Skip to content

Commit

Permalink
add ta and td descriptions
Browse files Browse the repository at this point in the history
  • Loading branch information
@peterjunpark
peterjunpark committed Nov 7, 2024
1 parent 4cd346b commit ab47f7b
Show file tree
Hide file tree
Showing 6 changed files with 568 additions and 136 deletions.
116 changes: 94 additions & 22 deletions src/rocprof_compute_soc/analysis_configs/gfx906/1500_TA_and_TD.yaml
View file
Original file line number Diff line number Diff line change
Expand Up @@ -2,6 +2,78 @@
# Add description/tips for each metric in this section.
# So it could be shown in hover.
Metric Description:
Address Processing Unit:
Busy: &TA_Busy_tip >-
Percent of the total CU cycles the address processor was busy.
Address Stall: &TA_Address_Stall_tip >-
Percent of the total CU cycles the address processor was stalled from sending address requests
further into the vL1D pipeline.
Data Stall: &TA_Data_Stall_tip >-
Percent of the total CU cycles the address processor was stalled from sending write/atomic
data further into the vL1D pipeline.
Data-Processor → Address Stall: &TA_Data-Processor_Address_Stall_tip >-
Percent of total CU cycles the address processor was stalled waiting to send command data to
the data processor.
Total Instructions: &TA_Total_Instructions_tip >-
Global/Generic Instructions: &TA_Global_Generic_Instructions_tip >-
The total number of global and generic memory instructions executed on all compute units on
the accelerator, per normalization unit.
Global/Generic Read Instructions: &TA_Global_Generic_Read_tip >-
The total number of global and generic memory read instructions executed on all compute units
on the accelerator, per normalization unit.
Global/Generic Write Instructions: &TA_Global_Generic_Write_tip >-
The total number of global and generic memory write instructions executed on all compute units
on the accelerator, per normalization unit.
Global/Generic Atomic Instructions: &TA_Global_Generic_Atomic_tip >-
The total number of global and generic memory atomic (with and without return) instructions
executed on all compute units on the accelerator, per normalization unit.
Spill/Stack Instructions: &TA_Spill_Stack_Instructions_tip >-
The total number of spill/stack memory instructions executed on all compute units on the
accelerator, per normalization unit.
Spill/Stack Read Instructions: &TA_Spill_Stack_Read_tip >-
The total number of spill/stack memory read instructions executed on all compute units on the
accelerator, per normalization unit.
Spill/Stack Write Instructions: &TA_Spill_Stack_Write_tip >-
The total number of spill/stack memory write instructions executed on all compute units on the
accelerator, per normalization unit.
Spill/Stack Atomic Instructions: &TA_Spill_Stack_Atomic_tip >-
The total number of spill/stack memory atomic (with and without return) instructions executed
on all compute units on the accelerator, per normalization unit. Typically unused as these
memory operations are typically used to implement thread-local storage.
Spill/Stack Total Cycles: &TA_Spill_Stack_Total_Cycles_tip >-
The number of cycles the address processing unit spent working on spill/stack instructions,
per normalization unit.
Spill/Stack Coalesced Read: &TA_Spill_Stack_Coalesced_Read_tip >-
The number of cycles the address processing unit spent working on coalesced spill/stack read
instructions, per normalization unit.
Spill/Stack Coalesced Write: &TA_Spill_Stack_Coalesced_Write_tip >-
The number of cycles the address processing unit spent working on coalesced spill/stack write
instructions, per normalization unit.
Data-Return Path:
Data-Return Busy: &TD_Busy_tip >-
Percent of the total CU cycles the data-return unit was busy processing or waiting on data to
return to the CU.
Cache RAM → Data-Return Stall: &TD_Cache_RAM_Data-Return_Stall_tip >-
Percent of the total CU cycles the data-return unit was stalled on data to be returned from
the vL1D Cache RAM.
Workgroup Manager → Data-Return Stall: &TD_Workgroup_Manager_Data-Return_Stall_tip >-
Percent of the total CU cycles the data-return unit was stalled by the workgroup manager due
to initialization of registers as a part of launching new workgroups.
Coalescable Instructions: &TD_Coalescable_Instructions_tip >-
The number of instructions submitted to the data-return unit by the address processor that
were found to be coalescable, per normalization unit.
Read Instructions: &TD_Read_tip >-
The number of read instructions submitted to the data-return unit by the address processor
summed over all compute units on the accelerator, per normalization unit. This is expected to
be the sum of global/generic and spill/stack reads in the address processor.
Write Instructions: &TD_Write_tip >-
The number of store instructions submitted to the data-return unit by the address processor
summed over all compute units on the accelerator, per normalization unit. This is expected to
be the sum of global/generic and spill/stack stores counted by the vL1D cache-front-end.
Atomic Instructions: &TD_Atomic_tip >-
The number of atomic instructions submitted to the data-return unit by the address processor
summed over all compute units on the accelerator, per normalization unit. This is expected to
be the sum of global/generic and spill/stack atomics in the address processor.

# Define the panel properties and properties of each metric in the panel.
Panel Config:
Expand All @@ -24,97 +96,97 @@ Panel Config:
min: MIN(((100 * TA_TA_BUSY_sum) / ($GRBM_GUI_ACTIVE_PER_XCD * $cu_per_gpu)))
max: MAX(((100 * TA_TA_BUSY_sum) / ($GRBM_GUI_ACTIVE_PER_XCD * $cu_per_gpu)))
unit: pct
tips:
tips: *TA_Busy_tip
Address Stall:
avg: AVG(((100 * TA_ADDR_STALLED_BY_TC_CYCLES_sum) / ($GRBM_GUI_ACTIVE_PER_XCD * $cu_per_gpu)))
min: MIN(((100 * TA_ADDR_STALLED_BY_TC_CYCLES_sum) / ($GRBM_GUI_ACTIVE_PER_XCD * $cu_per_gpu)))
max: MAX(((100 * TA_ADDR_STALLED_BY_TC_CYCLES_sum) / ($GRBM_GUI_ACTIVE_PER_XCD * $cu_per_gpu)))
unit: pct
tips:
tips: *TA_Address_Stall_tip
Data Stall:
avg: AVG(((100 * TA_DATA_STALLED_BY_TC_CYCLES_sum) / ($GRBM_GUI_ACTIVE_PER_XCD * $cu_per_gpu)))
min: MIN(((100 * TA_DATA_STALLED_BY_TC_CYCLES_sum) / ($GRBM_GUI_ACTIVE_PER_XCD * $cu_per_gpu)))
max: MAX(((100 * TA_DATA_STALLED_BY_TC_CYCLES_sum) / ($GRBM_GUI_ACTIVE_PER_XCD * $cu_per_gpu)))
unit: pct
tips:
tips: *TA_Data_Stall_tip
Data-Processor → Address Stall:
avg: AVG(((100 * TA_ADDR_STALLED_BY_TD_CYCLES_sum) / ($GRBM_GUI_ACTIVE_PER_XCD * $cu_per_gpu)))
min: MIN(((100 * TA_ADDR_STALLED_BY_TD_CYCLES_sum) / ($GRBM_GUI_ACTIVE_PER_XCD * $cu_per_gpu)))
max: MAX(((100 * TA_ADDR_STALLED_BY_TD_CYCLES_sum) / ($GRBM_GUI_ACTIVE_PER_XCD * $cu_per_gpu)))
unit: pct
tips:
tips: *TA_Data-Processor_Address_Stall_tip
Total Instructions:
avg: AVG((TA_TOTAL_WAVEFRONTS_sum / $denom))
min: MIN((TA_TOTAL_WAVEFRONTS_sum / $denom))
max: MAX((TA_TOTAL_WAVEFRONTS_sum / $denom))
unit: (Instructions + $normUnit)
tips:
tips: *TA_Total_Instructions_tip
Global/Generic Instructions:
avg: AVG((TA_FLAT_WAVEFRONTS_sum / $denom))
min: MIN((TA_FLAT_WAVEFRONTS_sum / $denom))
max: MAX((TA_FLAT_WAVEFRONTS_sum / $denom))
unit: (Instructions + $normUnit)
tips:
tips: *TA_Global_Generic_Instructions_tip
Global/Generic Read Instructions:
avg: AVG((TA_FLAT_READ_WAVEFRONTS_sum / $denom))
min: MIN((TA_FLAT_READ_WAVEFRONTS_sum / $denom))
max: MAX((TA_FLAT_READ_WAVEFRONTS_sum / $denom))
unit: (Instructions + $normUnit)
tips:
tips: *TA_Global_Generic_Read_tip
Global/Generic Write Instructions:
avg: AVG((TA_FLAT_WRITE_WAVEFRONTS_sum / $denom))
min: MIN((TA_FLAT_WRITE_WAVEFRONTS_sum / $denom))
max: MAX((TA_FLAT_WRITE_WAVEFRONTS_sum / $denom))
unit: (Instructions + $normUnit)
tips:
tips: *TA_Global_Generic_Write_tip
Global/Generic Atomic Instructions:
avg: AVG((TA_FLAT_ATOMIC_WAVEFRONTS_sum / $denom))
min: MIN((TA_FLAT_ATOMIC_WAVEFRONTS_sum / $denom))
max: MAX((TA_FLAT_ATOMIC_WAVEFRONTS_sum / $denom))
unit: (Instructions + $normUnit)
tips:
tips: *TA_Global_Generic_Atomic_tip
Spill/Stack Instructions:
avg: AVG((TA_BUFFER_WAVEFRONTS_sum / $denom))
min: MIN((TA_BUFFER_WAVEFRONTS_sum / $denom))
max: MAX((TA_BUFFER_WAVEFRONTS_sum / $denom))
unit: (Instructions + $normUnit)
tips:
tips: *TA_Spill_Stack_Instructions_tip
Spill/Stack Read Instructions:
avg: AVG((TA_BUFFER_READ_WAVEFRONTS_sum / $denom))
min: MIN((TA_BUFFER_READ_WAVEFRONTS_sum / $denom))
max: MAX((TA_BUFFER_READ_WAVEFRONTS_sum / $denom))
unit: (Instructions + $normUnit)
tips:
tips: *TA_Spill_Stack_Read_tip
Spill/Stack Write Instructions:
avg: AVG((TA_BUFFER_WRITE_WAVEFRONTS_sum / $denom))
min: MIN((TA_BUFFER_WRITE_WAVEFRONTS_sum / $denom))
max: MAX((TA_BUFFER_WRITE_WAVEFRONTS_sum / $denom))
unit: (Instructions + $normUnit)
tips:
tips: *TA_Spill_Stack_Write_tip
Spill/Stack Atomic Instructions:
avg: AVG((TA_BUFFER_ATOMIC_WAVEFRONTS_sum / $denom))
min: MIN((TA_BUFFER_ATOMIC_WAVEFRONTS_sum / $denom))
max: MAX((TA_BUFFER_ATOMIC_WAVEFRONTS_sum / $denom))
unit: (Instructions + $normUnit)
tips:
tips: *TA_Spill_Stack_Atomic_tip
Spill/Stack Total Cycles:
avg: AVG((TA_BUFFER_TOTAL_CYCLES_sum / $denom))
min: MIN((TA_BUFFER_TOTAL_CYCLES_sum / $denom))
max: MAX((TA_BUFFER_TOTAL_CYCLES_sum / $denom))
unit: (Cycles + $normUnit)
tips:
tips: *TA_Spill_Stack_Total_Cycles_tip
Spill/Stack Coalesced Read:
avg: AVG((TA_BUFFER_COALESCED_READ_CYCLES_sum / $denom))
min: MIN((TA_BUFFER_COALESCED_READ_CYCLES_sum / $denom))
max: MAX((TA_BUFFER_COALESCED_READ_CYCLES_sum / $denom))
unit: (Cycles + $normUnit)
tips:
tips: *TA_Spill_Stack_Coalesced_Read_tip
Spill/Stack Coalesced Write:
avg: AVG((TA_BUFFER_COALESCED_WRITE_CYCLES_sum / $denom))
min: MIN((TA_BUFFER_COALESCED_WRITE_CYCLES_sum / $denom))
max: MAX((TA_BUFFER_COALESCED_WRITE_CYCLES_sum / $denom))
unit: (Cycles + $normUnit)
tips:
tips: *TA_Spill_Stack_Coalesced_Write_tip

- metric_table:
id: 1502
Expand All @@ -132,13 +204,13 @@ Panel Config:
min: MIN(((100 * TD_TD_BUSY_sum) / ($GRBM_GUI_ACTIVE_PER_XCD * $cu_per_gpu)))
max: MAX(((100 * TD_TD_BUSY_sum) / ($GRBM_GUI_ACTIVE_PER_XCD * $cu_per_gpu)))
unit: pct
tips:
tips: *TD_Busy_tip
Cache RAM → Data-Return Stall:
avg: AVG(((100 * TD_TC_STALL_sum) / ($GRBM_GUI_ACTIVE_PER_XCD * $cu_per_gpu)))
min: MIN(((100 * TD_TC_STALL_sum) / ($GRBM_GUI_ACTIVE_PER_XCD * $cu_per_gpu)))
max: MAX(((100 * TD_TC_STALL_sum) / ($GRBM_GUI_ACTIVE_PER_XCD * $cu_per_gpu)))
unit: pct
tips:
tips: *TD_Cache_RAM_Data-Return_Stall_tip
Workgroup manager → Data-Return Stall:
avg: # No perf counter
min: # No perf counter
Expand All @@ -150,7 +222,7 @@ Panel Config:
min: MIN((TD_COALESCABLE_WAVEFRONT_sum / $denom))
max: MAX((TD_COALESCABLE_WAVEFRONT_sum / $denom))
unit: (Instructions + $normUnit)
tips:
tips: *TD_Coalescable_Instructions_tip
Read Instructions:
avg: AVG((((TD_LOAD_WAVEFRONT_sum - TD_STORE_WAVEFRONT_sum) - TD_ATOMIC_WAVEFRONT_sum)
/ $denom))
Expand All @@ -159,16 +231,16 @@ Panel Config:
max: MAX((((TD_LOAD_WAVEFRONT_sum - TD_STORE_WAVEFRONT_sum) - TD_ATOMIC_WAVEFRONT_sum)
/ $denom))
unit: (Instructions + $normUnit)
tips:
tips: *TD_Read_tip
Write Instructions:
avg: AVG((TD_STORE_WAVEFRONT_sum / $denom))
min: MIN((TD_STORE_WAVEFRONT_sum / $denom))
max: MAX((TD_STORE_WAVEFRONT_sum / $denom))
unit: (Instructions + $normUnit)
tips:
tips: *TD_Write_tip
Atomic Instructions:
avg: AVG((TD_ATOMIC_WAVEFRONT_sum / $denom))
min: MIN((TD_ATOMIC_WAVEFRONT_sum / $denom))
max: MAX((TD_ATOMIC_WAVEFRONT_sum / $denom))
unit: (Instructions + $normUnit)
tips:
tips: *TD_Atomic_tip
Loading

0 comments on commit ab47f7b

Please sign in to comment.