Discussion i9 9900K Benchmark Tests - First Run

Notorious^

Reputable
Feb 17, 2019
104
18
4,615
Got my new rig built and up and running and wanted to share my first set of benchmark tests with my new i9 9900K to see where i rank with other PC's running this CPU. So far i have not ran a Prime95 test, just CPU-Z and Cinebench. So please let me know what you think on my results and temp's.

i9 9900K 3.6 GHz - Stock Speed - Idle Temp 26c - CPU-Z Test Temp 53c

Cinebench - CPU = 2080 cb - OpenGL = 171.09 FPS

CPU-Z - Single Thread = 580 - Multi Thread (16) = 5662 - Multi Thread Rating = 9.76

===============================================================================

i9 9900K 3.6 GHz - OC'd 5 GHZ Voltage 1.32 Idle Temp 37c - CPU-Z Test Temp 85c

Cinebench - CPU=2184 cb - OpenGL = 193.33 FPS

CPU-Z - Single Thread = 594 - Multi Thread (16) = 5925 - Multi Thread Rating = 9.97
 

germanium

Distinguished
Jul 8, 2006
93
9
18,565
I have my Core I9-9900K in an Asus Strix Z390-E gaming motherboard with 5GHz all core boost with hyperthreading active. Running Optane memory RAID 0 with three drive connected to CPU so no discreet GPU.See https://forums.tomshardware.com/threads/my-testing-of-optane-memory-32-gb-in-raid-0.3424282/

My numbers for Cinebench aren't quite as good as yours but are nearly comparable. My CPU Z scores slightly better especially multithread.

Core I9-9900K 3.6GHz OC all cores @5GHz voltage 1.3000 except on Cinebench which is 1.35 volts peak test temp 87c Cinebench other wise 80c.

Cinebench best CPU=2117. Did not do Open GL No discreet GPU

CPU-Z single thread 595.7 Multi thread (16) 6130.2 Multi thread rating 10.29.

My Cinebench not stable under 3.5 volts. I have my voltages locked & only suffer a very small V Droop. Only a few thousandths of a volt droop.
 
Last edited:
  • Like
Reactions: Notorious^

germanium

Distinguished
Jul 8, 2006
93
9
18,565
My score on Cinebench were possibly as a result of throttling due to going over 85c on temperature. I could change the temp that it starts throttling to 90c but decided against it. Don't want to burn up my very expensive chip. Not worth it to me. Mine required more voltage than yours to get stable on Cinebench. Yours was on the ragged edge before throttling occurs. My CPU-Z temps were lower as I could run that stably at 1.3 volts & never went over 80c on temperature.

Our cooling systems are roughly equivalent even though mine is slightly smaller due to my fans running faster under load. Fully loaded my radiator fans run 3,000RPM.

Your CPU-Z scores probably were a result of throttling due to temperature being at 85c. Try backing down to 1.3 volts, if stable you will get a higher score possibly than me. Cinebench presents a much tougher load than CPU-Z tests do so you should be able to get away with lower voltage on CPU-z tests.
 
  • Like
Reactions: Notorious^

Notorious^

Reputable
Feb 17, 2019
104
18
4,615
My score on Cinebench were possibly as a result of throttling due to going over 85c on temperature. I could change the temp that it starts throttling to 90c but decided against it. Don't want to burn up my very expensive chip. Not worth it to me. Mine required more voltage than yours to get stable on Cinebench. Yours was on the ragged edge before throttling occurs. My CPU-Z temps were lower as I could run that stably at 1.3 volts & never went over 80c on temperature.

Our cooling systems are roughly equivalent even though mine is slightly smaller due to my fans running faster under load. Fully loaded my radiator fans run 3,000RPM.

Your CPU-Z scores probably were a result of throttling due to temperature being at 85c. Try backing down to 1.3 volts, if stable you will get a higher score possibly than me. Cinebench presents a much tougher load than CPU-Z tests do so you should be able to get away with lower voltage on CPU-z tests.

I will give that a try in just a minute. i busted out the 3D Mark last night as well. OC didn't really net any big gains to write home about as it was mostly GPU tests and one CPU benchmark.

I will report back on my CPU-Z benchmark at 1.3 volts.


3D Mark Time Spy

i9 9900K 3.6 GHz Stock Speed

Overall Score 6255

Graphics Score 5817
Test 1 37.80 FPS
Test 2 33.44 FPS

CPU Score 10 930
CPU Test 36.72 FPS

========================================================================

3D Mark Time Spy

i9 9900K 3.6 GHz OC'd 5 GHz

Overall Score 6338

Graphics Score 5873
Test 1 38.10 FPS
Test 2 33.81 FPS

CPU Score 11 516
CPU Test 38.69 FPS


========================================================================

3D Mark Fire Strike

i9 9900K 3.6GHz Stock Speed

Overall Score 16 039

Graphics Score 17 494
Graphics Test 1 82.44 FPS
Graphics Test 2 70.61 FPS

Physics Score 24 618
Physics Test 78.15 FPS

Combined Score 7 474
Combined Test 34.77 FPS

========================================================================

3D Mark Fire Strike

i9 9900K 3.6GHz OC'd 5GHz

Overall Score 16 197

Graphics Score 17 568
Graphics Test 1 82.82 FPS
Graphics Test 2 70.88 FPS

Physics Score 26 201
Physics Test 83.18 FPS

Combined Score 7 506
Combined Test 34.92 FPS
 
Last edited:

Notorious^

Reputable
Feb 17, 2019
104
18
4,615
Well i just could not break 6000 for a multi thread score on CPU-Z. i played with the Voltages a bit and could not hit the mark. So, i did what anyone would do, i overclocked more to 5.1 GHz @ 1.34v (LoL) and i finally broke through. I also ran another Cinebench @ 5.1 GHz and picked up a bit more performance.

Cinebench
CPU = 2219 cb
OpenGL = 197.82 FPS

CPU-Z
Single Thread = 607
Multi Thread (16)= 6048
Multi Thread Rating = 9.91
 
Last edited:
  • Like
Reactions: germanium

Notorious^

Reputable
Feb 17, 2019
104
18
4,615
f27eqr.jpg


nvnlnb.png
 
  • Like
Reactions: germanium

Notorious^

Reputable
Feb 17, 2019
104
18
4,615
Yeah i see that too. I'm using the MSI version. I'm going to go grab that version and do some more testing. I'm jealous you are cracking 6000 on 5GHz and i can't so far lol. But i will try that and see what happens.

So far i'm really happy with the build, i wished i would not have not "cheaped out" on a video card and got the 1080 Ti. The 1070 is no slouch but it's not a 1080 Ti either lol.
 
Last edited:

germanium

Distinguished
Jul 8, 2006
93
9
18,565
Hopefully you have a very healthy power supply for that 1080Ti.

Not only did I crack 6000 but by a healthy margin at that.

Also are you sure your doing an all core boost at 5.1GHz? If not you need to change some BIOS settings. Mine is truly running an all core boost at 5GHz. Only time it downclocks is if power goes over 165 for 127 seconds ( my setting, not stock set not to exeed185 watts) or if I go over 85c temp or doing AVX which I set for 4.8GHz all core boost (my setting).
 
Last edited:

germanium

Distinguished
Jul 8, 2006
93
9
18,565
Those settings look good.

One thing to look at is the FCLK which most companies set to 800MHz. Intel spec is 1000MHz. Note I also run my ring frequency at 4.8GHz. This the highest I can run it at current voltage of 1.300 setting. Going over this means ramping up voltage much quicker than ramping up the multiplier for core frequency.
 
Last edited:

germanium

Distinguished
Jul 8, 2006
93
9
18,565
It really depends on what benchmark you look at. Also motherboard plays a key roll as well. Sometimes A bios update cures initial release issues such as on my board excessive ramp up of voltage as CPU got hotter which in one case caused a boot failure due to overvoltage when on auto settings.

That was when I tried for 5GHz ring frequency. Did not try that again after bios update but did notice lower temps when on auto setting afterwards as well as lower power draw.

When tried for 4.9GHz voltage went up to 1.5 volts on auto but could run with reasonable stability except Cinebench @1.375 volts @4.9HGz with manual voltage settings. Idle temps were over 40c. Backed down to 4.8 @1.300 volts & idle temps stay below 40c

I made some mistakes when I posted these frequencies earlier. I'm actually running the ring frequency at 4.7GHz instead of 4.8GHz as I reported when I made this post. Sorry I was trying to go by my memory which is not all that great.
 
Last edited:

germanium

Distinguished
Jul 8, 2006
93
9
18,565

Notorious^

Reputable
Feb 17, 2019
104
18
4,615
Well i'm throwing in the towel on the 6000 multi thread score goal @5GHz lol. I just can't hit it without going to 5.1GHz.

Best it looks like i'm going to get is the 5920 range. 5929 is the best score i have been able to hit. I changed FCLK to 1000MHz and tried playing with the voltages and just can't make this puppy pull anything better than 5920's.


Maybe my chip doesn't like my mobo lol.