Perf metrics: CYCLE_ACTIVITY.STALLS_L1D_MISS is greater than CYCLE_ACTIVITY.CYCLES_L1D_MISS

[Snooker]

 

Number of CYCLE_ACTIVITY.STALLS_L1D_MISS (r0CA3) is greater than the number of CYCLE_ACTIVITY.CYCLES_L1D_MISS (r08A3) in 13th Gen Intel(R) Core(TM) i9-13900KS. 

How is this possible?

References:
https://perfmon-events.intel.com/index.html?pltfrm=ahybrid.html&evnt=CYCLE_ACTIVITY.CYCLES_L1D_MISS

https://perfmon-events.intel.com/index.html?pltfrm=ahybrid.html&evnt=CYCLE_ACTIVITY.STALLS_L1D_MISS

[Earl_Intel]

Hi Snooker,

Thank you for posting in the communities!

To better assist you and determine the most appropriate resolution, could you please provide the needed details listed below:

• May I know what Benchmarking application you're using?
• What is the make and model of your motherboard?
• What is the make and model of your RAM? how many RAM sticks do you have?

Once we have this information, we will be able to proceed with the next course of action and work towards resolving the issue you are facing.

Best regards,

Earl E.

Intel Customer Support Technician

[Snooker]

Hi @Earl_Intel 

The model is 13th Gen Intel(R) Core(TM) i9-13900KS (Raptor Lake). I was trying to benchmark L1d stalls and miss.

I found one bug that I wasn't considering CMask. I was only checking EventSel and UMask.

Let me know if you have any inputs here and if you can help me with any documentation on how to get raw codes correctly.

Thanks

[Earl_Intel]

Hi Snooker,

I wanted to check if you had the chance to review the questions I posted. Please let me know at your earliest convenience so that we can determine the best course of action to resolve this matter.

Best regards,

Earl E.

Intel Customer Support Technician

[Earl_Intel]

Hi Snooker,

Thank you for sharing this information with us.

Can you clarify what Benchmarking application you are using? this will help us to try and replicate the issue on our end.

Best regards,

Earl E.

Intel Customer Support Technician

[Earl_Intel]

Hi Snooker,

Just wondering if you had a chance to look over the questions I posted.

Best regards,

Earl E.

Intel Customer Support Technician

[Snooker]

Hi @Earl_Intel 

I used Google Benchmark to measure the performance of matrix multiplication.

[Snooker]

It's a standard C++ application for matrix-vector multiplication that uses AVX instructions.

[Earl_Intel]

Hi Snooker,

Thank you for sharing this information with us.

No worries, I will further check on this internally, I'll provide an update on this thread as soon as possible.

Best regards,

Earl E.

Intel Customer Support Technician

[Earl_Intel]

Hi Snooker,

Thank you for patiently waiting on us.

To further troubleshoot the issue, you can try using our Intel Processor Diagnostic Tool to check the AVX capability of your processor.

Intel® Processor Diagnostic Tool

Let me know if your processor will past the test.

Best regards,

Earl E.

Intel Customer Support Technician

[Earl_Intel]

Hi Snooker,

I just wanted to check if you had tried the Intel Processor Diagnostic Tool yet.

Here is the link for your convenience: Intel® Processor Diagnostic Tool

Best regards,

Earl E.

Intel Customer Support Technician

[Earl_Intel]

Hi Snooker,

I have not heard back from you so I will close this inquiry now. If you need further assistance, please submit a new question as this thread will no longer be monitored.

Best regards,

Earl E.

Intel Customer Support Technician