AI/rocm-systems
2
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

metrics fixes

Browse Source 
[ROCm/rocprofiler commit: 78c432aee2]
This commit is contained in:
Evgeny
2018-02-08 13:31:43 -06:00
parent 6960dcd77e
commit b9907b49ea
1 changed files with 5 additions and 5 deletions
Download Patch File Download Diff File Expand all files Collapse all files
projects/rocprofiler/test/metrics.xml
+5 -5
View File
@@ -110,10 +110,10 @@
# sum for 16 instances
<metric name="TA_FLAT_READ_WAVEFRONTS_sum" expr=sum(TA_FLAT_READ_WAVEFRONTS,16) ></metric>
<metric name="TA_FLAT_WRITE_WAVEFRONTS_sum" expr=sum(TA_FLAT_WRITE_WAVEFRONTS,16) ></metric>
<metric name=TCC_HIT_sum expr=sum(TCC_HIT,16) ></metric>
<metric name=TCC_MISS_sum expr=sum(TCC_MISS,16) ></metric>
<metric name=TCC_EA_RDREQ_sum expr=sum(TCC_EA_RDREQ,16) ></metric>
<metric name=TCC_EA_RDREQ_32B_sum expr=sum(TCC_EA_RDREQ_32B,16) ></metric>
<metric name="TCC_HIT_sum" expr=sum(TCC_HIT,16) ></metric>
<metric name="TCC_MISS_sum" expr=sum(TCC_MISS,16) ></metric>
<metric name="TCC_EA_RDREQ_sum" expr=sum(TCC_EA_RDREQ,16) ></metric>
<metric name="TCC_EA_RDREQ_32B_sum" expr=sum(TCC_EA_RDREQ_32B,16) ></metric>
# FETCH_SIZE, kilobytes
# The total kilobytes fetched from the video memory. This is measured with all extra fetches and any cache or memory effects taken into account.
@@ -269,7 +269,7 @@
# The percentage of GPUTime ALU units are stalled by the LDS input queue being full or the output queue being not ready. If there are LDS bank conflicts, reduce them. Otherwise, try reducing the number of LDS accesses if possible. Value range: 0% (optimal) to 100% (bad).
<metric
name="ALUStalledByLDS"
expr=100*SQ_WAIT_INST_LDS/SQ_WAVES/GRBM_GUI_ACTIVE/NUM_SHADER_ENGINES
expr=100*SQ_WAIT_INST_LDS*4/SQ_WAVES/GRBM_GUI_ACTIVE
descr="The percentage of GPUTime ALU units are stalled by the LDS input queue being full or the output queue being not ready. If there are LDS bank conflicts, reduce them. Otherwise, try reducing the number of LDS accesses if possible. Value range: 0% (optimal) to 100% (bad)."
></metric>