• Welcome to TechPowerUp Forums, Guest! Please check out our forum guidelines for info related to our community.

3900X poo ST Cinebench results

Joined
Mar 10, 2015
Messages
2,926 (1.58/day)
System Name Wut?
Processor 3900X
Motherboard ASRock Taichi X570
Cooling Water
Memory 32GB GSkill CL16 3600mhz
Video Card(s) Vega 56
Storage 2 x AData XPG 8200 Pro 1TB
Display(s) 3440 x 1440
Case Thermaltake Tower 900
Power Supply Seasonic Prime Ultra Platinum
I have just started getting around to checking my performance with my 3900X. My results are as follow:

ST: 497
MT: 7340

PBO settings are the default that came with my Taichi X570. I believe it is set to "Enabled" and all the values are 0. It so far has given me my best multi-thread value. So much to dig in to with PBO...

I am still on 1.0.0.3 ABBA so I am going to pop 1.0.0.4 up and see what kind of difference I get. I am on Win 10 1909 so I should have the new Ryzen scheduler rolling. I haven't done 1usmus's power plan yet either so I will likely try that out.

What I did notice is that when running Cinebench ST, the thread it was using was on the 2nd ccx which is the slowest on mine. Cores 0 - 5 all can boost to 4599 - to 4649 so I was fairly surprised that the second ccx was being used as I thought that scheduler knows to send to the best ccx regardless of 1usmus's power plan installed. My multi-thread score seems pretty inline with what a stock config score should be so perhaps throwing the new power plan and stepping up to 1.0.0.4 is all I really need to do.

What else have all the Zen 2 people done?

Edit:

Also, I am still reading through some of the Zen 2 threads so I may find what I need. Mods, also feel free to throw this into another thread if you see fit. I just didn't want to hijack another's.
 
Joined
Jun 24, 2015
Messages
1,474 (0.84/day)
Location
Western Canada
System Name Ol' Beastie | 雨上がりの空
Processor R7 3700X | i7-4790K
Motherboard B450I Aorus (F50) | H97N-WIFI (F7)
Cooling NH-U9S | NH-C14S
Memory 32GB Trident Z RGB 3600 (16-19-19-32, DJR) | 16GB Fury 1600 (10-11-10-30)
Video Card(s) NVIDIA RTX 2060 Super FE | EVGA GTX 1070 SC ACX3.0
Storage 1TB SX8200 - 750GB MX300 - 1TB Blue3D | 500GB 850 EVO
Display(s) BenQ GW2765HT | Samsung F2380
Case NCASE M1v5 + Pelican 1514 | TJ08T-E
Audio Device(s) DT 770 80Ω - Blue Snowball iCE
Power Supply Corsair SF600G/Cablemod + APC BR1500MS| Seasonic SSR-550PX
Mouse Endgame XM1 | Logitech MX Master
Keyboard Noxary X60 R - Noxary T60 Night Blue - Noxary T60 R2 Hyper Red - RAMA M10-B
Software Windows 10 1909
You can chime in on @lorry's thread if he don't mind; he has a 3900X that he's been trying to optimize for the past couple of months. Your MT score isn't anything out of the ordinary, which is good.

I'm not aware of any way you can force CB to run on a better core, just like how CPU-Z is still directed to use Core 0 for its ST test. 1.0.0.4B, latest chipset drivers, 1909, and whatever plan you like between Ryzen Balanced and 1usmus...there's not much you can do.

That said, your PBO may allow you to push boost clocks up a bit. Running either "balls-to-the-wall" manual PBO settings or a middle of the road 2X scalar and -7A EDC allows me to boost past rated boost occasionally (44.3x) on my best cores. You can play around with the scalar in there to see if it makes a difference. On mine, CPU-Z suffers as Core 0 is my worst, but PBO helps it boost a little higher as well (usually stuck at 42.8x).

On my 3700X, CB seems to gravitate towards one of my two best cores which run the test at close to 44x. Score is 507. Ideally, 3900X generally runs ST in the 520s.

The ST test also takes a little while. Make sure there's nothing to distract the system while the test completes, including sleep/screen off, Dropbox, and other background programs that might chime in with notifications/other unwanted usage. CB hates Dropbox on my rig.
 
Joined
Mar 23, 2016
Messages
3,643 (2.46/day)
Processor Ryzen 5 3600
Motherboard MSI B450 Tomahawk ATX
Cooling Scythe Kotetsu with AM4 bracket
Memory PNY Anarchy-X XLR8 Red DDR4 3200 MHz C15-17-17-17-35
Video Card(s) MSI GeForce RTX 2060 GAMING Z 6G
Storage Samsung 970 EVO NVMe M.2 500 GB, SanDisk Ultra II 480 GB
Display(s) Samsung SyncMaster C27H711 OC refresh rate 110Hz
Case Phantek Eclipse P400S (PH-EC416PS)
Audio Device(s) EVGA NU Audio
Power Supply EVGA 850 BQ
Mouse SteelSeries Rival 310
Keyboard Logitech G G413 Silver
Software Windows 10 Professional 64-bit v1903
What I did notice is that when running Cinebench ST, the thread it was using was on the 2nd ccx which is the slowest on mine. Cores 0 - 5 all can boost to 4599 - to 4649 so I was fairly surprised that the second ccx was being used as I thought that scheduler knows to send to the best ccx regardless of 1usmus's power plan installed.
Do you have the option/s in the BIOS called Collaborative Power Performance Control turned on?

It supposedly allows the Windows Scheduler to know which is the better core to boost.

I was wrong the Intel equivalent is Speedshift.

"With "Zen 2", AMD introduced a platform feature called UEFI CPPC2 (UEFI collaborative power and performance control). This capability puts the processor's on-die firmware in complete control of clock-speed/voltage selection at all times."

 
Last edited:
Joined
Mar 10, 2015
Messages
2,926 (1.58/day)
System Name Wut?
Processor 3900X
Motherboard ASRock Taichi X570
Cooling Water
Memory 32GB GSkill CL16 3600mhz
Video Card(s) Vega 56
Storage 2 x AData XPG 8200 Pro 1TB
Display(s) 3440 x 1440
Case Thermaltake Tower 900
Power Supply Seasonic Prime Ultra Platinum
You can chime in on @lorry's thread if he don't mind; he has a 3900X that he's been trying to optimize for the past couple of months. Your MT score isn't anything out of the ordinary, which is good.

I'm not aware of any way you can force CB to run on a better core, just like how CPU-Z is still directed to use Core 0 for its ST test. 1.0.0.4B, latest chipset drivers, 1909, and whatever plan you like between Ryzen Balanced and 1usmus...there's not much you can do.

That said, your PBO may allow you to push boost clocks up a bit. Running either "balls-to-the-wall" manual PBO settings or a middle of the road 2X scalar and -7A EDC allows me to boost past rated boost occasionally (44.3x) on my best cores. You can play around with the scalar in there to see if it makes a difference. On mine, CPU-Z suffers as Core 0 is my worst, but PBO helps it boost a little higher as well (usually stuck at 42.8x).

On my 3700X, CB seems to gravitate towards one of my two best cores which run the test at close to 44x. Score is 507. Ideally, 3900X generally runs ST in the 520s.

The ST test also takes a little while. Make sure there's nothing to distract the system while the test completes, including sleep/screen off, Dropbox, and other background programs that might chime in with notifications/other unwanted usage. CB hates Dropbox on my rig.
I am hesitant to mess with PBO too much as 6 cores will already boost at or above 4.6 and my others are 4.3 to 4.4.
Do you have the option/s in the BIOS called Collaborative Power Performance Control turned on?

It supposedly allows the Windows Scheduler to know which is the better core to boost.
I'll need to check that.
 
Joined
Sep 10, 2018
Messages
1,035 (1.79/day)
Location
California
System Name His & Hers
Processor i9 9900k @5ghz/ R9 3900X Stock
Motherboard Asus Maximus XI Code/ X570 Aorus Master
Cooling Corsair h150i Pro Push/Pull. 5x Corsair ML120 + 4x Corsair LL120/ Corsair h115i Platinum
Memory 32 GB 4x8GB 3600mhz Corsair RGB Pro/ 32 GB 3200 CL14 @3800 CL16
Video Card(s) Asus ROG Strix 2080 ti/ Titan XP
Storage lots of SSD.
Display(s) Asus ROG PG278QR / MSI 27 inch VA panel.
Case Corsair 500D SE RGB/ Phanteks Evolv X
Audio Device(s) Arctis Pro + gaming Dac/ Corsair sp 2500/ Logitech Z 625
Power Supply Seasonic Ultra Prime Titanium 850w x2
Mouse Logitech G903 Lightspeed/ Logitech G Pro Hero.
Keyboard Corsair K95 RGB Platinum/ Logitech G Pro
Your MT score looks pretty good for mostly stock ST should be around 520+ though so for whatever reason your system seems to be using a lesser core/thread. At the same time I wouldn't be overly concerned it will make zero difference on your MT performance.
It could also just be something running in the background or a specific process weighing you down. I've noticed that depending on how the boards auto settings are PBO can slightly boost MT performance while slightly decreasing ST performance. I have to manually disable it for best ST performance on my Aorus Master.
 
Joined
Mar 10, 2015
Messages
2,926 (1.58/day)
System Name Wut?
Processor 3900X
Motherboard ASRock Taichi X570
Cooling Water
Memory 32GB GSkill CL16 3600mhz
Video Card(s) Vega 56
Storage 2 x AData XPG 8200 Pro 1TB
Display(s) 3440 x 1440
Case Thermaltake Tower 900
Power Supply Seasonic Prime Ultra Platinum
Your MT score looks pretty good for mostly stock ST should be around 520+ though so for whatever reason your system seems to be using a lesser core/thread. At the same time I wouldn't be overly concerned it will make zero difference on your MT performance.
It could also just be something running in the background or a specific process weighing you down. I've noticed that depending on how the boards auto settings are PBO can slightly boost MT performance while slightly decreasing ST performance. I have to manually disable it for best ST performance on my Aorus Master.
Yeah, not so much concerned as I thought it was odd that the scheduler in 1909 was aware of which cores/ccxs should be scheduled first. It seems likely that I don't have the proper setting set in the bios.
 
Joined
Dec 3, 2019
Messages
647 (5.05/day)
Location
Crawley
System Name 3900x
Processor AMD Ryzen 3900x
Motherboard Gigabyte Aorus X470 gaming 7 WiFi
Cooling Noctua NH DI-15S soon Alphacool Eisbaer 360 with Noctua redux fans, then an open loop
Memory 32GB (4 x 8GB) Corsair 3200 C16 Dominator Platinum OC to 3733 16-16-16-32-48
Video Card(s) Gigabyte Aorus GeForce® RTX 2070 Super
Storage Samsung Evo 970 1TB NVMe M.2 SSD as boot drive, Intel 660P 2TB NVMe M.2 SSD second drive
Display(s) AOC Q3279VWFD8
Case Cooler Master H500M mesh front soon Thermaltaje View 7
Audio Device(s) Sound Blaster G6
Power Supply Seasonic Titanium 1000W with cable mods cables
Mouse Corsair Dark Core SE & Logitech G815 Keyboard
Keyboard Logitech G815
Software wIn 10 Pro possibly Manjaro at some point
You can chime in on @lorry's thread if he don't mind; he has a 3900X that he's been trying to optimize for the past couple of months. Your MT score isn't anything out of the ordinary, which is good.

I'm not aware of any way you can force CB to run on a better core, just like how CPU-Z is still directed to use Core 0 for its ST test. 1.0.0.4B, latest chipset drivers, 1909, and whatever plan you like between Ryzen Balanced and 1usmus...there's not much you can do.

That said, your PBO may allow you to push boost clocks up a bit. Running either "balls-to-the-wall" manual PBO settings or a middle of the road 2X scalar and -7A EDC allows me to boost past rated boost occasionally (44.3x) on my best cores. You can play around with the scalar in there to see if it makes a difference. On mine, CPU-Z suffers as Core 0 is my worst, but PBO helps it boost a little higher as well (usually stuck at 42.8x).

On my 3700X, CB seems to gravitate towards one of my two best cores which run the test at close to 44x. Score is 507. Ideally, 3900X generally runs ST in the 520s.

The ST test also takes a little while. Make sure there's nothing to distract the system while the test completes, including sleep/screen off, Dropbox, and other background programs that might chime in with notifications/other unwanted usage. CB hates Dropbox on my rig.
I've no objection at all, if you find anything of use to you on there, great.
Join in if you want, or whatever works best for you.
BTW, my best scores are 514/7414, but it took quite a while to get to that and your CB scores are very typical.
Theresa thread i think either on here or on AMD forums re CB score.
 
Joined
Sep 3, 2019
Messages
1,113 (5.08/day)
Location
Thessaloniki, Greece
System Name None
Processor Ryzen 5 3600 (PBO manual: PPT 105W, TDC 105A, EDC 1A, PBO Scalar auto)
Motherboard Gigabyte X570 Aorus Pro
Cooling Corsair H110i 280mm (Liquid metal for TIM)
Memory CorsairVengeanceLPX DDR4 2x8GB 3466MHz CL16-18-18-36 1T, cheap B-die, @3666 CL16-17-17-34 1T (1:1:1)
Video Card(s) MSI RX 5700XT Gaming X
Storage Samsung NVMe: 970Pro 512GB / SATA-III: 850Pro 1TB SSD, 860Evo 1TB SSD
Display(s) 24" EIZO FlexScan S2411W 1920x1200, 16:10 60Hz samsung S-PVA panel.
Case None
Audio Device(s) Sound Blaster Z <--optical link--> Logitech Z5500 5.1 500W
Power Supply Corsair HX750i
Mouse Logitech MX Master
Keyboard Logitech G15 v2
Software Windows 10 Home 64bit (v1909)
Do you have the option/s in the BIOS called Collaborative Power Performance Control turned on?

It supposedly allows the Windows Scheduler to know which is the better core to boost.

I was wrong the Intel equivalent is Speedshift.

"With "Zen 2", AMD introduced a platform feature called UEFI CPPC2 (UEFI collaborative power and performance control). This capability puts the processor's on-die firmware in complete control of clock-speed/voltage selection at all times."

This is a follow up to that article...

I have just started getting around to checking my performance with my 3900X. My results are as follow:

ST: 497
MT: 7340

PBO settings are the default that came with my Taichi X570. I believe it is set to "Enabled" and all the values are 0. It so far has given me my best multi-thread value. So much to dig in to with PBO...

I am still on 1.0.0.3 ABBA so I am going to pop 1.0.0.4 up and see what kind of difference I get. I am on Win 10 1909 so I should have the new Ryzen scheduler rolling. I haven't done 1usmus's power plan yet either so I will likely try that out.

What I did notice is that when running Cinebench ST, the thread it was using was on the 2nd ccx which is the slowest on mine. Cores 0 - 5 all can boost to 4599 - to 4649 so I was fairly surprised that the second ccx was being used as I thought that scheduler knows to send to the best ccx regardless of 1usmus's power plan installed. My multi-thread score seems pretty inline with what a stock config score should be so perhaps throwing the new power plan and stepping up to 1.0.0.4 is all I really need to do.

What else have all the Zen 2 people done?

Edit:

Also, I am still reading through some of the Zen 2 threads so I may find what I need. Mods, also feel free to throw this into another thread if you see fit. I just didn't want to hijack another's.
Regardless of 1usmus's power plan (if you install it or not) you should do the following if not already.

1. Letest AGESA (BIOS/UEFI version) for your board.
2. Win10 v1909
3. Latest chipset drivers (November) directly from AMD (not Asrock).

UEFI settings
  • Global C-state Control = Enabled
  • Power Supply Idle Control = Low Current Idle
  • CPPC = Enabled
  • CPPC Preferred Cores = Enabled
  • AMD Cool'n'Quiet = Enabled
  • PPC Adjustment = PState 0
optional:
4. 1usmus's custom PowerPlan v1.1 (for win v1909) for Ryzen (ZEN2) CPUs. (definately need those UEFI settings)

PBO settings are not as scary as it seems, and just raising all 3 individual limits of PPT/TDC/EDC accomplishes nothing (positive or negative).
PBO Scalar on the other hand does but first thing first...

Know you CPU...
After the above 1,2,3 at least (I strongly suggest 4. too) open HWiNFO64 (latest ver.) and look the "Core 1/2/3... clock (perf #x/x)"
Second (/x) is the order of best quality cores reported by the CPU itself. First (x/) is the order Win10 scheduler chooses to load. It may be the same in some CPUs but also very different to others.

Look screenshot below...
This is my 3600 after 3 hours of browsing and YT videos, nothing heavy load. If all is working as intented it should look like this. Notice the perf# order x/x and then the last (avg) column of effective clock. You can see what threads getting the most load.

HWiNFO_14_01_2020b.png


The CPU has 1 CCD with 2 CCXs.
Core 0/1/2 is CCX1
Core 3/4/5 is CCX2

Core 0 T0 is a physical core thread
Core 0 T1 is a logical core thread
and so on...

You can see in "eff clock" red box that cores 0/1/2 works more than 3/4/5 because win scheduler chooses so...
The effective clock in general contains the sleeping states of the cores. The higher eff clock a core has, the lesser in sleep state it is. (that sounds like yoda talking...)
The arising question is why win scheduler prefers cores 0/1/2 instead of 1/2/3, the 3 best of all. The answer is rather simple... Its trying to keep loads in 1 CCX as possible for lower latency.
Things get a lot more complicated when a CPU has 2 CCDs and 4 CCXs and the highest quality cores are scattered across all 4 CCXs...
This kind of behaviour is applied to every non all core load. From 1 single thread load, up to -1~2 threads from all threaded load.

What is the situation with yours?
 
Joined
Mar 10, 2015
Messages
2,926 (1.58/day)
System Name Wut?
Processor 3900X
Motherboard ASRock Taichi X570
Cooling Water
Memory 32GB GSkill CL16 3600mhz
Video Card(s) Vega 56
Storage 2 x AData XPG 8200 Pro 1TB
Display(s) 3440 x 1440
Case Thermaltake Tower 900
Power Supply Seasonic Prime Ultra Platinum
This is a follow up to that article...


Regardless of 1usmus's power plan (if you install it or not) you should do the following if not already.

1. Letest AGESA (BIOS/UEFI version) for your board.
2. Win10 v1909
3. Latest chipset drivers (November) directly from AMD (not Asrock).

UEFI settings
  • Global C-state Control = Enabled
  • Power Supply Idle Control = Low Current Idle
  • CPPC = Enabled
  • CPPC Preferred Cores = Enabled
  • AMD Cool'n'Quiet = Enabled
  • PPC Adjustment = PState 0
optional:
4. 1usmus's custom PowerPlan v1.1 (for win v1909) for Ryzen (ZEN2) CPUs. (definately need those UEFI settings)
2 and 3 are done. I am suspecting that some bios settings are not set (or HWInfo is a few versions behind) because I do not have entries for effective clock. I'll take a gander now.
 
Joined
Dec 3, 2019
Messages
647 (5.05/day)
Location
Crawley
System Name 3900x
Processor AMD Ryzen 3900x
Motherboard Gigabyte Aorus X470 gaming 7 WiFi
Cooling Noctua NH DI-15S soon Alphacool Eisbaer 360 with Noctua redux fans, then an open loop
Memory 32GB (4 x 8GB) Corsair 3200 C16 Dominator Platinum OC to 3733 16-16-16-32-48
Video Card(s) Gigabyte Aorus GeForce® RTX 2070 Super
Storage Samsung Evo 970 1TB NVMe M.2 SSD as boot drive, Intel 660P 2TB NVMe M.2 SSD second drive
Display(s) AOC Q3279VWFD8
Case Cooler Master H500M mesh front soon Thermaltaje View 7
Audio Device(s) Sound Blaster G6
Power Supply Seasonic Titanium 1000W with cable mods cables
Mouse Corsair Dark Core SE & Logitech G815 Keyboard
Keyboard Logitech G815
Software wIn 10 Pro possibly Manjaro at some point
2 and 3 are done. I am suspecting that some bios settings are not set (or HWInfo is a few versions behind) because I do not have entries for effective clock. I'll take a gander now.
latest version is 6.21-4040


Your MT score looks pretty good for mostly stock ST should be around 520+ though so for whatever reason your system seems to be using a lesser core/thread. At the same time I wouldn't be overly concerned it will make zero difference on your MT performance.
It could also just be something running in the background or a specific process weighing you down. I've noticed that depending on how the boards auto settings are PBO can slightly boost MT performance while slightly decreasing ST performance. I have to manually disable it for best ST performance on my Aorus Master.
I think that my MB is limiting my ST score (Gigabyte Aorus x470 gaming 7) as my best ST is 514, and/or a heavy process
 
Joined
Sep 10, 2018
Messages
1,035 (1.79/day)
Location
California
System Name His & Hers
Processor i9 9900k @5ghz/ R9 3900X Stock
Motherboard Asus Maximus XI Code/ X570 Aorus Master
Cooling Corsair h150i Pro Push/Pull. 5x Corsair ML120 + 4x Corsair LL120/ Corsair h115i Platinum
Memory 32 GB 4x8GB 3600mhz Corsair RGB Pro/ 32 GB 3200 CL14 @3800 CL16
Video Card(s) Asus ROG Strix 2080 ti/ Titan XP
Storage lots of SSD.
Display(s) Asus ROG PG278QR / MSI 27 inch VA panel.
Case Corsair 500D SE RGB/ Phanteks Evolv X
Audio Device(s) Arctis Pro + gaming Dac/ Corsair sp 2500/ Logitech Z 625
Power Supply Seasonic Ultra Prime Titanium 850w x2
Mouse Logitech G903 Lightspeed/ Logitech G Pro Hero.
Keyboard Corsair K95 RGB Platinum/ Logitech G Pro
latest version is 6.21-4040




I think that my MB is limiting my ST score (Gigabyte Aorus x470 gaming 7) as my best ST is 514, and/or a heavy process
That's less than 3% though of what I typically see which I'd consider margin of error
 
Joined
Dec 3, 2019
Messages
647 (5.05/day)
Location
Crawley
System Name 3900x
Processor AMD Ryzen 3900x
Motherboard Gigabyte Aorus X470 gaming 7 WiFi
Cooling Noctua NH DI-15S soon Alphacool Eisbaer 360 with Noctua redux fans, then an open loop
Memory 32GB (4 x 8GB) Corsair 3200 C16 Dominator Platinum OC to 3733 16-16-16-32-48
Video Card(s) Gigabyte Aorus GeForce® RTX 2070 Super
Storage Samsung Evo 970 1TB NVMe M.2 SSD as boot drive, Intel 660P 2TB NVMe M.2 SSD second drive
Display(s) AOC Q3279VWFD8
Case Cooler Master H500M mesh front soon Thermaltaje View 7
Audio Device(s) Sound Blaster G6
Power Supply Seasonic Titanium 1000W with cable mods cables
Mouse Corsair Dark Core SE & Logitech G815 Keyboard
Keyboard Logitech G815
Software wIn 10 Pro possibly Manjaro at some point
That's less than 3% though of what I typically see which I'd consider margin of error
Trouble is there are so many variables between users. One may have several manufacturers' software running whilst another has no RGB and controls their fans via the MB

Just out of interest, is there any software which would close multiple apps and background services for a benchmark and then turn them back on afterwards?
 
Joined
Sep 10, 2018
Messages
1,035 (1.79/day)
Location
California
System Name His & Hers
Processor i9 9900k @5ghz/ R9 3900X Stock
Motherboard Asus Maximus XI Code/ X570 Aorus Master
Cooling Corsair h150i Pro Push/Pull. 5x Corsair ML120 + 4x Corsair LL120/ Corsair h115i Platinum
Memory 32 GB 4x8GB 3600mhz Corsair RGB Pro/ 32 GB 3200 CL14 @3800 CL16
Video Card(s) Asus ROG Strix 2080 ti/ Titan XP
Storage lots of SSD.
Display(s) Asus ROG PG278QR / MSI 27 inch VA panel.
Case Corsair 500D SE RGB/ Phanteks Evolv X
Audio Device(s) Arctis Pro + gaming Dac/ Corsair sp 2500/ Logitech Z 625
Power Supply Seasonic Ultra Prime Titanium 850w x2
Mouse Logitech G903 Lightspeed/ Logitech G Pro Hero.
Keyboard Corsair K95 RGB Platinum/ Logitech G Pro
Trouble is there are so many variables between users. One may have several manufacturers' software running whilst another has no RGB and controls their fans via the MB

Just out of interest, is there any software which would close multiple apps and background services for a benchmark and then turn them back on afterwards?
Probably, I've usually never concern myself with that stuff even my 9900k and 3900X results I've posted here have icue and other crap running in the backround..... I was determined once to get over 5400 with my 9900k and turned everything else off but other than that I haven't cared much.
 
Joined
Dec 3, 2019
Messages
647 (5.05/day)
Location
Crawley
System Name 3900x
Processor AMD Ryzen 3900x
Motherboard Gigabyte Aorus X470 gaming 7 WiFi
Cooling Noctua NH DI-15S soon Alphacool Eisbaer 360 with Noctua redux fans, then an open loop
Memory 32GB (4 x 8GB) Corsair 3200 C16 Dominator Platinum OC to 3733 16-16-16-32-48
Video Card(s) Gigabyte Aorus GeForce® RTX 2070 Super
Storage Samsung Evo 970 1TB NVMe M.2 SSD as boot drive, Intel 660P 2TB NVMe M.2 SSD second drive
Display(s) AOC Q3279VWFD8
Case Cooler Master H500M mesh front soon Thermaltaje View 7
Audio Device(s) Sound Blaster G6
Power Supply Seasonic Titanium 1000W with cable mods cables
Mouse Corsair Dark Core SE & Logitech G815 Keyboard
Keyboard Logitech G815
Software wIn 10 Pro possibly Manjaro at some point
Probably, I've usually never concern myself with that stuff even my 9900k and 3900X results I've posted here have icue and other crap running in the backround..... I was determined once to get over 5400 with my 9900k and turned everything else off but other than that I haven't cared much.
I was thinking more about during gameplay than benchmarking, but effectively they are the same reason here
 
Joined
Sep 10, 2018
Messages
1,035 (1.79/day)
Location
California
System Name His & Hers
Processor i9 9900k @5ghz/ R9 3900X Stock
Motherboard Asus Maximus XI Code/ X570 Aorus Master
Cooling Corsair h150i Pro Push/Pull. 5x Corsair ML120 + 4x Corsair LL120/ Corsair h115i Platinum
Memory 32 GB 4x8GB 3600mhz Corsair RGB Pro/ 32 GB 3200 CL14 @3800 CL16
Video Card(s) Asus ROG Strix 2080 ti/ Titan XP
Storage lots of SSD.
Display(s) Asus ROG PG278QR / MSI 27 inch VA panel.
Case Corsair 500D SE RGB/ Phanteks Evolv X
Audio Device(s) Arctis Pro + gaming Dac/ Corsair sp 2500/ Logitech Z 625
Power Supply Seasonic Ultra Prime Titanium 850w x2
Mouse Logitech G903 Lightspeed/ Logitech G Pro Hero.
Keyboard Corsair K95 RGB Platinum/ Logitech G Pro
I was thinking more about during gameplay than benchmarking, but effectively they are the same reason here

Even with a ton of Crap running in the backround my 3900X doesn't feel any different than my 5ghz 9900k at 1440p with a 2080 ti sure if I use RivaTuner and look at a fps and graph the 9900k is a little higher at times but not enough for a blind test.... To me half the reason to get a high core count chip like this is so that you don't have to worry much about what is or isn't running in the background.

If you're a 1080p esport gamer and do this professionally then I could get behind limiting background stuff or have a pathetically slow cpu that barely meets the games requirements, but short of having a 2080 ti at 1080p it's just not going to matter in the majority of games.
 
Joined
Dec 3, 2019
Messages
647 (5.05/day)
Location
Crawley
System Name 3900x
Processor AMD Ryzen 3900x
Motherboard Gigabyte Aorus X470 gaming 7 WiFi
Cooling Noctua NH DI-15S soon Alphacool Eisbaer 360 with Noctua redux fans, then an open loop
Memory 32GB (4 x 8GB) Corsair 3200 C16 Dominator Platinum OC to 3733 16-16-16-32-48
Video Card(s) Gigabyte Aorus GeForce® RTX 2070 Super
Storage Samsung Evo 970 1TB NVMe M.2 SSD as boot drive, Intel 660P 2TB NVMe M.2 SSD second drive
Display(s) AOC Q3279VWFD8
Case Cooler Master H500M mesh front soon Thermaltaje View 7
Audio Device(s) Sound Blaster G6
Power Supply Seasonic Titanium 1000W with cable mods cables
Mouse Corsair Dark Core SE & Logitech G815 Keyboard
Keyboard Logitech G815
Software wIn 10 Pro possibly Manjaro at some point
Even with a ton of Crap running in the backround my 3900X doesn't feel any different than my 5ghz 9900k at 1440p with a 2080 ti sure if I use RivaTuner and look at a fps and graph the 9900k is a little higher at times but not enough for a blind test.... To me half the reason to get a high core count chip like this is so that you don't have to worry much about what is or isn't running in the background.

If you're a 1080p esport gamer and do this professionally then I could get behind limiting background stuff, but short of having a 2080 ti at 1080p it's just not going to matter in the majority of games.
I tend to use an app called FPS Monitor I think it's called, as you can configure it to show pretty much what you want at any time on screen, gameplay not needed.
I have one profile set up to show CPU, GPU & Ram fans speeds clock speeds, temps & and voltage.


FPS monitor.jpg
 
Last edited:
Joined
Sep 3, 2019
Messages
1,113 (5.08/day)
Location
Thessaloniki, Greece
System Name None
Processor Ryzen 5 3600 (PBO manual: PPT 105W, TDC 105A, EDC 1A, PBO Scalar auto)
Motherboard Gigabyte X570 Aorus Pro
Cooling Corsair H110i 280mm (Liquid metal for TIM)
Memory CorsairVengeanceLPX DDR4 2x8GB 3466MHz CL16-18-18-36 1T, cheap B-die, @3666 CL16-17-17-34 1T (1:1:1)
Video Card(s) MSI RX 5700XT Gaming X
Storage Samsung NVMe: 970Pro 512GB / SATA-III: 850Pro 1TB SSD, 860Evo 1TB SSD
Display(s) 24" EIZO FlexScan S2411W 1920x1200, 16:10 60Hz samsung S-PVA panel.
Case None
Audio Device(s) Sound Blaster Z <--optical link--> Logitech Z5500 5.1 500W
Power Supply Corsair HX750i
Mouse Logitech MX Master
Keyboard Logitech G15 v2
Software Windows 10 Home 64bit (v1909)
Really the typical user should not care about background unless there is some short of bottleneck, Ram usage (%) or CPU cores/threads. I dont care about background tasks even with ½ cores and threads of a 3900X. Cant tell in desktop or in gaming if the system is running 50 or 150 services. Only benchmark scores are affected by them and by a small margin. The system feels snappy and smooth on all scenarios.
 
Joined
Mar 10, 2015
Messages
2,926 (1.58/day)
System Name Wut?
Processor 3900X
Motherboard ASRock Taichi X570
Cooling Water
Memory 32GB GSkill CL16 3600mhz
Video Card(s) Vega 56
Storage 2 x AData XPG 8200 Pro 1TB
Display(s) 3440 x 1440
Case Thermaltake Tower 900
Power Supply Seasonic Prime Ultra Platinum
Alright, so after plenty of digging through the Bios....

Cppc was disabled but cppc preferred was enabled. Imagine that. Flipping the switch brought single threaded up to 524. Another mem tweak puts my multithreaded up to 7360.
 
Joined
Sep 3, 2019
Messages
1,113 (5.08/day)
Location
Thessaloniki, Greece
System Name None
Processor Ryzen 5 3600 (PBO manual: PPT 105W, TDC 105A, EDC 1A, PBO Scalar auto)
Motherboard Gigabyte X570 Aorus Pro
Cooling Corsair H110i 280mm (Liquid metal for TIM)
Memory CorsairVengeanceLPX DDR4 2x8GB 3466MHz CL16-18-18-36 1T, cheap B-die, @3666 CL16-17-17-34 1T (1:1:1)
Video Card(s) MSI RX 5700XT Gaming X
Storage Samsung NVMe: 970Pro 512GB / SATA-III: 850Pro 1TB SSD, 860Evo 1TB SSD
Display(s) 24" EIZO FlexScan S2411W 1920x1200, 16:10 60Hz samsung S-PVA panel.
Case None
Audio Device(s) Sound Blaster Z <--optical link--> Logitech Z5500 5.1 500W
Power Supply Corsair HX750i
Mouse Logitech MX Master
Keyboard Logitech G15 v2
Software Windows 10 Home 64bit (v1909)
Alright, so after plenty of digging through the Bios....

Cppc was disabled but cppc preferred was enabled. Imagine that. Flipping the switch brought single threaded up to 524. Another mem tweak puts my multithreaded up to 7360.
Nice!
Now if you like post a sreenshot of HWiNFO while CB-R20, like this

HWiNFO_06_01_2020_99_60_63_x3_newpaste_cool.png

To have it identical as mine you should open HWiNFO and R20, start the bench and as soon as it starts filling boxes and progress hit the clock down right on HW window to reset values. Right before it finishes take the screenshot (I believe yours will take about ~40sec to finish). This way you will isolate values to the bench time only...
 
Joined
Dec 3, 2019
Messages
647 (5.05/day)
Location
Crawley
System Name 3900x
Processor AMD Ryzen 3900x
Motherboard Gigabyte Aorus X470 gaming 7 WiFi
Cooling Noctua NH DI-15S soon Alphacool Eisbaer 360 with Noctua redux fans, then an open loop
Memory 32GB (4 x 8GB) Corsair 3200 C16 Dominator Platinum OC to 3733 16-16-16-32-48
Video Card(s) Gigabyte Aorus GeForce® RTX 2070 Super
Storage Samsung Evo 970 1TB NVMe M.2 SSD as boot drive, Intel 660P 2TB NVMe M.2 SSD second drive
Display(s) AOC Q3279VWFD8
Case Cooler Master H500M mesh front soon Thermaltaje View 7
Audio Device(s) Sound Blaster G6
Power Supply Seasonic Titanium 1000W with cable mods cables
Mouse Corsair Dark Core SE & Logitech G815 Keyboard
Keyboard Logitech G815
Software wIn 10 Pro possibly Manjaro at some point
Alright, so after plenty of digging through the Bios....

Cppc was disabled but cppc preferred was enabled. Imagine that. Flipping the switch brought single threaded up to 524. Another mem tweak puts my multithreaded up to 7360.
Pretty similar to mine - 514 single thread, 7410 multi

7-1-20-r20-best.jpg
 
Joined
Aug 5, 2019
Messages
235 (0.95/day)
System Name Neon Master
Processor AMD Ryzen 3900x
Motherboard x570 Aorus Master F10
Cooling H115i RGB Platinum 280mm/Case Fans NF-A14 x4
Memory 2x16GB 3733 C16 G.Skill Neo
Video Card(s) RTX 2070 Super Strix
Storage MP600 1tb+2tb / 860 Evo 1tb / WD4005FZBX
Display(s) LG 27GL850-b
Case Phanteks Evolv X
Audio Device(s) BeyerDynamic DT 770 Pro (250 Ohms)
Power Supply Seasonic Prime Ultra Titanium 1000w
Mouse Scimitar Pro rgb
Keyboard K95 Platinum rgb
Software W10Pro 1909
Looks like a case of poo 1st ccd, decent 2nd. (or.. at least peak core isnt exactly the greatest on your binned ccd)
Other way around for me, i'm getting ~530 SC, slightly sub 7400 (consistently, all stock, no pbo)
 
Joined
Dec 3, 2019
Messages
647 (5.05/day)
Location
Crawley
System Name 3900x
Processor AMD Ryzen 3900x
Motherboard Gigabyte Aorus X470 gaming 7 WiFi
Cooling Noctua NH DI-15S soon Alphacool Eisbaer 360 with Noctua redux fans, then an open loop
Memory 32GB (4 x 8GB) Corsair 3200 C16 Dominator Platinum OC to 3733 16-16-16-32-48
Video Card(s) Gigabyte Aorus GeForce® RTX 2070 Super
Storage Samsung Evo 970 1TB NVMe M.2 SSD as boot drive, Intel 660P 2TB NVMe M.2 SSD second drive
Display(s) AOC Q3279VWFD8
Case Cooler Master H500M mesh front soon Thermaltaje View 7
Audio Device(s) Sound Blaster G6
Power Supply Seasonic Titanium 1000W with cable mods cables
Mouse Corsair Dark Core SE & Logitech G815 Keyboard
Keyboard Logitech G815
Software wIn 10 Pro possibly Manjaro at some point
Looks like a case of poo 1st ccd, decent 2nd. (or.. at least peak core isnt exactly the greatest on your binned ccd)
Other way around for me, i'm getting ~530 SC, slightly sub 7400 (consistently, all stock, no pbo)
Lucky you
I'm still unsure that I have all the various BIOS settings correct

@Calmmo were you talking about mine or Zach's or moproblems ?
 
Joined
Sep 3, 2019
Messages
1,113 (5.08/day)
Location
Thessaloniki, Greece
System Name None
Processor Ryzen 5 3600 (PBO manual: PPT 105W, TDC 105A, EDC 1A, PBO Scalar auto)
Motherboard Gigabyte X570 Aorus Pro
Cooling Corsair H110i 280mm (Liquid metal for TIM)
Memory CorsairVengeanceLPX DDR4 2x8GB 3466MHz CL16-18-18-36 1T, cheap B-die, @3666 CL16-17-17-34 1T (1:1:1)
Video Card(s) MSI RX 5700XT Gaming X
Storage Samsung NVMe: 970Pro 512GB / SATA-III: 850Pro 1TB SSD, 860Evo 1TB SSD
Display(s) 24" EIZO FlexScan S2411W 1920x1200, 16:10 60Hz samsung S-PVA panel.
Case None
Audio Device(s) Sound Blaster Z <--optical link--> Logitech Z5500 5.1 500W
Power Supply Corsair HX750i
Mouse Logitech MX Master
Keyboard Logitech G15 v2
Software Windows 10 Home 64bit (v1909)
I
Looks like a case of poo 1st ccd, decent 2nd. (or.. at least peak core isnt exactly the greatest on your binned ccd)
Other way around for me, i'm getting ~530 SC, slightly sub 7400 (consistently, all stock, no pbo)
This is what better cooling can achieve...
What are your temps during R20? I bet under 70 for all core.
You could gain some more using PBO, but not just enabling it. Some tweaking must be made.
 
Joined
Mar 10, 2015
Messages
2,926 (1.58/day)
System Name Wut?
Processor 3900X
Motherboard ASRock Taichi X570
Cooling Water
Memory 32GB GSkill CL16 3600mhz
Video Card(s) Vega 56
Storage 2 x AData XPG 8200 Pro 1TB
Display(s) 3440 x 1440
Case Thermaltake Tower 900
Power Supply Seasonic Prime Ultra Platinum
1. Letest AGESA (BIOS/UEFI version) for your board.
2. Win10 v1909
3. Latest chipset drivers (November) directly from AMD (not Asrock).

UEFI settings
  • Global C-state Control = Enabled
  • Power Supply Idle Control = Low Current Idle
  • CPPC = Enabled
  • CPPC Preferred Cores = Enabled
  • AMD Cool'n'Quiet = Enabled
  • PPC Adjustment = PState 0
All of those are done now except I don't have the bottom two settings. Or they are named something else.

So, interestingly, after updating to 1.0.0.4b, my boost clocks dropped about 75-100mhz all core. My max all core is now 4075. All of my previously saved profiles were not usable anymore so I could have possibly missed a setting or something got reset that I didn't know.

I'm done for the night but I'll be back on it some tomorrow I think.
 
Joined
Dec 3, 2019
Messages
647 (5.05/day)
Location
Crawley
System Name 3900x
Processor AMD Ryzen 3900x
Motherboard Gigabyte Aorus X470 gaming 7 WiFi
Cooling Noctua NH DI-15S soon Alphacool Eisbaer 360 with Noctua redux fans, then an open loop
Memory 32GB (4 x 8GB) Corsair 3200 C16 Dominator Platinum OC to 3733 16-16-16-32-48
Video Card(s) Gigabyte Aorus GeForce® RTX 2070 Super
Storage Samsung Evo 970 1TB NVMe M.2 SSD as boot drive, Intel 660P 2TB NVMe M.2 SSD second drive
Display(s) AOC Q3279VWFD8
Case Cooler Master H500M mesh front soon Thermaltaje View 7
Audio Device(s) Sound Blaster G6
Power Supply Seasonic Titanium 1000W with cable mods cables
Mouse Corsair Dark Core SE & Logitech G815 Keyboard
Keyboard Logitech G815
Software wIn 10 Pro possibly Manjaro at some point
All of those are done now except I don't have the bottom two settings. Or they are named something else.

So, interestingly, after updating to 1.0.0.4b, my boost clocks dropped about 75-100mhz all core. My max all core is now 4075. All of my previously saved profiles were not usable anymore so I could have possibly missed a setting or something got reset that I didn't know.

I'm done for the night but I'll be back on it some tomorrow I think.
As far as I'm aware Coo'n'quiet is auto enabled if it isn't listed (it's not in my bios settings either)
 
Top