core. Can anyone confirm the validity of this score for me? The lower clock readings only happen during those 2 AIDA64 tests, ( ran CB 15, 20, 23, 3Dmark, with no strangeness).

But why the hell would synthetic Memory Bandwidth tests cause my locked CPU multiplier to briefly unlock and drop? And really why would the two tests each result in different drops? I felt better when I thought it was an ASUS program causing the clock change, I'm used to ASUS software trying to murder my rig! Anyway, I'd really like to understand how this is happening. I'll be reading everything I can find on the subject, but if someone has an answer, I'd be grateful to hear it!
 
1 - 7 of 7 Posts
Top