-
Notifications
You must be signed in to change notification settings - Fork 39
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
List of architectures with PERF_METRICS MSR support #253
Comments
After making a script to find all occurrences of each Supports TMA level 1 metrics only:
Supports TMA levels 1 & 2:
|
Hi @willowec , 10th generation Ice Lake core introduced the first iteration of 12th generation P-core (Golden Cove core) extends the previous The model specific section of the SDM will also document MSRs. Then the documentation links 11th generation TGL to the prior generation. Where table 2-44 corresponded to ICL. You're tables are correct.
|
Hello,
I am working on a pull request for PAPI in order to add support for directly reading the
PERF_METRICS
MSR to acquire level 1 and level 2 Topdown Micro-architecture Analysis (TMA) metrics on architectures that support them. I know that this can be detected on a model by model basis by reading theIA32_PERF_CAPABILITIES
MSR, but that MSR cannot be read from userspace. Therefore, I am trying to assemble a list of CPUID Family/Model ID's for what processors support thePERF_METRICS
MSR so that I can hardcode support.I have already tried Reading the Intel Software Developer Manual Volume 4. This got me a meager list of 4 Family/Model ID's that support
PERF_METRICS
:06_7DH
,06_7EH
06_AAH
, and06_BDH
. This list is obviously not comprehensive as it does not include my own machine where I have been developing the component (06_B7H
).I see that in this repository, there is a file for many architectures called '[architecture]_metrics.json'. These files do individually list the events that are used in combination to derive topdown metrics, and it seems like it might be possible to use them to discover whether the
PERF_METRICS
MSR is supported on that architecture. Am I correct in guessing that, for example, line 60 of alderlake_metrics_goldencove_core.json indicates that the alderlake architecture supports theFETCH_LATENCY
field of thePERF_METRICS
MSR, and therefore level 2 TMA metrics in hardware? If so, does there already exist a complete list of what architectures support the MSR or should I manually assemble it?Thank you,
Willow
The text was updated successfully, but these errors were encountered: