Asus TUF B550-PLUS + HWiNFO64 6.29-4230

[Zach]

Cores speed shows not correct, ASUS TUF GAMING B550-PLUS


View attachment 5161

You need to compare the Effective Clock: https://www.hwinfo.com/forum/threads/effective-clock-vs-instant-discrete-clock.5958/

@EKrava, in other words what you see on RyzenMaster is probably the effective clock or closer to that. But do not expect to see same numbers.
Open HWiNFO sensors window with RM together and compare.

Like this:

You will see 12 effective clock readings because eff. clock is calculated on all threads by HWiNFO, physical and logical cores

Current


Average


@Martin as you can see there is a lot of difference. I tried it with default polling period (2000ms) and shorter (1000, 500) and still the difference is a lot. I also did a readings reset on both with 1-2sec diff.
May it be very different the way RM calculate this? Has anything to do with 6/12 threads by RM/HWiNFO or is this another disclosed and proprietary "situation" by AMD, like RM's temp reporting?

 

[Martin]

@EKrava, in other words what you see on RyzenMaster is probably the effective clock or closer to that. But do not expect to see same numbers.
Open HWiNFO sensors window with RM together and compare.

Like this:

Current
View attachment 5162

Average
View attachment 5163

@Martin as you can see there is a lot of difference. I tried it with default polling period (2000ms) and shorter (1000, 500) and still the difference is a lot. I also did a readings reset on both with 1-2sec diff.
May it be very different the way RM calculate this? Is this another disclosed and proprietary "situation" by AMD, like RM's temp reporting?

RM uses a different method (which should not be affected by the operating system overhead) with more fine-grained averaging and handling of low-power states.
HWiNFO is theoretically able to use the same method as RM, but that would result in just another set of clock values in sensors window, which already contains lots of values and hence cause even more confusion to users. Moreover, the method used by RM is AMD Zen-specific and can't be used on other CPUs (i.e. Intel, older AMD models), while the Effective Clock currently reported is sort-of universal and supported by other vendors too. So I think it's better to use this universal method also for direct comparison with other vendors.

 

[Pictus]

Please try the new Beta build 4235, that should now report proper PCIe rate for the GPU when the rate is reduced at port.

Sorry for the late reply, but all seems fine with beta 6.31-4255, thanks!