Follow along with the video below to see how to install our site as a web app on your home screen.
Note: This feature may not be available in some browsers.
Competitor rules
Please remember that any mention of competitors, hinting at competitors or offering to provide details of competitors will result in an account suspension. The full rules can be found under the 'Terms and Rules' link in the bottom right corner of your screen. Just don't mention competitors in any way, shape or form and you'll be OK.
I generally don't rate most reviews of thermal paste, though to be fair it is complex and time consuming to do properly - many persist with using the same application method with all the pastes out of "fairness" despite there usually is a variation of what method is most optimal for any given paste, some pastes will give optimal results almost instantly, some will take far longer than a 30 minute burn-in period to reach their best performance.
I generally take exception to people hand spreading AS5 as you'd almost always get better results with a paste designed primarily for hand-spreading if doing that as AS5 is designed primarily to get best results itself under contact pressure and thermal cycles, though on today's larger dies it can struggle without hand-spreading but again you'd be better going to another paste - one of the reasons I've largely switched to MX-6 on modern CPUs.
Please refer to the table below to check which offset bars are required for which cooler model. Regular and chromax.black models use the same mounting hardware, so both regular and chromax.black offset mounting bars can be used depending on visua...
yea correct however what im saying is theres 2 verisons of AMB-12 1 for horizontal mounting and 1 for vertical and they both have the same product id "AMB-12".
I generally don't rate most reviews of thermal paste, though to be fair it is complex and time consuming to do properly - many persist with using the same application method with all the pastes out of "fairness" despite there usually is a variation of what method is most optimal for any given paste, some pastes will give optimal results almost instantly, some will take far longer than a 30 minute burn-in period to reach their best performance.
You have full right to your opinion and I am not out to change it, nor has mine changed from your post . I saw discussion going on, and I felt I'd add my take also . I was never calling out anyone or saying their methods are worse and mine better. That's why my post had no quote of another/mention of another forum members name.
Besides the mega review by THG, the other mega reviews that I've seen are similar outcomes. Der8auer had one many years ago, I can't find the site now. So I think these are indicative of performance you may get out of TIM.
I do a lot of hours of testing and thermal cycles. I'll be honest, yet to notice a vast difference between results at application and a week or few weeks later, when my equipment has reached ~100hrs+ usage time. An example of how quickly I may reach hours of usage is in this post (there where shutdowns of rig, so some thermal cycling).
I generally take exception to people hand spreading AS5 as you'd almost always get better results with a paste designed primarily for hand-spreading if doing that as AS5 is designed primarily to get best results itself under contact pressure and thermal cycles, though on today's larger dies it can struggle without hand-spreading but again you'd be better going to another paste - one of the reasons I've largely switched to MX-6 on modern CPUs.
I have compared hand spread vs just doing a cross, etc. The hand spread regardless of paste used ends up better. In that most of the spread has occurred already, so from get go I may (note I say may, as I found it hard to quantify) get better temperature, less chance of over spill, some mild does still occur. I'm not wondering did I use excessive or too little TIM. Contact pressure spreading is still going to occur on the hand spread, it's unavoidable. I will never by hand get the right level of TIM as the contact pressure results in.
Stock vs CO with RAM OC. CO+RAM OC the CPU is consuming more power vs just stock, but the CO is giving it a mild edge on power side. I have also wall plug meter reads for system and again very close between each setup.
Generally these optimizations dont do a ton, but we're a bit better in most use cases. The trade off of this optimizations was that it made reported latency from benchmarks like AIDA 64 higher.
This is because all modern processors use advanced prefetchers and cache retention policies that cloud dram latency measures. To accurately measure dram latency these features need to be accounted for (with the easiest way disabling them).
It's not fake result in say AIDA64, it just as AMD Employee explained. So far running Legacy (ie advanced prefetchers and cache retention polices disabled), out benchmarks I ran,only effects AIDA64 memory benchmark ns, not gamed on AM5 setup yet.
*** edit ***
I also think AMD guided ODMs to implement this feature as AIDA64 is quite popular with us normal users and some reviewers. And people may say "Oh look 9000 series is step backwards...", because AFAIK it's a 9000 series feature. IIRC from when I looked at old 7000 series AIDA64 benchmarks, they had lower ns then 9000 series.
It's not fake result in say AIDA64, it just as AMD Employee explained. So far running Legacy (ie advanced prefetchers and cache retention polices disabled), out benchmarks I ran,only effects AIDA64 memory benchmark ns, not gamed on AM5 setup yet.
*** edit ***
I also think AMD guided ODMs to implement this feature as AIDA64 is quite popular with us normal users and some reviewers. And people may say "Oh look 9000 series is step backwards...", because AFAIK it's a 9000 series feature. IIRC from when I looked at old 7000 series AIDA64 benchmarks, they had lower ns then 9000 series.
I have only owned ASUS since 2007. I purchased Crosshair VI Hero WiFi, Crosshair VIII Dark Hero, TUF Gaming X670E-Plus WiFi, Crosshair X670E Hero.
ASUS gave me Crosshair VII Hero WiFi, Zenith Extreme and Zenith Extreme Alpha.
I have owned/experienced Ryzen 1000/2000/3000/5000/9000 and Threadripper 1000/2000. I'm an avid user of Beta UEFI also .
I always do a POST on defaults, get screen shots and data for various aspects that interest me. I use now ASUS TurboV Core, Ryzen Master, ZenTimings; Nitro setup is seen via Tools > System Info, to see tPHYRDL sync, switch between channel A/B on main page.
For ASUS TurboV Core the version in OP has given me more values then newer ones even included with USB in X670E Hero or X870E Hero I got from a OCN forum member. Just be aware you can change voltages "on the fly" with that app.
I'll set default voltages manually based on what I see in those applications, as I don't like voltages increased unnecessarily by auto rules.
1.8V PLL Voltage [1.80000]
1.8V Standby Voltage [1.80000]
Misc_ALW [0.75000]
Chipset0 VDD Voltage [1.05000]
Chipset1 VDD Voltage [1.05000]
CPU 3.3V [3.30000]
CPU SOC Voltage [Manual Mode]
- VDDSOC Voltage Override [1.02500]
CPU VDDIO / MC Voltage [Manual Mode]
- VDDIO Override [1.09980]
Misc Voltage [Manual Mode]
- Misc Voltage Override [1.10000]
VDDP Voltage [0.800]
High DRAM Voltage Mode [Disabled]
DRAM VDD Voltage [1.10000]
DRAM VDDQ Voltage [1.10000]
VDDG CCD Voltage [900 mV]
VDDG IOD Voltage [900 mV]
PMIC Force Continuous Current Mode [Auto]
PMIC Voltages [Sync All PMICs]
SPD HUB VLDO (1.8V) [1.80000]
SPD HUB VDDIO (1.0V) [1.00000]
Memory VDD Voltage [1.10000]
Memory VDDQ Voltage [1.10000]
Memory VPP Voltage [1.80000]
On AM5 we can see AMD auto rules, if you have EXPO RAM. These rules may change based on CPU/AGESA in use.
These AMD auto rules can be seen if you leave motherboard maker menus on Auto and load EXPO from AMD CBS/OC menus.
AMD help strings from UEFI AMD Overclocking menu, as this section is AMD FW area, it's AMD guidance.
I like to set Ai Overclock Tuner to manual and 100MHz BCLK.
Extreme Tweaker > Tweaker's Paradise > Clock Spread Spectrum [Disabled], as otherwise board may adjust BCLK to comply with EMI regulations. Even through frequency isn't absolute, it stops the mild swing which may destabilize OC. As FCLK/MEMCLK and other buses base their frequency on BCLK it's also good to stop this variation. Extreme Tweaker > DIGI+ VRM > VRM Spread Spectrum [Disabled].
Advanced > CPU Configuration > SVM Mode [Disabled] if you don't use VM, as some things benchmark better with it off and in the past read some say had smoother experience in games.
I also set on Tools > Download & Install ARMOURY CRATE app [Disabled], I have no use for these apps.
That's about it really . Rest is personal preference/needs.
As I do PBO CO OC I don't have a need to adjust LLC. RAM OC is again based on RAM IC/Kit in use and what the CPU IMC allows.
In a past post you mentioned a dual CCD CPU, AFAIK Medium Load Boostit (link) set to [Enabled] can give gains. Skatterbencher has some good guides on 9000 series and also the 7000 series is good to reference.
I have only owned ASUS since 2007. I purchased Crosshair VI Hero WiFi, Crosshair VIII Dark Hero, TUF Gaming X670E-Plus WiFi, Crosshair X670E Hero.
ASUS gave me Crosshair VII Hero WiFi, Zenith Extreme and Zenith Extreme Alpha.
I have owned/experienced Ryzen 1000/2000/3000/5000/9000 and Threadripper 1000/2000. I'm an avid user of Beta UEFI also .
I always do a POST on defaults, get screen shots and data for various aspects that interest me. I use now ASUS TurboV Core, Ryzen Master, ZenTimings; Nitro setup is seen via Tools > System Info, for tPHYRDL sync switch between channel A/B on main page.
For ASUS TurboV Core the version in OP has given me more values then newer ones even included with USB in X670E Hero or X870E Hero I got from a OCN forum member. Just be aware you can change voltages "on the fly" with that app.
I'll set default voltages manually based on what I see in those applications, as I don't like voltages increased unnecessarily by auto rules.
1.8V PLL Voltage [1.80000]
1.8V Standby Voltage [1.80000]
Misc_ALW [0.75000]
Chipset0 VDD Voltage [1.05000]
Chipset1 VDD Voltage [1.05000]
CPU 3.3V [3.30000]
CPU SOC Voltage [Manual Mode]
- VDDSOC Voltage Override [1.02500]
CPU VDDIO / MC Voltage [Manual Mode]
- VDDIO Override [1.09980]
Misc Voltage [Manual Mode]
- Misc Voltage Override [1.10000]
VDDP Voltage [0.800]
High DRAM Voltage Mode [Disabled]
DRAM VDD Voltage [1.10000]
DRAM VDDQ Voltage [1.10000]
VDDG CCD Voltage [900 mV]
VDDG IOD Voltage [900 mV]
PMIC Force Continuous Current Mode [Auto]
PMIC Voltages [Sync All PMICs]
SPD HUB VLDO (1.8V) [1.80000]
SPD HUB VDDIO (1.0V) [1.00000]
Memory VDD Voltage [1.10000]
Memory VDDQ Voltage [1.10000]
Memory VPP Voltage [1.80000]
On AM5 we can see AMD auto rules, if you have EXPO RAM. These rules may change based on CPU/AGESA in use.
These AMD auto rules can be seen if you leave motherboard maker menus on Auto and load EXPO from AMD CBS/OC menus.
AMD help strings from UEFI AMD Overclocking menu, as this section is AMD FW area, it's AMD guidance.
I like to set Ai Overclock Tuner to manual and 100MHz BCLK.
Extreme Tweaker > Tweaker's Paradise > Clock Spread Spectrum [Disabled], as otherwise board may adjust BCLK to comply with EMI regulations. Even through frequency isn't absolute, it stops the mild swing which may destabilize OC. As FCLK/MEMCLK and other buses base their frequency on BCLK it's also good to stop this variation. Extreme Tweaker > DIGI+ VRM > VRM Spread Spectrum [Disabled].
Advanced > CPU Configuration > SVM Mode [Disabled] if you don't use VM, as some things benchmark better with it off and in the past read some say had smoother experience in games.
I also set on Tools > Download & Install ARMOURY CRATE app [Disabled], I have no use for these apps.
That's about it really . Rest is personal preference/needs. As I do PBO CO OC I don't have a need to adjust LLC. RAM OC is again based on RAM IC/Kit in use and what the CPU IMC allows.
In a past post you mentioned a dual CCD CPU, AFAIK Medium Load Boostit (link) set to [Enabled] can give gains.
I’ve read through some of your CO info you posted previously and I could have saved myself quite a bit of time….probably weeks in fact trying different methods of trying to tame the 7950x3d pretty much blind. Hopefully with what I now learned from what must have took you quite some time, I will have an easier time when I get my 9950x3d.
The silicon gods did answer my prayers . My first roll of the silicon lottery die on R7 9800X3D was fruitful .
Has nice PBO CO OC.
FCLK 2233MHz is stable in a lot of things, excluding some heavy CPU only oriented stability tests.
Has a decent IMC IMO, did 6000C28 1:1 similar SOC/VDDP as 2x other 9000 series I have which do 6400C28 1:1. 6200C28 1:1 has been fine on this 9800X3D, will be soon be trying 6400C28.
Comparing power efficiency to the R7 9700X I have, the R7 9800X3D is the better choice, this is disregarding price difference between each.
You are welcome , enjoy your current and future HW .
The silicon gods did answer my prayers . My first roll of the silicon lottery die on R7 9800X3D was fruitful .
Has nice PBO CO OC.
FCLK 2233MHz is stable in a lot of things, excluding some heavy CPU only oriented stability tests.
Has a decent IMC IMO, did 6000C28 1:1 similar SOC/VDDP as 2x other 9000 series I have which do 6400C28 1:1. 6200C28 1:1 has been fine on this 9800X3D, will be soon be trying 6400C28.
Comparing power efficiency to the R7 9700X I have, the R7 9800X3D is the better choice, this is disregarding price difference between each.
6400 C28 seems to be what most are aiming for. I’d be over the moon with that. Think @LtMatt has achieved 6400 C26 2233MHz, think he is using a GENE though if I remember right.
This site uses cookies to help personalise content, tailor your experience and to keep you logged in if you register.
By continuing to use this site, you are consenting to our use of cookies.
. I saw discussion going on, and I felt I'd add my take also 
.