Overclock.net banner

2661 - 2680 of 2704 Posts

·
Registered
Joined
·
8 Posts
Thanks again! I just noticed that my XMP profile sets the VCCIO and VCCSA to 1.315v, isn't that too high? Could it be having some impact?
 

·
Registered
Joined
·
308 Posts

·
Facepalm
Joined
·
8,202 Posts
Discussion Starter #2,663
Thanks for your reply mate.

I will try, but if it's hitting 96C at 1.36 you can imagine how it will be at 1.43? EDIT: tried it at 1.43, CN quits 3 seconds after the test starts..

The fact that my room is really warm (around 30C) is not helping at all, but this is how bad it looks:

View attachment 2458890
Ok, that high VF explains that. That explains why you can't do 5.2 ghz...my SP94 is 90mv there, even though I've seen some SP 80's with a lower VF (both point and range) than that.
150mv VF between 5.1 to 5.2 is one of the definitely worse ones, but not the record.
I usually see SP63 as low as 130, but I've seen some around 80-90 too. Usually a SP63 with a high VF but good scaling from 5.1 to 5.2 (like 60-70mv) is actually awesome--you just need to delid those and they can take off. The absolute worst one ever is 180, that's beyond all hope. I actually saw an SP63 with a worse VF (the 180mv one) from 5.2-5.3 ghz than a SP51! Yes, the SP51 clocked better...
 

·
Registered
Joined
·
308 Posts

·
Registered
Joined
·
95 Posts
Just sharing the V/F curve from my SP71 10900KA.




This has OC’d as well as my golden SP63 (will be selling that one) but at slightly lower voltage. Still have my 52/49 OC in place but now at 1.41v bios voltage instead of 1.43v. This chip as well runs the same coolness at the golden SP63, roughly 220amps under full load.

To recap, I have tested roughly 13 10900k/ka at this point (one more on the way). All but two were SP63. But there was a huge variance in OC potential and how cool they ran between those SP63s. Some would OC fine to 5.2/4.9 and run cool (relatively speaking), others that would hit the same clocks but would run much hotter and some would not even hit 5.2 or 5.1 either. In fact one of my non-SP63 was an SP85 that wouldn’t go over 5.0 and ran hot!

Since my golden SP63 was one of my first chips I use it’s settings as a bar by which I grade the other CPUs. Aka what can they run 5.2/4.9 stable at? What is stable? At least 12hrs p95 small fft no avx with no Windows hardware errors. I essentially found that some of these 10900k just ran cooler than others at the same settings, and one easy way to categorize a cool chip from a hot chip was their amperage draw. The cooler chips pull roughly 10% less amps than the hot chips at the identical frequency and voltages. For example a cooler chip under full load would only pull 220-230amps in P95 while the hotter chips were in the 240-250amp range.


Sent from my iPhone using Tapatalk Pro
 

·
Facepalm
Joined
·
8,202 Posts
Discussion Starter #2,666
Just sharing the V/F curve from my SP71 10900KA.




This has OC’d as well as my golden SP63 (will be selling that one) but at slightly lower voltage. Still have my 52/49 OC in place but now at 1.41v bios voltage instead of 1.43v. This chip as well runs the same coolness at the golden SP63, roughly 220amps under full load.

To recap, I have tested roughly 13 10900k/ka at this point (one more on the way). All but two were SP63. But there was a huge variance in OC potential and how cool they ran between those SP63s. Some would OC fine to 5.2/4.9 and run cool (relatively speaking), others that would hit the same clocks but would run much hotter and some would not even hit 5.2 or 5.1 either. In fact one of my non-SP63 was an SP85 that wouldn’t go over 5.0 and ran hot!

Since my golden SP63 was one of my first chips I use it’s settings as a bar by which I grade the other CPUs. Aka what can they run 5.2/4.9 stable at? What is stable? At least 12hrs p95 small fft no avx with no Windows hardware errors. I essentially found that some of these 10900k just ran cooler than others at the same settings, and one easy way to categorize a cool chip from a hot chip was their amperage draw. The cooler chips pull roughly 10% less amps than the hot chips at the identical frequency and voltages. For example a cooler chip under full load would only pull 220-230amps in P95 while the hotter chips were in the 240-250amp range.


Sent from my iPhone using Tapatalk Pro
You said you used 1.41v bios set at 5.2 ghz, what loadline calibration did you use for the chips?
And what is the LLC and load temp of that chip you posted at 5.2 / 4.9 @ 1.41v ?
 

·
Registered
Joined
·
2,315 Posts
“Golden sp63”. What do you call my sp104?
If the results isn't there, it isn't golden ;)

Does youre sp104 pass Cinebench r20 @ 5600mhz under water, or maybe 5700mhz?
 

·
Registered
Joined
·
95 Posts
“Golden sp63”. What do you call my sp104?
I guess we would need a larger sample size of SP104 chips to test. Let me know when you have 10+ to test:)


Sent from my iPhone using Tapatalk Pro
 

·
Registered
Joined
·
95 Posts
You said you used 1.41v bios set at 5.2 ghz, what loadline calibration did you use for the chips?
And what is the LLC and load temp of that chip you posted at 5.2 / 4.9 @ 1.41v ?
Here are some stats during 30mins of P95 small fft no avx. This is with system loaded. I went back to check BIOS vcore and forget I had actually dialed it back to 1.39v @LLC6. The Aquacomputer Octo temp sensors correspond to the following:

1-2: water temp sensors
4: ambient air temp
soft temp1: Air/water delta

2458954
 

·
Registered
Joined
·
308 Posts
To recap, I have tested roughly 13 10900k/ka at this point (one more on the way). All but two were SP63. But there was a huge variance in OC potential and how cool they ran between those SP63s. Some would OC fine to 5.2/4.9 and run cool (relatively speaking), others that would hit the same clocks but would run much hotter and some would not even hit 5.2 or 5.1 either. In fact one of my non-SP63 was an SP85 that wouldn’t go over 5.0 and ran hot!
Your findings correspond well with what Der8auer found out too. I have considered testing out a new 10900K, but seeing your resultes and what Der8auer found out, the chances are that I just will get another SP63 chip. I could only hope I get one that runs ~10C cooler at the same settings as my current one... Don't know if it's worth the hassle. Then again, I could get lucky...
 

·
Registered
Joined
·
104 Posts
If the results isn't there, it isn't golden ;)

Does youre sp104 pass Cinebench r20 @ 5600mhz under water, or maybe 5700mhz?
No
If the results isn't there, it isn't golden ;)

Does youre sp104 pass Cinebench r20 @ 5600mhz under water, or maybe 5700mhz?
no pass r20 @5.6 for any voltage I throw at. Maybe coz my ambient is high . 27c and 30c water.
 

·
Registered
Joined
·
8 Posts
Ok, that high VF explains that. That explains why you can't do 5.2 ghz...my SP94 is 90mv there, even though I've seen some SP 80's with a lower VF (both point and range) than that.
150mv VF between 5.1 to 5.2 is one of the definitely worse ones, but not the record.
I usually see SP63 as low as 130, but I've seen some around 80-90 too. Usually a SP63 with a high VF but good scaling from 5.1 to 5.2 (like 60-70mv) is actually awesome--you just need to delid those and they can take off. The absolute worst one ever is 180, that's beyond all hope. I actually saw an SP63 with a worse VF (the 180mv one) from 5.2-5.3 ghz than a SP51! Yes, the SP51 clocked better...
So I talked with the store that sold it to me here in the UK and they said I can RMA it. Is it worth it in your opinion? What would you say are the probablity of getting a better one?

Thanks
 

·
Registered
Joined
·
8 Posts
By the way, would there be a highter power compumption and less OC potential when I'm using the integrated gpu? I am waiting for my graphics card still, so no gpu is installed yet.
 

·
Registered
Joined
·
2 Posts
hi, im pretty new to overclocking and have limited knowledge but recently gave it a shot on my 10900K.
i dont have the best chip it seems (SP 63) but was hoping i could manage 5.2.
currently on 5.1 at 1.35v and LLC 6.
is it possible to get 5.2 out of a SP 63?
ive tried going up in voltage and i need to hit around 1.45 at LLC 7 to get 5.2 out of this. im worried that the voltage and LLC setting is dangerous and may degrade my chip.
any thoughts would be greatly appreciated.
 

·
Registered
Joined
·
797 Posts
I read that the 10850k doesn't overclock as well as the 10900k
10850K is for "stock".
10850K is the worst binned 10 core. That's why it's cheaper. It' garbage bins. Often way under SP50 LOL

Buy cheap, buy twice ;)
I'm building a rig for a friend with a 10850K and with Apex XII 0707 it's SP72.

Super easy 5.2ghz all cores, 0 avx offset at @ 1.3v LLC8 (no vdroop). Only tested 1 hour Realbench 2.56 so far, but it's barely above 80C on a 360mm so have no doubts this will be good as is for gaming already.
 

·
Registered
Joined
·
947 Posts
I'm building a rig for a friend with a 10850K and with Apex XII 0707 it's SP72.

Super easy 5.2ghz all cores, 0 avx offset at @ 1.3v LLC8 (no vdroop). Only tested 1 hour Realbench 2.56 so far, but it's barely above 80C on a 360mm so have no doubts this will be good as is for gaming already.
Nice to hear a happy story (y)
 

·
Not New to Overclock.net
Joined
·
3,334 Posts
FFT112

10900k - SP81
M12E - Bios 098

51|48 - 1.273 v/f L6 vmin 1.15xx
2x16gb 4400 18-18-18-38 @1.40v
vccio/vcssa - 1.3/1.3
txp/ppd - 0/0


52|49 -1.373 v/f L6 vmin 1.24xx
2x16gb 4400 18-18-18-38 @1.40v
vccio/vcssa - 1.35/1.35
txp/ppd - 0/0

 

·
Registered
Joined
·
947 Posts
FFT112

10900k - SP81
M12E - Bios 098


51|48 - 1.273 v/f L6 vmin 1.15xx
2x16gb 4400 18-18-18-38 @1.40v
vccio/vcssa - 1.3/1.3
txp/ppd - 0/0


52|49 -1.373 v/f L6 vmin 1.24xx
2x16gb 4400 18-18-18-38 @1.40v
vccio/vcssa - 1.35/1.35
txp/ppd - 0/0


So better ram overclock with lower CPU overclock?
 

·
Not New to Overclock.net
Joined
·
3,334 Posts
So better ram overclock with lower CPU overclock?
u mean vcore. nope. same.
fft112 was really difficult.

afaik i am the first with 4400

daily now gaming is 53|50. vmin 1.28v with powerlimit at 52|49 arnd 315 watt..
its the ram now. should i keep 4400c18 or 4400c16??? 🤔 c16 requires 1.5v.. its above the ram rated [email protected] will test next week the 4266c17 32gb dual rank two dimm kits. i hope they can do run as well as this.. but kindda doubt it. being lucky twice.. dats stretching it..

c16 is worth it for latency. c17 vs c18.. marginal difference.

its in prep for rtx 3080/3090/3080ti/3080 super..
 
2661 - 2680 of 2704 Posts
Top