First AMD build in 20 years!

I'm currently using -15 on my two fastest cores and -5 on the rest. Based on your post you are recommending a lower negative offset for the best cores vs higher?
To hit higher boost clocks the better cores need more voltage. Having a larger negative offset will mean they get fed less voltage and will be less likely to hit the higher speeds during light load scenarios.
 
If your on a 5950x, your not cooling that with anything less than a H2O AIO.

I think all AMD boards over-volt from what I have been reading. I have manually set almost all of my voltages now. My mobo was putting 1.2V on the SOC (NB/Chipset)! Way above normal, so I set that to 1.05V along with a bunch of other voltages that were quite a bit higher than stock.

I recommend starting maybe with my settings (or close) as a quick OC. The lower current limits seem to have helped a lot with maintaining boosts and keep temperatures in check.

Pretty happy with my CPUz scores for now :)

View attachment 348716
I have to disagree on one point - lots of people are using the NH-D15 and NH-D15s and getting good temps with overclocks. Keep in mind, that Noctua is generally regarded as the best air cooler on the market. I did consider a 360 or even 420 AIO and ultimately, will still go that route if I must. I already had the NH-D15s so based on the reading I did, I decided to give it a shot :)

I had an Arctic Freezer 240 on my 8700k and while it performed spectacularly, it died a month after the warranty expired and the system seemed to have intermittent stability issues after that. That kind of left a bad taste in my mouth about AIOs. I know the Arctic Freezer II now has a six year warranty and is near the top of the leaderboards in performance, so it’s something I can consider. Plus I have Amazon Reward points which would effectively make it “free” if I buy it.

I’ll use some of your settings as a starting point and see what happens. If I reapply the TIM next weekend and don’t see better results, I’ll probably go ahead with the AIO.
 
To hit higher boost clocks the better cores need more voltage. Having a larger negative offset will mean they get fed less voltage and will be less likely to hit the higher speeds during light load scenarios.

I will have to recheck my settings then because right now i'm doing 5.05Ghz single core.

cpuz max boost.png
 
If your on a 5950x, your not cooling that with anything less than a H2O AIO.

I think all AMD boards over-volt from what I have been reading. I have manually set almost all of my voltages now. My mobo was putting 1.2V on the SOC (NB/Chipset)! Way above normal, so I set that to 1.05V along with a bunch of other voltages that were quite a bit higher than stock.
SOC voltage of 1.2 seems excessive for auto. On my 5900x it is set to 1.1v on auto and hwinfo shows between 1.071 and 1.085 with dual rank 3600mhz CL16 memory. TRC is a value where the chipset regularly ignores the XMP values and can give some extra performance.

Using an air cooler is fine on a 5800x without PBO and the 5800x is much more difficult to cool than the 5900x and 5950x due to having only one CCD. For PBO the 5900x and 5950x will be fine on the best air coolers, but the 5800x is probably borderline. Temps are mostly dependent on how many watts you are pushing into a few cores on a single CCD and if the core is doing AVX or similar, total number of watts is more secondary on the large air towers, AIOs and custom loops. E.g. run 120 watts into 6 cores with AVX or similar will be much hotter than running 32 thread cinebench. Try prime95 small fft 6-7 threads with PBO on (5 threads with PBO off) and you will get an idea of your max temps. Doing undervolting based on curves does lower temps quite a bit for the same mhz, but the voltage each core can run at is very CPU sample dependant. Regular PBO runs with something that should be OK for even the worst samples so it produces more heat compared to fine tuning.

AIOs will be slightly better on manual fan curves or when running one off short tests, but will not be massively better than air, maybe 5-7 degrees for a 360AIO vs the D15 when both are at full fan speed after both reach steady state. Put the AIO on an auto profile and then run 20 minutes of lopped cinebench, OCCT stress test or similar and then measure for 2 minutes and the results for the AIO will be quite different than when doing a single cinebench run as there is around a 6 minute soak time for 360 AIOs before they reach steady state.
 
I have an update about the issues I was seeing. I did a little research and it appears the "PBO Fmax Enhancer" setting is causing issues for people using the Dark Hero board. I disabled that and ran Prime95 blended and it ran fine and temps hit about 62 degrees at roughly 4.5 Ghz all core. Prior to disabling Fmax Enhancer, all cores were hitting the 4.7-4.9 range and temps were soaring. I also ran the Small FFT and the CPU dropped clockspeeds to around 2.8 Ghz and had no issues running it, with a temp of around 59 degrees. I hadn't tried the Small FFT when I had the Fmax Enhancer enabled, so I'm not sure what the frequency and temps would've been.

I'm fine with those temps and will monitor system stability over the next week or so. I'll keep an eye out for BIOS updates which may fix this issue in the future - I'm on the latest standard BIOS release but there is a beta available, so I'm assuming that it will be available soon. Cinebench was hitting 88 degrees before but is now in the mid 60s, though I am seeing a big performance drop in that benchmark. If the system proves stable over the next week or so, I may start manually tweaking some values to overclock it in order to get some additional performance.
 
I have an update about the issues I was seeing. I did a little research and it appears the "PBO Fmax Enhancer" setting is causing issues for people using the Dark Hero board. I disabled that and ran Prime95 blended and it ran fine and temps hit about 62 degrees at roughly 4.5 Ghz all core. Prior to disabling Fmax Enhancer, all cores were hitting the 4.7-4.9 range and temps were soaring. I also ran the Small FFT and the CPU dropped clockspeeds to around 2.8 Ghz and had no issues running it, with a temp of around 59 degrees. I hadn't tried the Small FFT when I had the Fmax Enhancer enabled, so I'm not sure what the frequency and temps would've been.

I'm fine with those temps and will monitor system stability over the next week or so. I'll keep an eye out for BIOS updates which may fix this issue in the future - I'm on the latest standard BIOS release but there is a beta available, so I'm assuming that it will be available soon. Cinebench was hitting 88 degrees before but is now in the mid 60s, though I am seeing a big performance drop in that benchmark. If the system proves stable over the next week or so, I may start manually tweaking some values to overclock it in order to get some additional performance.
Did it drop to 2.8ghz with 5 threads? Small fft all core gives low temps while 5-7 threads is a different story on my dark hero with 5900x.
 
Did it drop to 2.8ghz with 5 threads? Small fft all core gives low temps while 5-7 threads is a different story on my dark hero with 5900x.

It looked like all cores dropped to 2.8 GHz on the small FFT. I need to re-enable the Fmax enhancer and see what happens.
 
It looked like all cores dropped to 2.8 GHz on the small FFT. I need to re-enable the Fmax enhancer and see what happens.
My all core on a 5900x at stock drops to 3.7 with all core small FFT so I do think it is intentional with max number of threads (24), but haven't seen it drop when using 5 or 6 threads.
 
My all core on a 5900x at stock drops to 3.7 with all core small FFT so I do think it is intentional with max number of threads (24), but haven't seen it drop when using 5 or 6 threads.

Do you have Fmax enhancer enabled? I can try to limit it to fewer threads and see what happens.
 
Do you have Fmax enhancer enabled? I can try to limit it to fewer threads and see what happens.
Nope, run D.O.C.P. and all else stock for day to day usage. Only OC occasionally for fun, but gains aren't worth it for permanent OC in my use cases.
 
Because I love to tweak... how are you all finding your best "cores" for single core targeted PBO offsets? Do you just let HWInfo or Monitor run in the background and whichever cores seems to boost the highest all the time, you pay attention to those cores? It seems my Core 0, 1, 3 and 7 are always hitting 5199Mhz; while all other cores will hit 4.8~5.0Ghz max. Does that mean I should start setting my PBO offsets per core and set the "higher boosting" cores with less of a negative curve offset? Right now I just have all cores at -20 offset and a Frequency boost max of 150Mhz.

Also, it seems my DDR4 will not budge past 3200Mhz without Prime95 errors... so I have moved on to extremely tight timings. Working on 2nd level and 3rd level right now. Fun fun... I so love tweaking stuff! I did this back on my x99 platform and there was a good 2~3% performance hidden in those timings. Not sure if AMD Zen3 will have that same effect. But, its fun nonetheless.
 
Nope, run D.O.C.P. and all else stock for day to day usage. Only OC occasionally for fun, but gains aren't worth it for permanent OC in my use cases.

I’ve been an overclocker since the 90s, but in the case of this system, I probably won’t do much other than whatever the board can do or possibly some light manual overclocking. I bought the 5950x specifically for work, virtualization, media work, and gaming - in roughly that order. Especially once I get a 3080 or 3090, an overclock is diminished in importance. Right now, I’m more interested in running stress tests just to make sure my cooling is holding up and it appears to be.
 
Because I love to tweak... how are you all finding your best "cores" for single core targeted PBO offsets? Do you just let HWInfo or Monitor run in the background and whichever cores seems to boost the highest all the time, you pay attention to those cores? It seems my Core 0, 1, 3 and 7 are always hitting 5199Mhz; while all other cores will hit 4.8~5.0Ghz max. Does that mean I should start setting my PBO offsets per core and set the "higher boosting" cores with less of a negative curve offset? Right now I just have all cores at -20 offset and a Frequency boost max of 150Mhz.

Also, it seems my DDR4 will not budge past 3200Mhz without Prime95 errors... so I have moved on to extremely tight timings. Working on 2nd level and 3rd level right now. Fun fun... I so love tweaking stuff! I did this back on my x99 platform and there was a good 2~3% performance hidden in those timings. Not sure if AMD Zen3 will have that same effect. But, its fun nonetheless.

Based on your previous post i've been testing more CO settings.

Core 0 and Core 3 are my best cores and have a -20 offset on them and the rest are -10 Max Frequency boost is set to 200

5800X best cores.png
 
I’ve been an overclocker since the 90s, but in the case of this system, I probably won’t do much other than whatever the board can do or possibly some light manual overclocking. I bought the 5950x specifically for work, virtualization, media work, and gaming - in roughly that order. Especially once I get a 3080 or 3090, an overclock is diminished in importance. Right now, I’m more interested in running stress tests just to make sure my cooling is holding up and it appears to be.
Overclocking 15-20 years ago was a lot more fun than it is nowdays. Back then 15-20% gains were common and occasionally 30%, but on the products from the last 4-5 years you are often stuck in the single digits :(
 
Overclocking 15-20 years ago was a lot more fun than it is nowdays. Back then 15-20% gains were common and occasionally 30%, but on the products from the last 4-5 years you are often stuck in the single digits :(
True, but most top tier products now are also so fast in stock form that the need/desire to overclock is no longer the difference between good and great performance... I'm OK with this and happy to just eek out a little more.
 
Overclocking 15-20 years ago was a lot more fun than it is nowdays. Back then 15-20% gains were common and occasionally 30%, but on the products from the last 4-5 years you are often stuck in the single digits :(

Exactly. I may notice a "big" drop in the benchmarks, but I loaded up a game and the FPS count was nearly the same. At this stage, I care more about stability since I'm using this for work and plan on building more VMs on it.
 
Overclocking 15-20 years ago was a lot more fun than it is nowdays. Back then 15-20% gains were common and occasionally 30%, but on the products from the last 4-5 years you are often stuck in the single digits :(

Due to node sizes getting smaller and difficulty getting rid of the heat. Those days of huge overclocks are gone.
 
I blame precision boost and other "opertunistic OC" features ;)

To be honest these days you have alot of new users that have never seen a old style bios and just use to UEFI. And they don't have the years of overclocking experience so for them these features are a god send. For us old school guys not as fun but its progress what can you do.
 
Broke 11k on CineBench R20 with this thing... pretty happy about that as at stock settings it benches under 10K for me. I'll take a 10% multi-core performance boost. :D

1618874684084.png
 
So am I missing something being an AMD "noob"? I can get extremely tight timings (see below) on my 4 sticks of 3200Mhz Samsung B-Die DDR4 32GB total (confirmed via part number and programs)... all at 1.38V on the DIMMs. SOC is at 1.1V (AMD recommended and default value hard set manually).

For the life of me, no matter how much voltage I throw at the DDR4, I can't even get 3400Mhz stable with MUCH looser timings, much less anything higher. I can run Prime95 all day at these tight timings, but go a notch above 3200Mhz and Prime95 will fail in 30 seconds and will BSOD my PC with "Memory Management" if I do not stop it quickly after a rounding error.

Was my ram likely binned to the max already (was a Quad G.Skill Kit from my x99 system)? I thought B-Die could OC nicely and much higher with voltage... Could it be some setting/voltage I am missing somewhere as I am not familiar with AMD? Did I just draw the short straw and my IMC sucks? Maybe Mobo/Ram/CPU/Bios not quite right?

I guess I have not tried a command rate of 2T, I have been using 1T, but even so, seems like having to go to 1T would be silly considering how tight i can run things at 3200Mhz.

I have FCLK=MCLK (1:1) locked as well.

Am I missing something here? :unsure:

1618928826257.png

cachemem_5950x_3200_8.png
 
So am I missing something being an AMD "noob"? I can get extremely tight timings (see below) on my 4 sticks of 3200Mhz Samsung B-Die DDR4 32GB total (confirmed via part number and programs)... all at 1.38V on the DIMMs. SOC is at 1.1V (AMD recommended and default value hard set manually).

For the life of me, no matter how much voltage I throw at the DDR4, I can't even get 3400Mhz stable with MUCH looser timings, much less anything higher. I can run Prime95 all day at these tight timings, but go a notch above 3200Mhz and Prime95 will fail in 30 seconds and will BSOD my PC with "Memory Management" if I do not stop it quickly after a rounding error.

Was my ram likely binned to the max already (was a Quad G.Skill Kit from my x99 system)? I thought B-Die could OC nicely and much higher with voltage... Could it be some setting/voltage I am missing somewhere as I am not familiar with AMD? Did I just draw the short straw and my IMC sucks? Maybe Mobo/Ram/CPU/Bios not quite right?

I guess I have not tried a command rate of 2T, I have been using 1T, but even so, seems like having to go to 1T would be silly considering how tight i can run things at 3200Mhz.

I have FCLK=MCLK (1:1) locked as well.

Am I missing something here? :unsure:

View attachment 349435
View attachment 349438

i'm currently on a 3200 B-die kit myself and can do 3600CL14 on my kit, however only at 16GB for now going up to 32GB's soon. Also have you tried the Ryzen Dram Calc?

Are you able to post a thaiphoon burner screenshot of your memory.

thaiphoon burner2.png


also for the timing's you are running in your screenshot which are low your latency is abit higher than what I see

Aida new bios.PNG


zen timing2.PNG
 
Last edited:
i'm currently on a 3200 B-die kit myself and can do 3600CL14 on my kit, however only at 16GB for now going up to 32GB's soon. Also have you tried the Ryzen Dram Calc?

Are you able to post a thaiphoon burner screenshot of your memory.

View attachment 349442

also for the timing's you are running in your screenshot which are low your latency is abit higher than what I see

View attachment 349443

View attachment 349444
I do notice you have "PowerDown" Disabled.... I believe mine is enabled. Does that even matter for the purpose of our conversation?

When you run at 3600Mhz, can you get a snapshot of that? Does it run at 2T timings or with Gear Down Mode enabled? I have yet to try either of those, but from my reading, it appears Gear Down is 1.5T.

Yes, I have used Ryzen Dram Calc... it actually spit out the numbers I am using right now at 3200Mhz; except I tightened them further with 13-13-13-26 (Calc spit out CL 14).

This is the Kit I have...

https://www.newegg.com/g-skill-32gb...820232219?Item=N82E16820232219&quicklink=true

Reading the part number on my individual sticks, I can 100% confirm it IS Samsung B-Die. (ends in 8810B)
 
I do notice you have "PowerDown" Disabled.... I believe mine is enabled. Does that even matter for the purpose of our conversation?

When you run at 3600Mhz, can you get a snapshot of that? Does it run at 2T timings or with Gear Down Mode enabled? I have yet to try either of those, but from my reading, it appears Gear Down is 1.5T.

Yes, I have used Ryzen Dram Calc... it actually spit out the numbers I am using right now at 3200Mhz; except I tightened them further with 13-13-13-26 (Calc spit out CL 14).

This is the Kit I have...

https://www.newegg.com/g-skill-32gb...820232219?Item=N82E16820232219&quicklink=true

Reading the part number on my individual sticks, I can 100% confirm it IS Samsung B-Die. (ends in 8810B)

These are the settings that I used when testing 3600

3600 CL14 Dram.PNG


And yes those sticks are B-die

b-die finder.PNG
 
does ryzen with fclk effect usual order of oc cpu first then memory? is it a constant back and forth now between cpu > mem > cpu until you got everything locked?
 
Based on your previous post i've been testing more CO settings.

Core 0 and Core 3 are my best cores and have a -20 offset on them and the rest are -10 Max Frequency boost is set to 200

View attachment 349106

This is interesting to me. Everything I've read has said that you need more voltage on the "good cores" and less on the "bad" cores. You're doing it backwards seemingly and getting great clock speeds. Can you screenshot your effective clocks? Curious if you're clock stretching.
 
These are the settings that I used when testing 3600

View attachment 349461

And yes those sticks are B-die

View attachment 349462
So I guess either my CPU or RAM can't run at CR1 above 3200Mhz, but by changing it to CR2, the world has opened up!

Not sure how high it is worth pushing on these things; I am at 1.42V on the DRAM (1.44V according to Mobo sensors, it overvolts I guess). But I got 3600Mhz stable at 14-14-14-28 @ 2T. Bandwidth is up and latency is down!

I could probably push 3800~4000 I bet if I start to loosen timings, but I have also read that 3600Mhz is the sweet spot on these CPUs. I do not really want to push my ram past 1.45V for 24/7 use.

1618943436990.png


cachemem_5950x_3600_3.png
 
This is interesting to me. Everything I've read has said that you need more voltage on the "good cores" and less on the "bad" cores. You're doing it backwards seemingly and getting great clock speeds. Can you screenshot your effective clocks? Curious if you're clock stretching.

Ran Geekbench 5 to get it to push the clocks and this is what I see.

i'm still playing with CO but so far i've done -15 on best cores and -5 on the rest and now -20 and -10 and my speeds have been good.

hwinfo geekbench.PNG
 
So I guess either my CPU or RAM can't run at CR1 above 3200Mhz, but by changing it to CR2, the world has opened up!

Not sure how high it is worth pushing on these things; I am at 1.42V on the DRAM (1.44V according to Mobo sensors, it overvolts I guess). But I got 3600Mhz stable at 14-14-14-28 @ 2T. Bandwidth is up and latency is down!

I could probably push 3800~4000 I bet if I start to loosen timings, but I have also read that 3600Mhz is the sweet spot on these CPUs. I do not really want to push my ram past 1.45V for 24/7 use.

View attachment 349510

View attachment 349511

Very nice.

What i'm curious about is why the more I tweak memory why does it seem there is a drop in write performance at L3 cache level. I see this in both our screenshots. Need to find out which memory settings is affecting that.
 
Very nice.

What i'm curious about is why the more I tweak memory why does it seem there is a drop in write performance at L3 cache level. I see this in both our screenshots. Need to find out which memory settings is affecting that.
I don't have the slightest idea... lol. The funny thing is; I am using the same exact timings for secondary that I used at 3200Mhz, so technically, the timings are even tighter. The only thing i can image is during memory training at boot; maybe the mobo is overwriting something that loosens things up for stability?

I have not touched the 3rd level timings quite yet, I have left those on Auto... however those (at least on Intel) typically only effected the Memory Bandwidth.
 
Very nice.

What i'm curious about is why the more I tweak memory why does it seem there is a drop in write performance at L3 cache level. I see this in both our screenshots. Need to find out which memory settings is affecting that.
So, I did a CMOS reset and reloaded my profile for the CPU and memory overclocks I had... all of a sudden, same exact settings (tested multiple times) and I finally got below 60ns latency and my L3 Cache numbers make sense again!

So maybe just try a reset and things work themselves out? lol.

cachemem_5950x_3600_5.png
 
Close to breaking 16K CPU score alone now... LOL.

I got to admit, learning to tweak and OC this AMD has been a blast so far... so much to learn and so much performance to gain past what everyone tells you these things can do! Certainly harder than tweaking an Intel system, but I love messing with this stuff.

http://www.3dmark.com/spy/19830166
 
So, I did a CMOS reset and reloaded my profile for the CPU and memory overclocks I had... all of a sudden, same exact settings (tested multiple times) and I finally got below 60ns latency and my L3 Cache numbers make sense again!

So maybe just try a reset and things work themselves out? lol.

View attachment 349578

lol nice.

Nah not going to do a reset. I can already hit 60ns at 3200 with my current setup. The numbers vary alittle depending on bios version, will find the setting that is affecting L3 writes just need to more investigating.

This is what I see at 3600CL14 on a slightly older bios will need to retest it with my current bios

PBO2 3600 CL14v2.PNG
 
lol nice.

Nah not going to do a reset. I can already hit 60ns at 3200 with my current setup. The numbers vary alittle depending on bios version, will find the setting that is affecting L3 writes just need to more investigating.

This is what I see at 3600CL14 on a slightly older bios will need to rest it with my current bios

View attachment 349590
Yeah, the latest bios for my MSI (beta bios) fixed the abysmal L3 cache performance issues these chips were having I guess. I didn't know any better when I built the system 2 weeks ago, but saw 100% speed increase in the L3 after testing the beta bios last weekend. I never did a cmos reset after the flash, so maybe something was iffy I didn't have access too. Its flying now though, so im happy. 😀
 
Yeah, the latest bios for my MSI (beta bios) fixed the abysmal L3 cache performance issues these chips were having I guess. I didn't know any better when I built the system 2 weeks ago, but saw 100% speed increase in the L3 after testing the beta bios last weekend. I never did a cmos reset after the flash, so maybe something was iffy I didn't have access too. Its flying now though, so im happy. 😀
If I remember correctly the Aida L3 fix was in Agesa 1.2.0.1 which was bios 3602 for my board. I'm currently on bios 3603 which is AGESA 1.2.0.1 Patch A with the USB fix. Glad you are happy always nice to have the new machine running optimal.
 
If I remember correctly the Aida L3 fix was in Agesa 1.2.0.1 which was bios 3602 for my board. I'm currently on bios 3603 which is AGESA 1.2.0.1 Patch A with the USB fix. Glad you are happy always nice to have the new machine running optimal.
Yeah, I'm on Agesa 1.2.0.2 now.. but now AMD and/or Creative Labs needs to fix this PCIe crap on the internal PCIe sound cards. There are so many reddits on creative's channel about the issue. Basically, unless I run my video card at PCIe 3.0 instead of 4.0, my sound card will randomly swap channels when the video card and system are under load (like gaming). Mobo sound is ridiculously bad (to me anyway, i tried)... so I need my SBXz to work as it goes to my z906 5.1 PC surround sound system. I guess all the new AE cards have the same issue too.
 
So i think i have my memory and IF pretty well tuned. I can do 4000CL18 / 2000 IF however i get WHEA errors regardless of what I have voltages set at. I just can't get my cpu to clock very well. Jealous of you guys hitting 5.0+ ghz. Any tips for me? Currently running PBO2, +200 mhz, with custom CO.

mem.png
 
So i think i have my memory and IF pretty well tuned. I can do 4000CL18 / 2000 IF however i get WHEA errors regardless of what I have voltages set at. I just can't get my cpu to clock very well. Jealous of you guys hitting 5.0+ ghz. Any tips for me? Currently running PBO2, +200 mhz, with custom CO.

View attachment 349690
Impressive latency and for AMD, I assume those are 1.45V sticks? You may want to update your bios, looks like you are on an older version based on your L3 cache numbers. I tried 3800Mhz last night on my kit; but my RAM did not like it, would not even post past 07 DeBug without 1.5V (too high for my 24/7 gaming blood). Then again, seeing as my sticks are rated for 3200Mhz, I'm happy at 3600Mhz with tight timings I guess... lol.

As for 5.0+; maybe just silicon lottery? I am hitting 5.2Ghz on Single Core boost now (4.9Ghz all core boost, 4.7Ghz average) with nothing more than PBO curve all core offset of -20 with +150Mhz FM. Getting close to 700 points for single core, but haven't quite cracked it yet!

Try setting your current limits, keeping temps lower allows for higher and longer boosts. I set my PBO advanced as follows:
PPT: 200
TDC: 200
EDC: 150

Capture7.JPG
 
Last edited:
Impressive latency and for AMD, I assume those are 1.45V sticks? You may want to update your bios, looks like you are on an older version based on your L3 cache numbers. I tried 3800Mhz last night on my kit; but my RAM did not like it, would not even post past 07 DeBug without 1.5V (too high for my 24/7 gaming blood). Then again, seeing as my sticks are rated for 3200Mhz, I'm happy at 3600Mhz with tight timings I guess... lol.

As for 5.0+; maybe just silicon lottery? I am hitting 5.2Ghz on Single Core boost now (4.9Ghz all core boost, 4.7Ghz average) with nothing more than PBO curve all core offset of -20 with +150Mhz FM. Getting close to 700 points for single core, but haven't quite cracked it yet!

View attachment 349699
That's insane, congrats on those clocks. Very impressive. I think you might be right on silicone lottery for my CPU. Although, I've been trying this morning to do the opposite of what I've read with regards to CO in that I am giving my "bad" cores LESS negative offset and my "good" cores MORE negative offset. Seeing higher clocks, so apparently Makaveli@BETA was on to something. It's just so weird that everywhere else has said to do the opposite.

1619009020236.png


As for my L3, I am on the latest stable for the Dark Hero. Previously I was getting in the 400 gb/s range, L3 has gone up over each successive bios update. There is a beta, but I am going to wait till it goes stable to update. Ram is at 1.35 volts, so pretty happy with those sticks.
 
Back
Top