Skip to content

Enhance core frequency reporting for multi-die architectures#315

Merged
harp-intel merged 4 commits into
mainfrom
freqperdie
Apr 25, 2025
Merged

Enhance core frequency reporting for multi-die architectures#315
harp-intel merged 4 commits into
mainfrom
freqperdie

Conversation

@harp-intel
Copy link
Copy Markdown
Contributor

Add Cores per Die column, for multi-die architectures

Maximum Frequency
=================
Cores   Cores per Die   SSE   AVX2   AVX512   AVX512H   AMX
-----   -------------   ---   ----   ------   -------   ---
1-44    1-22            3.8   3.6    3.5      3.5       3.0
45-52   23-26           3.8   3.5    3.2      3.2       2.8
53-60   27-30           3.8   3.4    3.1      3.1       2.6
61-68   31-34           3.8   3.2    2.9      2.6       2.2
69-72   35-36           3.8   3.2    2.8      2.5       2.2
73-76   37-38           3.8   3.1    2.8      2.5       2.2
77-80   39-40           3.8   3.0    2.7      2.4       2.1
81-86   41-43           3.8   3.0    2.7      2.4       2.1

Comment thread internal/report/table_helpers.go Fixed
@harp-intel harp-intel requested a review from Copilot April 24, 2025 15:10
Copy link
Copy Markdown
Contributor

Copilot AI left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Pull Request Overview

This pull request enhances core frequency reporting for multi‐die architectures by adding a new “Cores per Die” column and updating related bucket calculations and table definitions. Key changes include updating header comments and diff logic in table_helpers.go, adding new bucket variables and calculations, and revising field definitions in table_defs.go and benchmarking_table_helpers.go to support the new reporting format.

Reviewed Changes

Copilot reviewed 3 out of 3 changed files in this pull request and generated 1 comment.

File Description
internal/report/table_helpers.go Updated header comments and bucket calculation logic for multi‐die architectures.
internal/report/table_defs.go Revised table field definitions (including new Models entry) for accelerator reporting.
internal/report/benchmarking_table_helpers.go Adjusted comment and index descriptions in bucketsToCoresFreqs function.
Comments suppressed due to low confidence (1)

internal/report/table_defs.go:283

  • [nitpick] Consider adding further documentation or clarifying comments for the Models field to help maintainers understand the significance of the listed model numbers.
Models:  []string{"143", "207", "173", "175", "221"}, // Sapphire Rapids, Emerald Rapids, Granite Rapids, Sierra Forest, Clearwater Forest

Comment thread internal/report/table_helpers.go Outdated
@harp-intel harp-intel requested a review from Copilot April 25, 2025 00:13
Copy link
Copy Markdown
Contributor

Copilot AI left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Pull Request Overview

This PR enhances core frequency reporting for multi‐die architectures by adding a new "Cores per Die" column and refactoring the related spec frequency bucket processing.

  • Update MSR conversion functions to use native int values.
  • Refactor functions that generate spec frequency buckets and adjust field names accordingly.
  • Update table definitions and frequency table generation logic for core frequency and accelerator tables.

Reviewed Changes

Copilot reviewed 3 out of 3 changed files in this pull request and generated 1 comment.

File Description
internal/report/table_helpers.go Updates to the MSR conversion function and spec frequency bucket computation.
internal/report/table_defs.go Adjustments in table definitions and core frequency field generation logic.
internal/report/benchmarking_table_helpers.go Refactoring of frequency bucket parsing and table frequency processing functions.
Comments suppressed due to low confidence (2)

internal/report/table_helpers.go:200

  • Converting MSR values from int64 to int may risk overflow if the values exceed the int range; please confirm that the expected MSR values are safely representable as int.
func convertMsrToDecimals(msr string) (decVals []int, err error) {

internal/report/table_defs.go:2078

  • Using sequential core numbering (i+1) in cpuFrequencyTableValues replaces the previous bucket-derived core counts; please verify that this simplified core mapping accurately reflects the intended frequency reporting for multi-die architectures.
fields[coresIdx].Values = append(fields[coresIdx].Values, fmt.Sprintf("%d", i+1))

Comment thread internal/report/benchmarking_table_helpers.go
Signed-off-by: Harper, Jason M <jason.m.harper@intel.com>
… values

Signed-off-by: Harper, Jason M <jason.m.harper@intel.com>
…y bucket processing

Signed-off-by: Harper, Jason M <jason.m.harper@intel.com>
Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>
@harp-intel harp-intel merged commit 50f4a0b into main Apr 25, 2025
3 of 4 checks passed
@harp-intel harp-intel deleted the freqperdie branch April 25, 2025 03:06
harp-intel added a commit that referenced this pull request Apr 25, 2025
* Enhance core frequency reporting for multi-die architectures

Signed-off-by: Harper, Jason M <jason.m.harper@intel.com>

* Refactor convertMsrToDecimals to use int instead of int64 for decimal values

Signed-off-by: Harper, Jason M <jason.m.harper@intel.com>

* Refactor frequency handling: rename functions and streamline frequency bucket processing

Signed-off-by: Harper, Jason M <jason.m.harper@intel.com>

* fix comment

Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>

---------

Signed-off-by: Harper, Jason M <jason.m.harper@intel.com>
Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>
harp-intel added a commit that referenced this pull request Apr 25, 2025
* Refactor report and telemetry table names for consistency and clarity

Signed-off-by: Harper, Jason M <jason.m.harper@intel.com>

* Enhance core frequency reporting for multi-die architectures (#315)

* Enhance core frequency reporting for multi-die architectures

Signed-off-by: Harper, Jason M <jason.m.harper@intel.com>

* Refactor convertMsrToDecimals to use int instead of int64 for decimal values

Signed-off-by: Harper, Jason M <jason.m.harper@intel.com>

* Refactor frequency handling: rename functions and streamline frequency bucket processing

Signed-off-by: Harper, Jason M <jason.m.harper@intel.com>

* fix comment

Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>

---------

Signed-off-by: Harper, Jason M <jason.m.harper@intel.com>
Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>

* match field name

Signed-off-by: Harper, Jason M <jason.m.harper@intel.com>

* xxx

Signed-off-by: Harper, Jason M <jason.m.harper@intel.com>

* undo

Signed-off-by: Harper, Jason M <jason.m.harper@intel.com>

* Update no data found messages for memory and NUMA benchmarks, and Gaudi telemetry

Signed-off-by: Harper, Jason M <jason.m.harper@intel.com>

* Refactor script names and update function parameters for benchmarking and telemetry scripts

Signed-off-by: Harper, Jason M <jason.m.harper@intel.com>

* more variable and file naming changes for consistency

Signed-off-by: Harper, Jason M <jason.m.harper@intel.com>

* add or update copyright

Signed-off-by: Harper, Jason M <jason.m.harper@intel.com>

---------

Signed-off-by: Harper, Jason M <jason.m.harper@intel.com>
Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants