|Topic Review (Newest First)|
|03-24-2020 05:37 AM|
Your CPU is not defective, you are just getting a lower score in R20 than you should. This indicates something maybe not enabled in the BIOS. PBO, Precision Boost Overdrive, is not normally enabled on motherboards by default. Mine was set to Auto which seemed to never use it. You would most likely get scores that are closer to what everyone else gets if you enable it in the BIOS. My 3900X scored lower in R20 bench tests as well compared to what others were getting and it was my PBO settings. By default A-XMP, memory settings, and PBO seem to be not enabled(turned on) by default. Being that you are building your own system you will find that somethings you need to enable or change to get the best performance.
|03-24-2020 01:23 AM|
|03-23-2020 01:12 PM|
|03-23-2020 01:09 PM|
|03-23-2020 12:45 PM|
|03-23-2020 09:38 AM|
|Alex0401||plz tell me, my processor is not a factory defect if it passed the test successfully?|
|03-10-2020 06:07 AM|
|cosminmocan||something is not right with your score, it should be at least 5200.|
|03-10-2020 05:57 AM|
Same situation here, end of last year: Coming from Blue-Team (i4790K), I had to learn some things first. Maybe check the following guide first, even it's only a google-translation, it's ok to read:
|03-09-2020 09:06 PM|
updated my bios and upgraded from a 1700X to a 3800X
Thermaltake P3 White//AMD RYZEN 7 3800X 8-Core 3.9 GHz (4.5 GHz Turbo)
Corsair H100i-PRO//GIGABYTE GA-AX370-Gaming K5 MB
EVGA RTX-2080TI (11GB) 1TB-SSD//8TBHDD(x2)
G.SKILL TridentZ Series 32GB//1000 Watt Corsair PSU
|03-09-2020 10:28 AM|
Hi everyone! New owner of the 3800x here! Really glad coming back to AMD after over a decade! And boy, is this an upgrade over my trusty old, but aging 4790k. I mean, Cinebench R20 sums it up nicely. The 4790k balls to the wall overclocked (4.8GHz core at 1.344V, cache at 4.3GHz and RAM at 2400MHz CL10) nets me about 2300 points, best case scenario with almost nothing running in the background. The first run of the 3800x, nothing tweaked, just enabled PBO and set XMP, boom, 5000 flat. After a bit of overclocking to 4.4GHz all core it gets 5300 points. Insane.
Now, I have a few questions that scouring the web doesn't seem to hold any answers to.
1) Besides the basics, like messing with core multiplier, voltage and soc voltage, are there any other settings that are important to mess with to get a better overclock? I'm currently at 4.4GHz all core at 1.33125V in BIOS (somewhere around 1.325 with droop, maybe a tad lower), but I can't get 4.5GHz to work for the life of me at up to 1.425V. Other settings include 1.075V SoC voltage, and the other IO voltages whose names escape me right now are at 1.000v for the first two and 0.9V for the third one. RAM is currently set to 3733MHz CL16 at 1.35V and IF is 1:1 with the RAM at 1866MHz.
2) What is the maximum safe DDR4 voltage for the CPU's IMC? I mean, ok, Samsung B-die chips like on my G.Skill memory can take 1.5V easy from what I read, but what about the CPU itself? I could justify a RAM upgrade to 32GB if my RAM breaks, but I'd prefer that my CPU stays alive.
3) Does manually overclocking break C-states? As my system is set up, my cores barely enter C6 according to HWiNFO64. Ryzen Master does show the cores going to sleep, but usually shortly.
4) When manually overclocking, does the chip always stay at the voltage you set it to? In HWiNFO64 I get a lot of conflicting info. Core VID is always ~1.1V, core voltage stays at exactly what I set it to and the VCore that's I think reported by the motherboard shows as a tad lower (I'll assume due to droop). What gives? Ryzen Master shows the 1.1V VID reading, but a lower average core voltage. It's confusing to say the least. See the screenshot below of an x265 stress test I did to see what I mean, HWiNFO64 is open on the right.
5) Speaking of that screenshot, are clocks never actually reaching the value set in BIOS? Or is it misreporting the clocks in the "Effective core clock" readings? Because these usually hover at 4.1-4.3GHz under load and never ever reach 4.4GHz. Although they do sometimes show spikes to over 4.5GHz, but I don't believe those.
6) Is x265 stress test (from here) a good stress test for Ryzen 3rd gen? It was the most reliable for my old Intel, but I don't know if things have changed.
Thanks in advance for any input guys!
|This thread has more than 10 replies. Click here to review the whole thread.|