Intensely consuming power from AMD Ryzen 9 5950X processors
Intensely consuming power from AMD Ryzen 9 5950X processors
This setup would cause issues, as mixing matched all sticks with XMP together isn't supported. The IMC would dislike this arrangement. It can't handle such configurations. The system only allows up to 3200mt/s initially, and combining memory types hasn't been typical.
I would dismiss memory problems since he achieved 2133mhz during the third RMA.
This has no relevance to me. I’d use the Corsair memory and execute just the G.Skill B-Die. Likely fewer or none of the problems.
I just tried two different memory kits here without mixing them. Apologies for any confusion. The Corsair kit was my first DDR4 set since I moved from AMD Phenom to Intel Haswell, and it still functions at the rated speeds. It uses Samsung B-die memory, which is a bit outdated and only supports 2133 MHz SPD. Since it’s a 4-stick kit, I chose the ASRock board because of its T-Topology DRAM design. I switched to the G.Skill kit simply because they were marked as a QVL kit for this board—both are Samsung B-die and rated for 2666 MHz SPD. Both kits are performing well with XMP and a 1:1 FCLK ratio on my 3950X. I ordered an MSI MAG X570 Tomahawk WIFI board today as a replacement. It looks like it has a better layout and all the features I need. Just a few missing conveniences, but honestly: Nobody’s perfect ^^ B550 is out because the PCIe 3.0 interface might affect its connectivity speeds.
Thank you for your message. I was uncertain at first. Hoping the new board works out. I'll monitor this discussion and wish for a favorable outcome.
I updated my discussion about the new setup. I swapped in the MSI MAG X570 Tomahawk WiFi and tried some tests to see if problems arose. The supposed dead 5950X worked perfectly—no crashes, restarts, or BSODs on the new board. AMD isn’t keen on swapping the existing chip, is it? What stood out was that the MSI unit caps at around 5.0 GHz single-threaded boost and doesn’t exceed that in the factory setting. In contrast, the ASRock X570 Phantom Gaming X model had all chips running roughly 50-100 MHz higher across most tasks (up to 5.1 GHz on Cinebench single-thread), with an average VCore of about 0.05V and SoC/VDDP/VDDG up by 0.1V. It looks like the ASRock board pushed the chips beyond their limits—possibly overclocking until instability set in. Overall, I think the MSI unit performs better than the ASRock version, and maybe even edges out the ASRock X570 Taichi (same design) despite missing features like power/reset buttons, a debug LCD, and an extra Intel Ethernet controller. The MSI board also stays cooler—about 10°C less with a smaller heat sink, better wiring layout, and WiFi antennas with stronger signals. The ASRock model ran at higher voltages and temperatures, likely due to its simpler but more efficient cooling system. The only drawbacks I noticed were the lack of basic RGB controls in the BIOS and occasional issues with the HP Reverb G2 on rear USB ports (except one front header).