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

Trying to understand Ryzen 3000 series boost speed variations

tabascosauz

Moderator
Supporter
Staff member
Joined
Jun 24, 2015
Messages
7,678 (2.37/day)
Location
Western Canada
System Name ab┃ob
Processor 7800X3D┃5800X3D
Motherboard B650E PG-ITX┃X570 Impact
Cooling NH-U12A + T30┃AXP120-x67
Memory 64GB 6000CL30┃32GB 3600CL14
Video Card(s) RTX 4070 Ti Eagle┃RTX A2000
Storage 8TB of SSDs┃1TB SN550
Case Caselabs S3┃Lazer3D HT5
The 3600 cant match 3700X whatever you do, in any way... Its due to binning and we cant do anything about it, other than try to find best settings and get the most out of it. This is new for all of us, and I just searching the best way to do it.
Did you try to raise PBO scalar when decreasing EDC? Whats your max temp during test before and after EDC reduction?
All things matter, especially temp. The Tj of ZEN2 while may be 95C, the clock reduction is starting way below that. Also the 80C point that is rumored to be the barrier does not exist either. I saw clock raising when drop temp from 65C to 60C when using all those PBO settings.
I'm starting to believe that in order for PBO to work properly the temp must be max 50C. I'm not talking for PrecisionBoost, but PrecisionBoostOverdrive which is a separate function.

That is true, but I wouldn't be so quick to discount what's possible for six-cores, clocks wise not score wise. You do have 2 less cores, but that's not the point. I was sustaining the same multi core clocks as you not long ago, on 1.0.0.3, when PBO wouldn't give me any gains either. I still do, some freak days, without PBO. I've seen a 3600 ST result clock up 535+ in CPU-Z, and a 4.3GHz 1.36v 3700X, so the possibility is there, it's just not common.

A good example of a 3700X would probably be nojuan's chip, one that can both do 500+ ST and 5000+ MT. 3700X binning is real bad compared to the SKUs above it, which is the exact opposite of what people anticipated prior to launch.
 
Joined
Sep 3, 2019
Messages
2,998 (1.75/day)
Location
Thessaloniki, Greece
System Name PC on since Aug 2019, 1st CPU R5 3600 + ASUS ROG RX580 8GB >> MSI Gaming X RX5700XT (Jan 2020)
Processor Ryzen 9 5900X (July 2022), 150W PPT limit, 79C temp limit, CO -9~14
Motherboard Gigabyte X570 Aorus Pro (Rev1.0), BIOS F37h, AGESA V2 1.2.0.B
Cooling Arctic Liquid Freezer II 420mm Rev7 with off center mount for Ryzen, TIM: Kryonaut
Memory 2x16GB G.Skill Trident Z Neo GTZN (July 2022) 3600MHz 1.42V CL16-16-16-16-32-48 1T, tRFC:288, B-die
Video Card(s) Sapphire Nitro+ RX 7900XTX (Dec 2023) 314~465W (390W current) PowerLimit, 1060mV, Adrenalin v24.4.1
Storage Samsung NVMe: 980Pro 1TB(OS 2022), 970Pro 512GB(2019) / SATA-III: 850Pro 1TB(2015) 860Evo 1TB(2020)
Display(s) Dell Alienware AW3423DW 34" QD-OLED curved (1800R), 3440x1440 144Hz (max 175Hz) HDR1000, VRR on
Case None... naked on desk
Audio Device(s) Astro A50 headset
Power Supply Corsair HX750i, 80+ Platinum, 93% (250~700W), modular, single/dual rail (switch)
Mouse Logitech MX Master (Gen1)
Keyboard Logitech G15 (Gen2) w/ LCDSirReal applet
Software Windows 11 Home 64bit (v23H2, OSB 22631.3155)
That is true, but I wouldn't be so quick to discount what's possible for six-cores, clocks wise not score wise. You do have 2 less cores, but that's not the point. I was sustaining the same multi core clocks as you not long ago, on 1.0.0.3, when PBO wouldn't give me any gains either.
No, of course I'm not talking about cores, but clocks and boosting. When your 3700X was sustaining clocks same as my current (AGESA 1.0.0.3 ABB), mine was even further down. So the difference was the same. Later on boost was fixed by 1003 ABBA for every CPU.

Its binning...
3600 is worst that 3700X
3700X is worst than 3800X
3800X is worst than 3900X
3900X is worst than 3950X
...in terms of clocking and boosting by them selves.

AMD is using as more chiplets as it can be used and segmented them accordingly. This is a fact.
 
Joined
Dec 3, 2019
Messages
700 (0.43/day)
Location
Crawley
System Name 3900x
Processor AMD Ryzen 3900x
Motherboard Gigabyte Aorus X470 gaming 7 WiFi
Cooling Alphacool XT45 420mm Rad, 3 noctua 140 industrial 2000, EKWB EK-Quantum Kinetic TBE 200 D5 PWM
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, Intel 660P 2TB NVMe M.2 SSD , Samsung 2TB 860 QVC
Display(s) AOC Q3279VWFD8
Case Thermaltake View 71
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
Okay @Zach_01 first round of EDC changes done, you tell me what's happening :laugh:

EDC and PPT changed in BIOS (but again Only EDC actually changes, Looks to me like my PPT Is fixed?)

EDC-change-6-1-20.jpg

I haven't screenshot R20 results, but they are going up from today's baseline of 7020 with EDC at 130 it then hit 7030, with scaler x2 it was 7049. Still way below my best of 7347 though!

HW cpu 'properties', you can see that PPT wasn't raised
EDC-130.jpg

HW figures when at idle
EDC-130-1-at idle.jpg

HW during R20 run
EDC-130-2-during R20.jpg

Then applied scaler x2
EDC-change-scaler-x2-6-1-20.jpg

EDC-130-scaler-x2.jpg

HW when at idle
EDC-130-scaler-x2-idle.jpg

HW during R20
EDC-130-scaler-x2-during-R20.jpg

Hope that all makes sense - was it worth doing for me do you think?
Sorry sh1tty day here, had my neuro appt changed to a hosp further away etc etc.
 
D

Deleted member 185158

Guest
Okay @Zach_01 first round of EDC changes done, you tell me what's happening :laugh:

EDC and PPT changed in BIOS (but again Only EDC actually changes, Looks to me like my PPT Is fixed?)

View attachment 141435

I haven't screenshot R20 results, but they are going up from today's baseline of 7020 with EDC at 130 it then hit 7030, with scaler x2 it was 7049. Still way below my best of 7347 though!

HW cpu 'properties', you can see that PPT wasn't raised
View attachment 141437

HW figures when at idle
View attachment 141438

HW during R20 run
View attachment 141439

Then applied scaler x2
View attachment 141436

View attachment 141440

HW when at idle
View attachment 141441

HW during R20
View attachment 141442

Hope that all makes sense - was it worth doing for me do you think?
Sorry sh1tty day here, had my neuro appt changed to a hosp further away etc etc.

I am curios if you are running CB for a score or using it for testing stability. If the later, why such short tests? Dont think its reaching a max temp.... So assuming for the score?
Just wondering.....
 
Joined
Dec 3, 2019
Messages
700 (0.43/day)
Location
Crawley
System Name 3900x
Processor AMD Ryzen 3900x
Motherboard Gigabyte Aorus X470 gaming 7 WiFi
Cooling Alphacool XT45 420mm Rad, 3 noctua 140 industrial 2000, EKWB EK-Quantum Kinetic TBE 200 D5 PWM
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, Intel 660P 2TB NVMe M.2 SSD , Samsung 2TB 860 QVC
Display(s) AOC Q3279VWFD8
Case Thermaltake View 71
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
I am curios if you are running CB for a score or using it for testing stability. If the later, why such short tests? Dont think its reaching a max temp.... So assuming for the score?
Just wondering.....

coz @Zach_01 asked me too try it. I'm assuming curiousity to see how his idea works on a totally different CPU?


I'm doing some geekbench 5 and P95 runs as well, post those in a bit

Oh and also @tabascosauz trying your custom settings for P95 as well, not doing a thing. I'll post pics as well
 
D

Deleted member 185158

Guest
Ok Ive been trying to follow along and kinda got lost. Was just wondering thanks :)

Looks good though, you are learning tons!! Me too cause I dont have a 3000 series, they are a little different.
 
Joined
Sep 3, 2019
Messages
2,998 (1.75/day)
Location
Thessaloniki, Greece
System Name PC on since Aug 2019, 1st CPU R5 3600 + ASUS ROG RX580 8GB >> MSI Gaming X RX5700XT (Jan 2020)
Processor Ryzen 9 5900X (July 2022), 150W PPT limit, 79C temp limit, CO -9~14
Motherboard Gigabyte X570 Aorus Pro (Rev1.0), BIOS F37h, AGESA V2 1.2.0.B
Cooling Arctic Liquid Freezer II 420mm Rev7 with off center mount for Ryzen, TIM: Kryonaut
Memory 2x16GB G.Skill Trident Z Neo GTZN (July 2022) 3600MHz 1.42V CL16-16-16-16-32-48 1T, tRFC:288, B-die
Video Card(s) Sapphire Nitro+ RX 7900XTX (Dec 2023) 314~465W (390W current) PowerLimit, 1060mV, Adrenalin v24.4.1
Storage Samsung NVMe: 980Pro 1TB(OS 2022), 970Pro 512GB(2019) / SATA-III: 850Pro 1TB(2015) 860Evo 1TB(2020)
Display(s) Dell Alienware AW3423DW 34" QD-OLED curved (1800R), 3440x1440 144Hz (max 175Hz) HDR1000, VRR on
Case None... naked on desk
Audio Device(s) Astro A50 headset
Power Supply Corsair HX750i, 80+ Platinum, 93% (250~700W), modular, single/dual rail (switch)
Mouse Logitech MX Master (Gen1)
Keyboard Logitech G15 (Gen2) w/ LCDSirReal applet
Software Windows 11 Home 64bit (v23H2, OSB 22631.3155)
I am curios if you are running CB for a score or using it for testing stability. If the later, why such short tests? Dont think its reaching a max temp.... So assuming for the score?
Just wondering.....
No its not for stability test. At this point we are trying to determine CPU boost behaviour by EDC reduction. So far, by my testing and by the last screenshot I see from @lorry it seems that reacts as expected.
Current (EDC) reduction (capped) gives headroom from PPT raise. Thus meaning auto raise clocks/voltage be the FIT.

@lorry I see in your last screenshot that you are capped by PPT at 142 meaning that CPU wants to further raise clock/voltage. I dont know why its not taking values above 142. Mine easily took greater values from stock 88 PPT. I bet if you could actually set it to PPT 145~150 you would see more clocking
 
Joined
Dec 3, 2019
Messages
700 (0.43/day)
Location
Crawley
System Name 3900x
Processor AMD Ryzen 3900x
Motherboard Gigabyte Aorus X470 gaming 7 WiFi
Cooling Alphacool XT45 420mm Rad, 3 noctua 140 industrial 2000, EKWB EK-Quantum Kinetic TBE 200 D5 PWM
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, Intel 660P 2TB NVMe M.2 SSD , Samsung 2TB 860 QVC
Display(s) AOC Q3279VWFD8
Case Thermaltake View 71
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
No its not for stability test. At this point we are trying to determine CPU boost behaviour by EDC reduction. So far, by my testing and by the last screenshot I see from @lorry it seems that reacts as expected.
Current (EDC) reduction (capped) gives headroom from PPT raise. Thus meaning auto raise clocks/voltage be the FIT.

@lorry I see in your last screenshot that you are capped by PPT at 142 meaning that CPU wants to further raise clock/voltage. I dont know why its not taking values above 142. Mine easily took greater values from stock 88 PPT. I bet if you could actually set it to PPT 145~150 you would see more clocking

I cannot change PPT though, as you saw, it's changed in BIOS settings (the 145 was showing there when I went to change them this time), but as you can see from the first HW specs, it isn't being altered in practise.
Now whether that's a peculiarity of the 3900x or the x470 board I cannot determine without another MB.
 
Joined
Sep 3, 2019
Messages
2,998 (1.75/day)
Location
Thessaloniki, Greece
System Name PC on since Aug 2019, 1st CPU R5 3600 + ASUS ROG RX580 8GB >> MSI Gaming X RX5700XT (Jan 2020)
Processor Ryzen 9 5900X (July 2022), 150W PPT limit, 79C temp limit, CO -9~14
Motherboard Gigabyte X570 Aorus Pro (Rev1.0), BIOS F37h, AGESA V2 1.2.0.B
Cooling Arctic Liquid Freezer II 420mm Rev7 with off center mount for Ryzen, TIM: Kryonaut
Memory 2x16GB G.Skill Trident Z Neo GTZN (July 2022) 3600MHz 1.42V CL16-16-16-16-32-48 1T, tRFC:288, B-die
Video Card(s) Sapphire Nitro+ RX 7900XTX (Dec 2023) 314~465W (390W current) PowerLimit, 1060mV, Adrenalin v24.4.1
Storage Samsung NVMe: 980Pro 1TB(OS 2022), 970Pro 512GB(2019) / SATA-III: 850Pro 1TB(2015) 860Evo 1TB(2020)
Display(s) Dell Alienware AW3423DW 34" QD-OLED curved (1800R), 3440x1440 144Hz (max 175Hz) HDR1000, VRR on
Case None... naked on desk
Audio Device(s) Astro A50 headset
Power Supply Corsair HX750i, 80+ Platinum, 93% (250~700W), modular, single/dual rail (switch)
Mouse Logitech MX Master (Gen1)
Keyboard Logitech G15 (Gen2) w/ LCDSirReal applet
Software Windows 11 Home 64bit (v23H2, OSB 22631.3155)
You can try changing it from the other menu in AMD Overclocking.

Leave the menu of XFR enhancement like this

200106171513.png

And do the other like this...

200106171310.png
 
Joined
Dec 3, 2019
Messages
700 (0.43/day)
Location
Crawley
System Name 3900x
Processor AMD Ryzen 3900x
Motherboard Gigabyte Aorus X470 gaming 7 WiFi
Cooling Alphacool XT45 420mm Rad, 3 noctua 140 industrial 2000, EKWB EK-Quantum Kinetic TBE 200 D5 PWM
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, Intel 660P 2TB NVMe M.2 SSD , Samsung 2TB 860 QVC
Display(s) AOC Q3279VWFD8
Case Thermaltake View 71
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
Okay, geekbench 5. Interestingly this was showing better scores in general, both in single and multi core? Now showing around 1310 in single and 1330 ish in multi.
Sadly I totally forgot to get any screenshots for that though! It is a general improvement though.

P95 small, figures at the start or the run

EDC-130-scaler-x2-P95-small-start.jpg

and after 20 minutes
EDC-130-scaler-x2-P95-small-after-20mins.jpg

smallest at the start
EDC-130-scaler-x2-P95-smallest-start.jpg

after 20 minutes
EDC-130-scaler-x2-P95-smallest-after-20mins.jpg

@tabascosauz custom figures
EDC-130-scaler-x2-P95-custom.jpg

HW figures for custom at the start
EDC-130-scaler-x2-P95-custom-start.jpg

and after 20 mins - note that there was still masses of available ram!
EDC-130-scaler-x2-P95-custom-after-20mins.jpg

and largest at the start
EDC-130-scaler-x2-P95-largest-start.jpg

and after 20 mins
EDC-130-scaler-x2-P95-custom-after-20mins.jpg

hope they are useful @Zach_01

-----------------------------------------------------------------------------------------------------------------------------
Oh, and @Fry178 at general web browsing/social media etc, this rig uses around 110W and during stress testing 220W (not yet had a chance to game and see as yet).

And I verified it by testing it out on my kettle which was 3kw and my mini oven which is 1kw and both were reported with their correct figures. Which then means that the rest of my home uses about 200W on something, Hmmmm!
 
Joined
Sep 3, 2019
Messages
2,998 (1.75/day)
Location
Thessaloniki, Greece
System Name PC on since Aug 2019, 1st CPU R5 3600 + ASUS ROG RX580 8GB >> MSI Gaming X RX5700XT (Jan 2020)
Processor Ryzen 9 5900X (July 2022), 150W PPT limit, 79C temp limit, CO -9~14
Motherboard Gigabyte X570 Aorus Pro (Rev1.0), BIOS F37h, AGESA V2 1.2.0.B
Cooling Arctic Liquid Freezer II 420mm Rev7 with off center mount for Ryzen, TIM: Kryonaut
Memory 2x16GB G.Skill Trident Z Neo GTZN (July 2022) 3600MHz 1.42V CL16-16-16-16-32-48 1T, tRFC:288, B-die
Video Card(s) Sapphire Nitro+ RX 7900XTX (Dec 2023) 314~465W (390W current) PowerLimit, 1060mV, Adrenalin v24.4.1
Storage Samsung NVMe: 980Pro 1TB(OS 2022), 970Pro 512GB(2019) / SATA-III: 850Pro 1TB(2015) 860Evo 1TB(2020)
Display(s) Dell Alienware AW3423DW 34" QD-OLED curved (1800R), 3440x1440 144Hz (max 175Hz) HDR1000, VRR on
Case None... naked on desk
Audio Device(s) Astro A50 headset
Power Supply Corsair HX750i, 80+ Platinum, 93% (250~700W), modular, single/dual rail (switch)
Mouse Logitech MX Master (Gen1)
Keyboard Logitech G15 (Gen2) w/ LCDSirReal applet
Software Windows 11 Home 64bit (v23H2, OSB 22631.3155)
Well, for me to understand and to come to any conclusion I have to know your starting point, and see how the CPU runs those same tests with default (auto) PBO settings.
What I can see from those screenshots is that between "start" and "20mins later" your temps are going up, so the CPU decreases clock and voltage, while PPT is going up. This means that its trying to control the dynamics according to temp.
Is there any (easy) way to improve cooling of the system? Not only CPU but system in general. Better ventilation for example...

Temp is one of the reasons that I use R20 for the EDC testing. Because its short and the heat is not soaking.
 
Joined
Dec 3, 2019
Messages
700 (0.43/day)
Location
Crawley
System Name 3900x
Processor AMD Ryzen 3900x
Motherboard Gigabyte Aorus X470 gaming 7 WiFi
Cooling Alphacool XT45 420mm Rad, 3 noctua 140 industrial 2000, EKWB EK-Quantum Kinetic TBE 200 D5 PWM
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, Intel 660P 2TB NVMe M.2 SSD , Samsung 2TB 860 QVC
Display(s) AOC Q3279VWFD8
Case Thermaltake View 71
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
Well, for me to understand and to come to any conclusion I have to know your starting point, and see how the CPU runs those same tests with default (auto) PBO settings.
What I can see from those screenshots is that between "start" and "20mins later" your temps are going up, so the CPU decreases clock and voltage, while PPT is going up. This means that its trying to control the dynamics according to temp.
Is there any (easy) way to improve cooling of the system? Not only CPU but system in general. Better ventilation for example...

Temp is one of the reasons that I use R20 for the EDC testing. Because its short and the heat is not soaking.

About the only way would be to set the fans at 100%, which is somewhat noisy but okay just for a test time period.
I am already using the mesh front rather than the glass so the front 200mm fans get plenty of room air, could remove the side panel I guess, although wouldn't that then redirect some of the intake air to a degree?

I cannot say with Any certainty, as the few screenshots that I have aren't specific enough, but it looks as if the temp has dropped a few degrees (3 roughly) during a R20 run, Now
 
Joined
Nov 7, 2017
Messages
1,538 (0.65/day)
Location
Ibiza, Spain.
System Name Main
Processor R7 5950x
Motherboard MSI x570S Unify-X Max
Cooling D5 clone, 280 rad, two F14 + three F12S bottom/intake, two P14S + F14S (Rad) + two F14 (top)
Memory 2x8 GB Corsair Vengeance bdie 3600@CL16 1.35v
Video Card(s) GB 2080S WaterForce WB
Storage six M.2 pcie gen 4
Display(s) Sony 50X90J
Case Tt Level 20 HT
Audio Device(s) Asus Xonar AE, modded Sennheiser HD 558, Klipsch 2.1 THX
Power Supply Corsair RMx 750w
Mouse Logitech G903
Keyboard GSKILL Ripjaws
VR HMD NA
Software win 10 pro x64
Benchmark Scores TimeSpy score Fire Strike Ultra SuperPosition CB20
decent consumption for a 3900.
mine sits at 84w, similar to the 3770K+gtx1080@2ghz i had before, even with the 2080
(i like to consume least amount when its in power savings, as i dont use sleep etc).

You have a nice rig, but cant tell if u have fans on the top of it.

especially with temps being more and more the limiting factor (def on gpu),
i would recommend getting LC for either gpu or cpu (or both, long run).

Even a value AIO (cpu) dumping the heat outside the case, will lower you case/gpu temps enough so it wont throttle (starts at 50C),
or put the gpu under water, as its usually produces more heat than most cpus (not much outside gaming stresses both at the same time),
and both times you will also see about 30C less on all other temps (chip/vrm/pwm/ram/drives).
one reason i will never look at air cooling until i can have a case that houses gpu and cpu in two (physically) seperate compartments/areas
(not really possible because of MB and/or wires in the way).

just as idea. in case you come across one at a good deal and some change to waste :D
gonna link the US sites, just look who has the cheapest.

arctic is non asetek design (with pump failures) and good cooling for low price.

corsair covers ANY damaged parts if the aio fails/leaks etc.

or alphacool if u want the ability to add things/flush/refill/top off, as its more like a custom aio.
 
Joined
Dec 3, 2019
Messages
700 (0.43/day)
Location
Crawley
System Name 3900x
Processor AMD Ryzen 3900x
Motherboard Gigabyte Aorus X470 gaming 7 WiFi
Cooling Alphacool XT45 420mm Rad, 3 noctua 140 industrial 2000, EKWB EK-Quantum Kinetic TBE 200 D5 PWM
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, Intel 660P 2TB NVMe M.2 SSD , Samsung 2TB 860 QVC
Display(s) AOC Q3279VWFD8
Case Thermaltake View 71
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
decent consumption for a 3900.
mine sits at 84w, similar to the 3770K+gtx1080@2ghz i had before, even with the 2080
(i like to consume least amount when its in power savings, as i dont use sleep etc).

You have a nice rig, but cant tell if u have fans on the top of it.

especially with temps being more and more the limiting factor (def on gpu),
i would recommend getting LC for either gpu or cpu (or both, long run).

Even a value AIO (cpu) dumping the heat outside the case, will lower you case/gpu temps enough so it wont throttle (starts at 50C),
or put the gpu under water, as its usually produces more heat than most cpus (not much outside gaming stresses both at the same time),
and both times you will also see about 30C less on all other temps (chip/vrm/pwm/ram/drives).
one reason i will never look at air cooling until i can have a case that houses gpu and cpu in two (physically) seperate compartments/areas
(not really possible because of MB and/or wires in the way).

just as idea. in case you come across one at a good deal and some change to waste :D
gonna link the US sites, just look who has the cheapest.

arctic is non asetek design (with pump failures) and good cooling for low price.

corsair covers ANY damaged parts if the aio fails/leaks etc.

or alphacool if u want the ability to add things/flush/refill/top off, as its more like a custom aio.

It will go as low as 100W when it's a mind to :laugh:

I'm looking at an open loop eventually, just the fact that's going to be £400+ LoL
There are no fans on top but I can fit a 360 either top or front - I'm leaning towards the front as that has more leeway with radiator thickness but I've not as yet got down to doing any measuring with regard to pump and reservoir placement as yet, I don't see any point when I am still no way close to buying (who knows what might come out between now and then?).
If I do go the front route I may well move the front 200mm to the top (yes the case can fit them up top) pushing air downwards, it's a thought anyway.

right now my GPU is sitting at 25c and during testing it wasn't above 40c or 45c i think it was, but yes a water cooled GPU would be great I know, again something for the future.

This is my wish list for ekwb parts - included various rads, pumps, etc until I get down to measuring

water-cooled.jpg

Oh and also looking at the Noctua industrial fans, as their static pressure figures look insane, but they do ramp up to 45 decibels o_O

Noctua-industrial.jpgwater-cooled.jpg
 

Attachments

  • Noctua-industrial-1.jpg
    Noctua-industrial-1.jpg
    88.7 KB · Views: 143
Joined
Nov 7, 2017
Messages
1,538 (0.65/day)
Location
Ibiza, Spain.
System Name Main
Processor R7 5950x
Motherboard MSI x570S Unify-X Max
Cooling D5 clone, 280 rad, two F14 + three F12S bottom/intake, two P14S + F14S (Rad) + two F14 (top)
Memory 2x8 GB Corsair Vengeance bdie 3600@CL16 1.35v
Video Card(s) GB 2080S WaterForce WB
Storage six M.2 pcie gen 4
Display(s) Sony 50X90J
Case Tt Level 20 HT
Audio Device(s) Asus Xonar AE, modded Sennheiser HD 558, Klipsch 2.1 THX
Power Supply Corsair RMx 750w
Mouse Logitech G903
Keyboard GSKILL Ripjaws
VR HMD NA
Software win 10 pro x64
Benchmark Scores TimeSpy score Fire Strike Ultra SuperPosition CB20
since you dont get your rig covered when going custom, i would look into the eisbaer.
much easier to install/deal with, and has cooling that matches my custom loop (passive, with 5L coolant) that had a more powerful pump and even a better block,
even that im using throttled silent fans.
going custom would only "help" if you plan on spending money on a top block,
as ryzen is hot in small space and heat transfer is priority.

ek is good stuff, but you can get the "best" block for a bit more

i recommend not using plexi (on the pump (but acrylic or other), and dont waste money on coolant.
just get deionized water (not just distilled) , and add dead water treatment to it (5%Copper Sulfate, 2 drops per liter).

and dont waste your money on noctua/noiseblocker (like i did), as the arctic (are offering identical models for a lot less (6-8$)
they have P series (pressure, for heatsinks) and F series (airflow/case), and are pretty much silent (0.3 sone about 20 dba)
and even inaudible (0.08 sone), dont make clicking noise when throttled (4-6v depending on unit), and offer same warranty time.

fixed speed units still allow for rpm control with bios
 
Joined
Dec 3, 2019
Messages
700 (0.43/day)
Location
Crawley
System Name 3900x
Processor AMD Ryzen 3900x
Motherboard Gigabyte Aorus X470 gaming 7 WiFi
Cooling Alphacool XT45 420mm Rad, 3 noctua 140 industrial 2000, EKWB EK-Quantum Kinetic TBE 200 D5 PWM
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, Intel 660P 2TB NVMe M.2 SSD , Samsung 2TB 860 QVC
Display(s) AOC Q3279VWFD8
Case Thermaltake View 71
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
since you dont get your rig covered when going custom, i would look into the eisbaer.
much easier to install/deal with, and has cooling that matches my custom loop (passive, with 5L coolant) that had a more powerful pump and even a better block,
even that im using throttled silent fans.
going custom would only "help" if you plan on spending money on a top block,
as ryzen is hot in small space and heat transfer is priority.

ek is good stuff, but you can get the "best" block for a bit more

i recommend not using plexi (on the pump (but acrylic or other), and dont waste money on coolant.
just get deionized water (not just distilled) , and add dead water treatment to it (5%Copper Sulfate, 2 drops per liter).

and dont waste your money on noctua/noiseblocker (like i did), as the arctic (are offering identical models for a lot less (6-8$)
they have P series (pressure, for heatsinks) and F series (airflow/case), and are pretty much silent (0.3 sone about 20 dba)
and even inaudible (0.08 sone), dont make clicking noise when throttled (4-6v depending on unit), and offer same warranty time.

fixed speed units still allow for rpm control with bios

I didn't get what you meant by "since you dont get your rig covered when going custom" ? If you mean leaks I'm practised enough on rebuilding enough bikes and car engines and rads etc. i was also under the impression that open loop beats AIO in the majority of cases? (pun intended).
I want something that will last And that can be added to down the road (GPU). i thought the idea was to use deionized (I used to make gallons of pure glass distilled H2) at the labs decades ago that was deionized as well. As for biocide additive i thought there were plenty of variations that you simply mixed 1:10 ?

The artic P fans have a static pressure of 2.4 mm H2O (@ 1 700 RPM), the Noctua has 7.63 mm H₂O, I'm assuming that higher is better surely?

And if not ekwg which others then re rads, pumps reservoirs etc?

This one I know is good

 
Joined
Jun 3, 2010
Messages
2,540 (0.50/day)
Did you only change EDC? Something surely hit the fan, because you're back at JEDEC on RAM, RAM is uncoupled, Vcore goes up to 1.9V which is clearly a bug, and all your TDC/EDC/PPT values appear to be exactly stock.

Clear your CMOS and restore your previous settings, if you've saved them as a BIOS profile?

-----

@mtcn77 see these results at 83A EDC and no offset. I put LLC back on Auto for more droop at load (I had a nasty habit of leaving it on Turbo, a remnant of when I was testing fixed freq).

View attachment 141235
tabascosauz, I've been off the grid, though very appreciative and collaborative of your efforts.
What I have come across by previous generations of AMD's current Ryzen Master FIT controls is, the power monitor cuts across at every attempt its algorithm meets its requirements that suggest current temperature slope will cross-sect with Tj Max. It is most shorthand of all these discussions to just watch the current TDP use and monitor that FIT does not try to curb its TDP budget. I have found by looking at near and above operating heat threshold, we can profile the whole temperature behaviour.
I would suggest you disable LLC and let voltage droop. LLC stabilizes performance, but that is not why EDC is concurrently important. We are trying to limit off-current voltage spikes that needlessly heat up the cpu, so that FIT doesn't detect such a discrepancy. EDC has the possibility to reflect that in the cpu frequency profiles. If you reduce EDC, the cpu successfully profiles itself for a vdroop event. Therefore, EDC should best be used singlehandedly to compensate vrm phase stability. If we use it to complement with LLC, we'd constantly be overshooting TDP current targets which would reflect as a lower available overclocked TDP.

You could also take starting voltage as 'LLC' starting point and by changing cpu voltage offset, you could differ the stable overclock target.
You haven't specified your undervolt targets to the best of my knowledge, but if you expect EDC to have such a vdroop that it runs like an undervolted setting, either the motherboard phases need be weak, or current should run so high.
 
Last edited:

tabascosauz

Moderator
Supporter
Staff member
Joined
Jun 24, 2015
Messages
7,678 (2.37/day)
Location
Western Canada
System Name ab┃ob
Processor 7800X3D┃5800X3D
Motherboard B650E PG-ITX┃X570 Impact
Cooling NH-U12A + T30┃AXP120-x67
Memory 64GB 6000CL30┃32GB 3600CL14
Video Card(s) RTX 4070 Ti Eagle┃RTX A2000
Storage 8TB of SSDs┃1TB SN550
Case Caselabs S3┃Lazer3D HT5
@lorry yeah I've been looking everywhere and can't seem to find any results for anyone experiencing a similar bug. Sorry mate. I don't know if there's anything else that you can change in the P95 settings. Something tells me it's either the way P95 optimizes for the 3900X/3950X, or something to do with how current AMD firmware interacts with P95? Either way, if you've been playing demanding games, running memtest86 and there haven't been any stability issues, probably just call it a day.

I will say, though, if you aren't in a terribly urgent rush to move to a custom loop, there are better options than the NF-F12. The regular PWM is, although powerful, quite loud at the upper end when up against a restrictive surface. The iPPC F12s are even noisier. Between the F12 PWM and the A12x25 PWM, the newer fan looks weaker on paper but delivers better airflow across the RPM band, and is noticeably "quieter"; compared to the F12, there's less motor hum/whine even at max speed.

I've been waiting for Noctua to release iPPC and Chromax based on the A12x25, but it appears that they want to keep the F12 lineage around for a while longer.

Something else to keep in mind is that the regular version of the A12x25 (and I'm guessing the regular F12 as well, I bought my F12 waaayyyy back when the gasket didn't exist) come with a full coverage square gasket to seal the edges of the fan to the radiator in the event of use for watercooling, a very important step to maintain static pressure. To the best of my knowledge, most if not all the iPPCs come in a barebones OEM-esque box with just the fan. No LNA/ULNA adapters, no splitters, no rubber mounts and no gaskets.

And I'm no expert on other fan brands; I'm sure there are other competitors out there worth a look.
 
Joined
Dec 3, 2019
Messages
700 (0.43/day)
Location
Crawley
System Name 3900x
Processor AMD Ryzen 3900x
Motherboard Gigabyte Aorus X470 gaming 7 WiFi
Cooling Alphacool XT45 420mm Rad, 3 noctua 140 industrial 2000, EKWB EK-Quantum Kinetic TBE 200 D5 PWM
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, Intel 660P 2TB NVMe M.2 SSD , Samsung 2TB 860 QVC
Display(s) AOC Q3279VWFD8
Case Thermaltake View 71
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
@lorry yeah I've been looking everywhere and can't seem to find any results for anyone experiencing a similar bug. Sorry mate. I don't know if there's anything else that you can change in the P95 settings. Something tells me it's either the way P95 optimizes for the 3900X/3950X, or something to do with how current AMD firmware interacts with P95? Either way, if you've been playing demanding games, running memtest86 and there haven't been any stability issues, probably just call it a day.

I will say, though, if you aren't in a terribly urgent rush to move to a custom loop, there are better options than the NF-F12. The regular PWM is, although powerful, quite loud at the upper end when up against a restrictive surface. The iPPC F12s are even noisier. Between the F12 PWM and the A12x25 PWM, the newer fan looks weaker on paper but delivers better airflow across the RPM band, and is noticeably "quieter"; compared to the F12, there's less motor hum/whine even at max speed.

I've been waiting for Noctua to release iPPC and Chromax based on the A12x25, but it appears that they want to keep the F12 lineage around for a while longer.

Something else to keep in mind is that the regular version of the A12x25 (and I'm guessing the regular F12 as well, I bought my F12 waaayyyy back when the gasket didn't exist) come with a full coverage square gasket to seal the edges of the fan to the radiator in the event of use for watercooling, a very important step to maintain static pressure. To the best of my knowledge, most if not all the iPPCs come in a barebones OEM-esque box with just the fan. No LNA/ULNA adapters, no splitters, no rubber mounts and no gaskets.

And I'm no expert on other fan brands; I'm sure there are other competitors out there worth a look.

I think that it is something to do with P95 itself and how it is coded, as Intel Burn will take me right down to 0MB of available ram left. Either way, it seems perfectly stable enough and nothing lost in the end.

It's not that I am in a rush, it's money LoL. It's just going to take me a while to get enough together. What I will likely do is how I bought this rig originally - decide on the parts and then buy them separately. That will cost me slightly more I know (extra delivery costs) but I am then committed to buying it, so not tempted to spend on something else until it is all together. Also it allows me more time to possibly find a better/more suitable part.

I bought the Noctua NH DI-15S and then swapped out its fan for the Noctua NF-A15 HS-PWM chromax.black.swap & a Noctua NF-F12 PWM chromax.black.swap

if you haven't checked lately, this is the range of chromax fans Now - their chromax range has been expanded and includes rubber mounts, splitters etc


Is this the one you are talking about?

 

tabascosauz

Moderator
Supporter
Staff member
Joined
Jun 24, 2015
Messages
7,678 (2.37/day)
Location
Western Canada
System Name ab┃ob
Processor 7800X3D┃5800X3D
Motherboard B650E PG-ITX┃X570 Impact
Cooling NH-U12A + T30┃AXP120-x67
Memory 64GB 6000CL30┃32GB 3600CL14
Video Card(s) RTX 4070 Ti Eagle┃RTX A2000
Storage 8TB of SSDs┃1TB SN550
Case Caselabs S3┃Lazer3D HT5
I think that it is something to do with P95 itself and how it is coded, as Intel Burn will take me right down to 0MB of available ram left. Either way, it seems perfectly stable enough and nothing lost in the end.

It's not that I am in a rush, it's money LoL. It's just going to take me a while to get enough together. What I will likely do is how I bought this rig originally - decide on the parts and then buy them separately. That will cost me slightly more I know (extra delivery costs) but I am then committed to buying it, so not tempted to spend on something else until it is all together. Also it allows me more time to possibly find a better/more suitable part.

I bought the Noctua NH DI-15S and then swapped out its fan for the Noctua NF-A15 HS-PWM chromax.black.swap & a Noctua NF-F12 PWM chromax.black.swap

if you haven't checked lately, this is the range of chromax fans Now - their chromax range has been expanded and includes rubber mounts, splitters etc


Is this the one you are talking about?


Good things come to those who wait :)

I saw you linked a screenshot of the NF-F12 iPPC-3000, that's why I thought you were looking at the iPPCs. The iPPCs are tougher and louder fans, some are water-resistant, and all of them have frames made from a different fibreglass material than the regular line (a total PITA to use with self-tapping case fan screws). The Chromax fans are just black versions of the regular fans, and have swappable coloured corner pieces. Only reason Chromax exists is because most people apparently hate Noctua brown; they're just coloured differently. Same goes for the Chromax versions of the U12S and L9i coolers; if the U12A wasn't so damn expensive, it'd be really hard to recommend the U12S Chromax.

No, the A12x15 is also a new fan but it's a thinner 15mm fan. Performance I'd expect would be closer to the S12A than the A12x25 and F12. All the other 120mms like the A12x25, F12, S12 and P12 are all standard 25mm thickness. Only reason to use the A12x15 is if you only have between 15mm-25mm of clearance.

The A12x25 had a Chromax prototype shown last year, but nothing yet.
 
Joined
Dec 3, 2019
Messages
700 (0.43/day)
Location
Crawley
System Name 3900x
Processor AMD Ryzen 3900x
Motherboard Gigabyte Aorus X470 gaming 7 WiFi
Cooling Alphacool XT45 420mm Rad, 3 noctua 140 industrial 2000, EKWB EK-Quantum Kinetic TBE 200 D5 PWM
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, Intel 660P 2TB NVMe M.2 SSD , Samsung 2TB 860 QVC
Display(s) AOC Q3279VWFD8
Case Thermaltake View 71
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
Good things come to those who wait :)

I saw you linked a screenshot of the NF-F12 iPPC-3000, that's why I thought you were looking at the iPPCs. The iPPCs are tougher and louder fans, some are water-resistant, and all of them have frames made from a different fibreglass material than the regular line (a total PITA to use with self-tapping case fan screws). The Chromax fans are just black versions of the regular fans, and have swappable coloured corner pieces. Only reason Chromax exists is because most people apparently hate Noctua brown; they're just coloured differently. Same goes for the Chromax versions of the U12S and L9i coolers; if the U12A wasn't so damn expensive, it'd be really hard to recommend the U12S Chromax.

No, the A12x15 is also a new fan but it's a thinner 15mm fan. Performance I'd expect would be closer to the S12A than the A12x25 and F12. All the other 120mms like the A12x25, F12, S12 and P12 are all standard 25mm thickness. Only reason to use the A12x15 is if you only have between 15mm-25mm of clearance.

The A12x25 had a Chromax prototype shown last year, but nothing yet.


erm, this?
Size 120x120x15 mm

 

tabascosauz

Moderator
Supporter
Staff member
Joined
Jun 24, 2015
Messages
7,678 (2.37/day)
Location
Western Canada
System Name ab┃ob
Processor 7800X3D┃5800X3D
Motherboard B650E PG-ITX┃X570 Impact
Cooling NH-U12A + T30┃AXP120-x67
Memory 64GB 6000CL30┃32GB 3600CL14
Video Card(s) RTX 4070 Ti Eagle┃RTX A2000
Storage 8TB of SSDs┃1TB SN550
Case Caselabs S3┃Lazer3D HT5
erm, this?
Size 120x120x15 mm


Yes. This is the A12x15. Chromax version available.
noctua_nf_a12_15_1_4.jpgnf_a12_15_chromax_black_1_1.jpg

This is the A12x25. Chromax and iPPC not available.
noctua_nf_a12x25_1_8.jpg

Most fans on the consumer market are 25mm, which makes the A12x15 a bit of a niche product.
 
Joined
Dec 3, 2019
Messages
700 (0.43/day)
Location
Crawley
System Name 3900x
Processor AMD Ryzen 3900x
Motherboard Gigabyte Aorus X470 gaming 7 WiFi
Cooling Alphacool XT45 420mm Rad, 3 noctua 140 industrial 2000, EKWB EK-Quantum Kinetic TBE 200 D5 PWM
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, Intel 660P 2TB NVMe M.2 SSD , Samsung 2TB 860 QVC
Display(s) AOC Q3279VWFD8
Case Thermaltake View 71
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
--------------------------------------------------------------------------------------------------------
and The NF-A12x25 does come with a gasket
Anti-vibration gasket for watercooling radiators


Anti-vibration gasket for watercooling radiators
The NF-A12x25 includes a silicone gasket that can be used when installing the fan on watercooling radiators or other applications that benefit from a tighter seal with the fan. The gasket helps to damp minute vibrations and to reduce air loss between the fan and the radiator for optimal cooling performance.
 
Joined
Nov 7, 2017
Messages
1,538 (0.65/day)
Location
Ibiza, Spain.
System Name Main
Processor R7 5950x
Motherboard MSI x570S Unify-X Max
Cooling D5 clone, 280 rad, two F14 + three F12S bottom/intake, two P14S + F14S (Rad) + two F14 (top)
Memory 2x8 GB Corsair Vengeance bdie 3600@CL16 1.35v
Video Card(s) GB 2080S WaterForce WB
Storage six M.2 pcie gen 4
Display(s) Sony 50X90J
Case Tt Level 20 HT
Audio Device(s) Asus Xonar AE, modded Sennheiser HD 558, Klipsch 2.1 THX
Power Supply Corsair RMx 750w
Mouse Logitech G903
Keyboard GSKILL Ripjaws
VR HMD NA
Software win 10 pro x64
Benchmark Scores TimeSpy score Fire Strike Ultra SuperPosition CB20
corsair offers 5y warranty incl replacing parts that die if it goes bad,
e.g. you get the pc replaced if it explodes tmrw.

when you get anything else, you pay for it, so there would be no difference if custom or different aio.
yes, de-ionized water (type I) is the cleanest, and you want one drop of 5% biozide solution per liter of fluid.
i used it in a loop with a "dirty" (slimy residue/fishy smell) resorator to test it, just flushed solids out,
then filled with (only) distilled water and added the biozide.
after 1y nothing in the loop whatsoever, so with deionized water, its gonna be fine for a while.


I just realized that the cost (money/time) nowadays the decent aio dont get worse perf than a custom loop that cost more.
unless utter silence on pump is needed (which really isnt the case, there are just better solutions to decouple them from case, pump still making noise),
because they are separate (from the block), i dont see any difference between my eisbaer and a custom loop that (can) cost almost 3 times as much.

the arctic is a good unit if its just to use water/better than asetek design (high pump failures, very low flow),
but the additional money on something like the eisbaer is worth it if you want options for later (another rad/block/pump).
e.g:
the 2080 i have comes with an (asetek) cooler that is identical (block/pump) with those corsair and the like (CPU aio),
its not just loud, but annoying as hell (pump has high freq whine), especially since no other noise.
so ill swap the card for one with water block, add another 120 rad, and a D5 pump/res later.
since all are "reg" LC parts, i have the same option as on a custom loop, incl no need to tear down everything when expanding.


yes, they have less pressure, but they do that at less than half dba (and beats the NF P12 as cooler fan)
and i cant find any noctua/noiseblocker that can match an arctic below 12 dba.

i prefer to have less noisy fans running at almost full speed, less ramping up/down with cpu load shifting a lot,
and with gpu and cpu both under water, temps for each are about 30-50*C less (depending on load),
which means there is no need for high rpm/air flow in the first place.
but sure, going up if you can stand the louder fans is fine, but i can even do push&pull (2x fans)
for less than the other brands doing only one.

at least for me, with 6-8$/per arctic fan, i just dont see it making sense to spend 2-3 times for "10%" gains (for silent fans).

maybe look at the bionics, i like them as they are powerful but silent, and the blade material doesnt attract ANY dust.
ran two for about 1y, looked like new, when i had to dust others after 2 month).


im a silent "whore" and haven't used a seal/gasket in the past 5y on any fan below 20 dba.
proper cases usually dont need it (less resonating than those cheap flimsy cases).

edited for typo and adding info
 
Last edited:
Joined
Dec 3, 2019
Messages
700 (0.43/day)
Location
Crawley
System Name 3900x
Processor AMD Ryzen 3900x
Motherboard Gigabyte Aorus X470 gaming 7 WiFi
Cooling Alphacool XT45 420mm Rad, 3 noctua 140 industrial 2000, EKWB EK-Quantum Kinetic TBE 200 D5 PWM
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, Intel 660P 2TB NVMe M.2 SSD , Samsung 2TB 860 QVC
Display(s) AOC Q3279VWFD8
Case Thermaltake View 71
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
@Fry178 you lost me there somewhat mate, which are you advising / suggesting please?

@tabascosauz remember I'm still new to all of this so i looked at what might give the best performance and the ippc fans seemed to have that, only reason i looked at them.
Still very unsure what to go for TBH, an AIO or open loop, if open - hardware labs for the radiator ?
 
Top