-
Notifications
You must be signed in to change notification settings - Fork 128
SVID IMON Current monitoring? #486
Replies: 3 comments · 5 replies
-
Helpful!
|
Beta Was this translation helpful? Give feedback.
All reactions
-
I don't think these other ones are particularly helpful, but they were posted together with the other document in the original source: |
Beta Was this translation helpful? Give feedback.
All reactions
-
👀 1
-
Not sure if Purley is your processor architecture ?
But no CSR found to trigger So far, using latest version of CoreFreq, what are you getting from:
|
Beta Was this translation helpful? Give feedback.
All reactions
-
After building from the develop branch, my output is as follows:
|
Beta Was this translation helpful? Give feedback.
All reactions
-
👍 1
-
I think there's a regression on Memory Controller, DIMM topology and DDR4 Speed, with previous decoding |
Beta Was this translation helpful? Give feedback.
All reactions
-
Has your PC been upgraded to 4 populated DIMMs ? Please post |
Beta Was this translation helpful? Give feedback.
All reactions
-
I've had 4 DIMMs for a good while.
|
Beta Was this translation helpful? Give feedback.
All reactions
-
Thanks. Since DDR5, have been introduced those DDR4 issues:
|
Beta Was this translation helpful? Give feedback.
-
Does this document have enough information for adding current monitoring and other low-level SVID-related information for Intel CPUs? Reading model-specific registers is simple even via existing console tools, but how would accessing these work?
621792-svid-protocol-training-101-rev0p9.pdf
I got interested because recently a popular monitoring application on Windows (HWInfo) added a "SVID IOUT" sensor, and by polling it at a fast enough rate (e.g. 20ms) it looks like it might be possible to measure very brief current spikes in the CPU. It would be curious to see what happens at even shorter polling rates (although CoreFreq natively can only go as low as 100 ms).
In the end, the above document might or might not be helpful (I stumbled upon it by accident), but overall it seems like it should be possible to obtain current values directly from the processor's SVID protocol.
Beta Was this translation helpful? Give feedback.
All reactions