Recent content by stormlight

You are correct, I left cLDO VDGG at .9 after successful at .95. VDGG has to do with graphic bus but since 3900x has no graphic, I tried with .9 and it works. In your case, .95 should had been more stable but it did not boot. Weird! Well, i am glad it works out for you! You should run a...

MRAB54, I noticed alot of your values are not from XMP profile. This is high priority. Please see my attachment. It shows where XMP extracted information goes. Most of the other values are from DRAM Calc. Good luck and let me know your result! You can always tighten the timing after...

Hey, when you are booting with 3333, would you please upload a screenshot of your ryzen master? I want to look at your setting. I doubt memory is the problem. Have you tried to boot IF@1800 and ram at 3333MHz? Just to make sure IF can run at 1800MHz. The .9V was for CLDO VDDP and VDGG...

Hey, what RAM do you have? Rated speed/timing I am surprise Hynix C die has the exact XMP profiles as a D-Die. Did you also set the .900V setting?

Thank you for the input! I but share my knowledge through reading, understanding, and also "personally witnessing/experiencing it" and feedback to your own reading knowledge. Keep getting fed with old information or not to further own verification is more like brainwashing. Well I guess it is...

I strongly disagree with statement. I strongly believe the majority of folks can tightening their Vcore offset to approximate -.1V and reduce excess heat without losing any performance of PB2 or PBO on single or multi-core. As long as folks are informed and happy then that is all that matter...

Here is my 200W 3900X with no voltage off set.

Make sure you click on the report button after you read your ram. There are alot more settings. See my attachment!

Here are my data for single core bench with .1V offset!

Just a head up, 1.3V on my 3900X works just fine for multi-core but it drop performance on single core. Single core needs aroudn 1.38 to work properly. I think the internal chip voltage boost is limited to how much current it can boost. Try Stock multi-core and stock single-core and note this...

I agree with you there and thank your emphasizing it!!! But there are people out there looking for the cheap and just want good enough to get by and It can be done. For those few folks in the North does have the advantages! :-) I disagree on reduced performance and that is my personal...

This is ~15-20% in performance not clock. As I had said earlier in post, most program cannot take the full advantage of all the thread. CPU utilization was not at 100%. I have to run two instances of handbrake to witness this! Hey, what is your package power, cooler and fan rpm? Yes, I get...

Do you have another powersupply to test with? If you are a system builder, you should always have a handy spare power supply. Imo, PS is the number one failure or culprit to other components. Multi-rail is fine as long as they are redundant and not counted toward power supply rated wattage...

Kamikazi, You are EDC limited at the moment. Increase your Vcore (voltage) should help you reduce EDC (current) but you are close to maxing out your PPT (core power limit) anyway. But higher VCore will help you with high frequency boost! These processor is power limited by default. Go to bios...

According to the screen shot from Dan_D, it uses more power. PPT was at 205W + ~40W (larger L-Cache) = ~245W package. I was guessing around ~260W package power in my math for scaling. I am not sure on the thread count of handbrakes, but do 2-3 instance of handbrake in parallel and you will see...