Possible.That's a false narrative that you've created based on hearsay.
At that point they would need to be annoyed about themselves for wasting time trying to overclock where there's no additional room to exploit anymore. I'm not sure many can be honest with themselves like this.Or better yet, instead of being annoyed, they could be pleased about the fact that the factory overclock is already that good that manual overclocking beyond it is inevitably getting really difficult.
That's very true. When Blackberry was done as a phone, my ex-landlord who I was on friendly terms with, bought it cheap and wanted to set it up and experience it coz he had never used it before but had heard a lot of good things about over the years. Being a non-techie, he brings it to me on my off day and we had to install some software to set it up. I don't remember exactly what happened but we kept going at it for three hours straight but something would keep going wrong.I'm not sure many can be honest with themselves like this.
1.25v was as low as I could go with my 7700, otherwise it wouldn't post.I remember seeing some benchmarks that the X3D chips work best at 6000 ddr speeds, and don't benefit from anything beyond that.
I also noticed that the lower 1.25V vsoc brings down idle power use a bit, which is normally too high. I mjght try reducing it further to 1.2V and see if it remains stable.
Could be due to your particular RAM kit. I've read that some people are able to go as low as 1.15V.1.25v was as low as I could go with my 7700, otherwise it wouldn't post.
Mine is 1.02 on 7800x3D and has been stable. No Expo.
The way bigger L3$ hides better memory frequencies and timings in most cases. As a result of that X3D chips likely perform the best at the sweet spot of the IMC, meaning most memory (and as such IMC) overclocking has at best no and at worst averse effects.I remember seeing some benchmarks that the X3D chips work best at 6000 ddr speeds, and don't benefit from anything beyond that.
Do people still use memtest86? Obviously, you do. Most people have moved on from it years ago. It might still be good for something quick, but it is far from comprehensive.I currently have 3 AM5 CPU's.
7800X3D is memtest86 stable at 1v SOC with 2x16 6000 C30 expo
7900X is memtest86 stable at 1.05v SOC with 2x32 6000 C30 expo
7600 is memtest86 stable at 1.1v SOC with 2x16 6000 C36 expo
These aren't even really optimized. 7800X3D wasn't tested lower than 1v and I jumped from 1.05 to 1.1 on the 7600 without testing anything in between. All of them post at 1v SOC with expo enabled.
I just use it to quickly rule out hardware problems, rough in voltages, and establish a baseline of memory stability before I boot into an OS. It’s still a good tool for that purpose. The reason I qualified my voltages as memtest86 stable was so people reading my post didn’t get the impression that I’d adequately stress tested at those voltages.Do people still use memtest86? Obviously, you do. Most people have moved on from it years ago. It might still be good for something quick, but it is far from comprehensive.
There's TestMem5, Karhu's RAM Test, AIDA64, Prime95 Large FFT, OCCT, Y-Cruncher, and anything Linpack. I'd say even HCI MemTest Pro is better. Maybe even looping 3DMark demos would be better.
Totally agree with you there on Memtest86. I think a lot of folks like to rush their OCs & can't be bothered running through at least a couple of runs of MT86. But I'm probably even more cautious than some on here & straight after MT86, I'll run Windows memory diagnostics before running Memtest Pro or some other memory test program. I've stuck to this process for yrs already & never corrupted my OS if & only if Windows memory diagnostics does not report errors with the default run.I just use it to quickly rule out hardware problems, rough in voltages, and establish a baseline of memory stability before I boot into an OS. It’s still a good tool for that purpose. The reason I qualified my voltages as memtest86 stable was so people reading my post didn’t get the impression that I’d adequately stress tested at those voltages.
Edit: I thought I should also mention that I used memtest86 to test a fresh AM4 build over the weekend before installing windows and it was throwing errors. I put the memory kit into another CPU mobo combo and it was still throwing errors leading me to conclude I have a bad module and need to RMA the kit. It’s still an extremely useful testing and troubleshooting tool even if it’s not the best memory stress test in the world.
I didn't think I needed this vs 2.0 but I boot up my Linux drive today and keyboard inputs had crazy latency in GRUB. Worked fine after boot but annoying nonetheless. This appears to have fixed it.ASrock has a bios 2.02 out that: Optimize USB port compatibility
I have some problems when sometimes my keyboard is not on, when power on, so hopefully it fixes that.
I've noticed this behavior on both my MSI and Gigabyte boards. Certain USB ports have this issue before booting into the OS. Hopefully this fix rolls out to the other vendors as well. It makes booting from a USB drive very extremely painful - Fortunately the front panel USB C port and one of the rear USB C ports on all my boards works correctly, so I just have to use one of my type C drives.I didn't think I needed this vs 2.0 but I boot up my Linux drive today and keyboard inputs had crazy latency in GRUB. Worked fine after boot but annoying nonetheless. This appears to have fixed it.
Also really nice that the save user defaults feature now actually saves everything, and it was cross-compatible between the save I made in 2.0 and loading it in 2.02. It wasn't always like that with earlier BIOS releases.
All the USB ports need to have something plugged in. Otherwise the free ports get jealous and refuse to let the active port work properly. At least, that's my hypothesis.It took like several days for problems to develop, that required trying different USB port for stability.
Up to four 10Gb/s ports (with type C support), plus one lone USB2 port, can be driven by AMD's own CPU IOD. The rest would be attached to the X670/ B650/ A620 southbridge from ASMedia. Maybe the good ports are the ones from the IOD.Fortunately the front panel USB C port and one of the rear USB C ports on all my boards works correctly,
Is it possible for a bad mobo to overvolt the IOD?Maybe the good ports are the ones from the IOD.
That would be negatory. I've checked the block diagrams of the boards and it's not always CPU ports that function properly. On one board the only rear panel USB port that works is from the second/lower chipset.Up to four 10Gb/s ports (with type C support), plus one lone USB2 port, can be driven by AMD's own CPU IOD. The rest would be attached to the X670/ B650/ A620 southbridge from ASMedia. Maybe the good ports are the ones from the IOD.
SourceAMD’s David McAfee has confirmed AMD’s plans to support their AM5 CPU socket long-term. When asked about AMD’s future plans for AM5 and their Ryzen CPU lineup, McAfee reconfirmed that AMD plans to support AM5 through 2025.
2 yrs is long enough till' my next upgrade.Well, that's only two years away.
Yeah but it wouldn't be anything special. AM5 existing from late 2022 to late 2025 would possibly miss Zen 6. Even LGA1700 is going to last 3 years. It needs to support Zen 6 for it to be a true long-term support socket like AM4 was.2 yrs is long enough till' my next upgrade.
Curios to see how the 8000 series goes on X670E chipset boards!
The question is, what reasons would there be to move to a new socket?Yeah but it wouldn't be anything special. AM5 existing from late 2022 to late 2025 would possibly miss Zen 6. Even LGA1700 is going to last 3 years. It needs to support Zen 6 for it to be a true long-term support socket like AM4 was.