victorluisgr

Fixed it, apparently the PBO I set on BIOS was not working (for some reason), but when I enabled it on Ryzen Master and set the CO values, it worked. Really don't know what happened, but it is working now so, your intuition to use Ryzen Master was right. Thanks.

So, after a 30 minute gaming session on a variety of CPU intensive games, I noticed in HWinfo that every core on the CPU is on the same frequency instead of one or two cores boosting to 5.05ghz, and others parking. I use process lasso on bitsum highest performance but I don't think that power profile affect frequencies in this way. There are some games that are really single-core performance dependent that I noticed a drop in performance. Anyone knows why is this happening or how to "fix" it? Edit: noticed that the "enhanced" on HWinfo is not there anymore, this happened after a BIOS reset. Edit: cinnebench R23 scores also dropped from 18100 to 16300 after this BIOS reset.

Just a quick question, because I cannot find this anywhere online. During medium load such as playing a fairly CPU intensive game, my 7800x3d hits 3.8ghz or 4ghz on some cores and randomly shoot to 4.8ghz to 5ghz only to come back and another core shoot to 5ghz again. It looks like the cores that are being used are random and those clocks fluctuate A LOT. Is this something I should be worrying about or is this some dynamic core parking? I don't know if it is relevant but, I did undervolt it to -20 fastest and second fastest and -25 the rest.

Went with -20 fastest, -25 second fastest and -30 6 cores and saw a tiny improvement in performance with effectively the same temps than the previous config. I think I undervolted too much. Thanks for the advice!

I've been using my 7800x3D for about a week and a half with -25 fastest, -30 second fastest and -40 6 cores, with no instability. But recently I saw some info that undervolting even if stable, may cause a core or two to not boost to it's full potential, but other sources say that undervolting increases performance because less heat means that the x3d chip can boost higher and for longer. Should I ease up on my undervolt? Or if this is specific from chip to chip is there any way I can test if my undervolt is reducing performance? As a side note, I did 9 hours of OCCT and hit only 4.55ghz all cores, I know that the 7800x3d only hits 5ghz on single core loads, but isn't 4.55ghz a little too low? Cinnebench R23: 18253

Heya! Just wanted to clarify, the memory in question has SK Hynix M-Die, not Samsung B-Die. The reason it is CL40 is that here in brazil it is A LOT cheaper to buy memory with high CAS latency and overclock it later rather than buying a guarateed overclock. Since I was made sure that the memory I was buying had Hynix Dies, I thought it was a great deal. I'm just bogged down by thoes 200mhz I'm missing, but considering that I payed around 30% - 35% less than a 6000mhz CL30, it is still worth it.

That's what I've been thinking. Those latencies that I got were waay above spec and 5800 to 6000 is not a whole lot of perfomance left on the table, maybe a frame or two in the end. Thanks anyway, gonna stop worrying about this.

I tried DOCP 6000mhz and it did not work. Only when I change it to 5800mhz, then it works perfectly.

I got a TUF GAMING B650M-Plus with version 1616 BIOS.

So, I got a CORSAIR 2x16 6000mhz CL40 memory kit (CMH32GX5M2B6000C40). It has SK Hynix M-die and I've successfully been able to overclock it to 5800mhz CL 30-36-36-28 with buildzoid timings (but I left FLCK on auto because I can't do 6000). Did several memory stability tests and passed all of them. But, the second I put 6000mhz instead of 5800mhz, I get a blue screen before I can even launch a stress test. Does my 7800x3d have a bad memory controller? Because a lot of posts says that 7800x3d should be able to reach 6000mhz. I'm using 1.25V SOC and 1.35V VDD and VDDQ.

Oh, so it would be a 13% loss in performance. That is huge, thank you, I'll probably just use a ethernet cable then.

I got a B650M-Plus, a micro-ATX board and was thinking about changing my USB wifi adapter to a PCI-e one. Right now the x16_1 is populated by a RTX 4070 that covers all other PCI-e slots, so if I want to install anything on the board I'll have to put the GPU on the x16_2 slot instead of the top x16_1 one. I've heard that the top slot is directly connected to the CPU whereas the second one goes through a slower path through the motherboard. Is that difference really noticeable?

Thank you very much, I got through the week without much problem using this approach. Got a new screw kit today from DeepCool and the installation went perfectly fine.

The screw simply does not go through the hole, It came from the same bag as the other three, but sadly that single one have a manufacturing error (I think).

Damn, I needed a working computer this week. Is there any other way just to make a temporary fix?