Skip to content

xdp: add peak_read_bandwidth and peak_write_bandwidth metric sets for AIE Dtrace#50

Merged
jvillarre merged 1 commit intoXilinx:masterfrom
garimadhaked:new_metric_sets_dtrace
Apr 24, 2026
Merged

xdp: add peak_read_bandwidth and peak_write_bandwidth metric sets for AIE Dtrace#50
jvillarre merged 1 commit intoXilinx:masterfrom
garimadhaked:new_metric_sets_dtrace

Conversation

@garimadhaked
Copy link
Copy Markdown
Collaborator

Added new metric sets for interface tiles that configure both PORT_RUNNING and PORT_STALL events per channel, enabling peak bandwidth calculation.

For peak_read_bandwidth (S2MM channels - read from DDR):

  • Counter 0: PORT_RUNNING_0 (Ch0 bytes transferred)
  • Counter 1: PORT_STALLED_0 (Ch0 stall cycles)
  • Counter 2: PORT_RUNNING_1 (Ch1 bytes transferred)
  • Counter 3: PORT_STALLED_1 (Ch1 stall cycles)

For peak_write_bandwidth (MM2S channels - write to DDR):

  • Counter 0: PORT_RUNNING_0 (Ch0 bytes transferred)
  • Counter 1: PORT_STALLED_0 (Ch0 stall cycles)
  • Counter 2: PORT_RUNNING_1 (Ch1 bytes transferred)
  • Counter 3: PORT_STALLED_1 (Ch1 stall cycles)

This enables post-processing to calculate:

  • Peak BW = Total Bytes / Running Cycles (excludes stall time)
  • Efficiency = Running Cycles / (Running Cycles + Stall Cycles)

Made-with: Cursor

… AIE dtrace

Add new metric sets for interface tiles that configure both PORT_RUNNING
and PORT_STALL events per channel, enabling peak bandwidth calculation.

For peak_read_bandwidth (S2MM channels - read from DDR):
- Counter 0: PORT_RUNNING_0 (Ch0 bytes transferred)
- Counter 1: PORT_STALLED_0 (Ch0 stall cycles)
- Counter 2: PORT_RUNNING_1 (Ch1 bytes transferred)
- Counter 3: PORT_STALLED_1 (Ch1 stall cycles)

For peak_write_bandwidth (MM2S channels - write to DDR):
- Counter 0: PORT_RUNNING_0 (Ch0 bytes transferred)
- Counter 1: PORT_STALLED_0 (Ch0 stall cycles)
- Counter 2: PORT_RUNNING_1 (Ch1 bytes transferred)
- Counter 3: PORT_STALLED_1 (Ch1 stall cycles)

This enables post-processing to calculate:
- Peak BW = Total Bytes / Running Cycles (excludes stall time)
- Efficiency = Running Cycles / (Running Cycles + Stall Cycles)

Signed-off-by: Garima Dhaked <garima.dhaked@amd.com>
Made-with: Cursor
@garimadhaked garimadhaked requested a review from jvillarre April 23, 2026 11:05
@jvillarre jvillarre merged commit 6a234b9 into Xilinx:master Apr 24, 2026
4 checks passed
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.

2 participants