From 6a4321c4e30b44915b86ea4af9ef405db2f65a4e Mon Sep 17 00:00:00 2001
From: Mythreya <Mythreya.Kuricheti@amd.com>
Date: Fri, 27 Oct 2023 10:59:11 -0700
Subject: [PATCH] Add MI200/MI300 counters

Revision - Addition [Impact SoC: MI200, MI300]
    Note: this set of counters are important help understand the
          bottleneck.
    1.	TCC_TAG_STALL
        a.	Metric:  TCC_TAG_STALL/TCC_CYCLE: percentage of time TCC
            tag lookup pipeline is stalled
    2.	TCP_TCR_TCP_STALL_CYCLES
        a.	Metric: TCP_TCR_TCP_STALL_CYCLES/TCP_GATE_EN1: percentage
            of time TCP is stalled by TCR

Revision - Addition [Impact SoC: MI300]
    3.	TCC_BUBBLE:
        a.	Definition: Number of 128-byte read requests sent to EA
        b.	Revised Metric #1, TCC-EA Read BW:
        ReadBW =   128 * TCC_BUBBLE
                + 64 * (TCC_EA0_RDREQ - TCC_BUBBLE - TCC_EA0_RDREQ_32B)
                + 32 * TCC_EA0_RDREQ_32B
        c.	Revised Metric #2: TCC_EA Read Latency
        ReadLatency = TCC_EA0_RDREQ_LEVEL / (TCC_BUBBLE + TCC_EA0_RDREQ)

    /* [Fineprint] More detailed arithmetic:
    * ReadLatency = TCC_EA0_RDREQ_LEVEL / (#32B_req + #64B_req + #128B_req * 2)
    */

Change-Id: I0a2dfc1b64ca97023b1e8ba0f9830330b3034946


[ROCm/rocprofiler commit: 46e02a98666794821b7ddc5895115070c10ee4da]
---
 .../rocprofiler/src/core/counters/basic/gfx_metrics.xml    | 1 +
 projects/rocprofiler/src/core/counters/derived/metrics.xml | 7 +++++++
 .../src/core/counters/metrics/basic_counters.xml           | 1 +
 .../src/core/counters/metrics/derived_counters.xml         | 7 +++++++
 projects/rocprofiler/test/tool/gfx_metrics.xml             | 1 +
 projects/rocprofiler/test/tool/metrics.xml                 | 7 +++++++
 6 files changed, 24 insertions(+)

diff --git a/projects/rocprofiler/src/core/counters/basic/gfx_metrics.xml b/projects/rocprofiler/src/core/counters/basic/gfx_metrics.xml
index a2e1efc756..80b877b2f3 100755
--- a/projects/rocprofiler/src/core/counters/basic/gfx_metrics.xml
+++ b/projects/rocprofiler/src/core/counters/basic/gfx_metrics.xml
@@ -603,6 +603,7 @@
     <metric name="TCC_EA0_RD_UNCACHED_32B" block="TCC" event="40" descr="Number of 32-byte TCC/EA read due to uncached traffic. A 64-byte request will be counted as 2"></metric>
     <metric name="TCC_EA0_RDREQ_LEVEL" block="TCC" event="44" descr="The sum of the number of TCC/EA read requests in flight. This is primarily meant for measure average EA read latency. Average read latency = TCC_PERF_SEL_EA_RDREQ_LEVEL/TCC_PERF_SEL_EA_RDREQ."></metric>
     <metric name="TCC_TAG_STALL" block="TCC" event="45" descr="Number of cycles the normal request pipeline in the tag was stalled for any reason. Normally, stalls of this nature are measured exactly from one point the pipeline, but that is not the case for this counter. Probes can stall the pipeline at a variety of places, and there is no single point that can reasonably measure the total stalls accurately."></metric>
+    <metric name="TCC_BUBBLE" block="TCC" event="56" descr="Number of 128-byte read requests sent to EA."></metric>
     <metric name="TCC_NORMAL_WRITEBACK" block="TCC" event="68" descr="Number of writebacks due to requests that are not writeback requests."></metric>
     <metric name="TCC_ALL_TC_OP_WB_WRITEBACK" block="TCC" event="73" descr="Number of writebacks due to all TC_OP writeback requests."></metric>
     <metric name="TCC_NORMAL_EVICT" block="TCC" event="74" descr="Number of evictions due to requests that are not invalidate or probe requests."></metric>
diff --git a/projects/rocprofiler/src/core/counters/derived/metrics.xml b/projects/rocprofiler/src/core/counters/derived/metrics.xml
index 83403437fb..33793b0f7e 100755
--- a/projects/rocprofiler/src/core/counters/derived/metrics.xml
+++ b/projects/rocprofiler/src/core/counters/derived/metrics.xml
@@ -213,6 +213,7 @@
     <metric name="TCP_TCC_RW_WRITE_REQ_sum" expr="sum(TCP_TCC_RW_WRITE_REQ,16)" descr="Total write requests with RW mtype from this TCP to all TCCs. Sum over TCP instances."></metric>
     <metric name="TCP_TCC_RW_ATOMIC_REQ_sum" expr="sum(TCP_TCC_RW_ATOMIC_REQ,16)" descr="Total atomic requests with RW mtype from this TCP to all TCCs. Sum over TCP instances."></metric>
     <metric name="TCP_PENDING_STALL_CYCLES_sum" expr="sum(TCP_PENDING_STALL_CYCLES,16)" descr="Stall due to data pending from L2. Sum over TCP instances."></metric>
+    <metric name="TCP_TCR_TCP_STALL_CYCLES_PERCENT" expr="100*TCP_TCR_TCP_STALL_CYCLES_sum/TCP_GATE_EN1_sum" descr="Percentage of time TCP is stalled by TCR."></metric>
     <metric name="TCA_CYCLE_sum" expr="sum(TCA_CYCLE,16)" descr="Number of cycles. Sum over all TCA instances "></metric>
     <metric name="TCA_BUSY_sum" expr="sum(TCA_BUSY,16)" descr="Number of cycles we have a request pending. Sum over all TCA instances."></metric>
     <metric name="TCC_BUSY_avr" expr="avr(TCC_BUSY,32)" descr="TCC_BUSY avr over all memory channels."></metric>
@@ -252,6 +253,7 @@
     <metric name="TCC_EA_RDREQ_GMI_CREDIT_STALL_sum" expr="sum(TCC_EA_RDREQ_GMI_CREDIT_STALL,32)" descr="Number of cycles there was a stall because the read request interface was out of GMI credits. Stalls occur regardless of whether a read needed to be performed or not. Sum over TCC instances."></metric>
     <metric name="TCC_EA_RDREQ_DRAM_CREDIT_STALL_sum" expr="sum(TCC_EA_RDREQ_DRAM_CREDIT_STALL,32)" descr="Number of cycles there was a stall because the read request interface was out of DRAM credits. Stalls occur regardless of whether a read needed to be performed or not. Sum over TCC instances."></metric>
     <metric name="TCC_TAG_STALL_sum" expr="sum(TCC_TAG_STALL,32)" descr="."></metric>
+    <metric name="TCC_TAG_STALL_PERCENT" expr="100*TCC_TAG_STALL_sum/TCC_CYCLE_sum" descr="Percentage of time the TCC tag lookup pipeline is stalled."></metric>
     <metric name="TCC_NORMAL_WRITEBACK_sum" expr="sum(TCC_NORMAL_WRITEBACK,32)" descr="Number of writebacks due to requests that are not writeback requests. Sum over TCC instances."></metric>
     <metric name="TCC_ALL_TC_OP_WB_WRITEBACK_sum" expr="sum(TCC_ALL_TC_OP_WB_WRITEBACK,32)" descr="Number of writebacks due to all TC_OP writeback requests. Sum over TCC instances."></metric>
     <metric name="TCC_NORMAL_EVICT_sum" expr="sum(TCC_NORMAL_EVICT,32)" descr="Number of evictions due to requests that are not invalidate or probe requests. Sum over TCC instances."></metric>
@@ -342,6 +344,7 @@
     <metric name="TCP_TCC_RW_WRITE_REQ_sum" expr="sum(TCP_TCC_RW_WRITE_REQ,16)" descr="Total write requests with RW mtype from this TCP to all TCCs. Sum over TCP instances."></metric>
     <metric name="TCP_TCC_RW_ATOMIC_REQ_sum" expr="sum(TCP_TCC_RW_ATOMIC_REQ,16)" descr="Total atomic requests with RW mtype from this TCP to all TCCs. Sum over TCP instances."></metric>
     <metric name="TCP_PENDING_STALL_CYCLES_sum" expr="sum(TCP_PENDING_STALL_CYCLES,16)" descr="Stall due to data pending from L2. Sum over TCP instances."></metric>
+    <metric name="TCP_TCR_TCP_STALL_CYCLES_PERCENT" expr="100*TCP_TCR_TCP_STALL_CYCLES_sum/TCP_GATE_EN1_sum" descr="Percentage of time TCP is stalled by TCR."></metric>
     <metric name="TCA_CYCLE_sum" expr="sum(TCA_CYCLE,16)" descr="Number of cycles. Sum over all TCA instances "></metric>
     <metric name="TCA_BUSY_sum" expr="sum(TCA_BUSY,16)" descr="Number of cycles we have a request pending. Sum over all TCA instances."></metric>
     <metric name="TCC_BUSY_avr" expr="avr(TCC_BUSY,16)" descr="TCC_BUSY avr over all memory channels."></metric>
@@ -375,12 +378,16 @@
     <metric name="TCC_EA0_RDREQ_32B_sum" expr="sum(TCC_EA0_RDREQ_32B,16)" descr="Number of 32-byte TCC/EA read requests Sum over TCC instances."></metric>
     <metric name="TCC_EA0_RD_UNCACHED_32B_sum" expr="sum(TCC_EA0_RD_UNCACHED_32B,16)" descr="Number of 32-byte TCC/EA read due to uncached traffic. A 64-byte request will be counted as 2 Sum over TCC instances."></metric>
     <metric name="TCC_TAG_STALL_sum" expr="sum(TCC_TAG_STALL,16)" descr="."></metric>
+    <metric name="TCC_TAG_STALL_PERCENT" expr="100*TCC_TAG_STALL_sum/TCC_CYCLE_sum" descr="Percentage of time the TCC tag lookup pipeline is stalled."></metric>
+    <metric name="TCC_BUBBLE_sum" expr=sum(TCC_BUBBLE,16) descr="Number of 128-byte read requests sent to EA. Sum over TCC instances."></metric>
     <metric name="TCC_NORMAL_WRITEBACK_sum" expr="sum(TCC_NORMAL_WRITEBACK,16)" descr="Number of writebacks due to requests that are not writeback requests. Sum over TCC instances."></metric>
     <metric name="TCC_ALL_TC_OP_WB_WRITEBACK_sum" expr="sum(TCC_ALL_TC_OP_WB_WRITEBACK,16)" descr="Number of writebacks due to all TC_OP writeback requests. Sum over TCC instances."></metric>
     <metric name="TCC_NORMAL_EVICT_sum" expr="sum(TCC_NORMAL_EVICT,16)" descr="Number of evictions due to requests that are not invalidate or probe requests. Sum over TCC instances."></metric>
     <metric name="TCC_ALL_TC_OP_INV_EVICT_sum" expr="sum(TCC_ALL_TC_OP_INV_EVICT,16)" descr="Number of evictions due to all TC_OP invalidate requests. Sum over TCC instances."></metric>
     <metric name="TCC_EA0_RDREQ_DRAM_sum" expr="sum(TCC_EA0_RDREQ_DRAM,16)" descr="Number of TCC/EA read requests (either 32-byte or 64-byte) destined for DRAM (MC). Sum over TCC instances."></metric>
     <metric name="TCC_EA0_WRREQ_DRAM_sum" expr="sum(TCC_EA0_WRREQ_DRAM,16)" descr="Number of TCC/EA write requests (either 32-byte of 64-byte) destined for DRAM (MC). Sum over TCC instances."></metric>
+    <metric name="TCC_EA_READBW" expr="(128*TCC_BUBBLE_sum)+64*(TCC_EA0_RDREQ_sum-TCC_BUBBLE_sum-TCC_EA0_RDREQ_32B_sum)+32*TCC_EA0_RDREQ_32B_sum" descr="EA read bandwidth."></metric>
+    <metric name="TCC_EA_READ_LATENCY" expr="TCC_EA0_RDREQ_LEVEL_sum/(TCC_BUBBLE_sum+TCC_EA0_RDREQ_sum)" descr="Lantency of an EA read."></metric>
     <metric name="FETCH_SIZE" expr="(TCC_EA0_RDREQ_32B_sum*32+(TCC_EA0_RDREQ_sum-TCC_EA0_RDREQ_32B_sum)*64)/1024" descr="The total kilobytes fetched from the video memory. This is measured with all extra fetches and any cache or memory effects taken into account."></metric>
     <metric name="WRITE_SIZE" expr="((TCC_EA0_WRREQ_sum-TCC_EA0_WRREQ_64B_sum)*32+TCC_EA0_WRREQ_64B_sum*64)/1024" descr="The total kilobytes written to the video memory. This is measured with all extra fetches and any cache or memory effects taken into account."></metric>
     <metric name="WRITE_REQ_32B" expr="TCC_EA0_WRREQ_64B_sum*2+(TCC_EA0_WRREQ_sum-TCC_EA0_WRREQ_64B_sum)" descr="The total number of 32-byte effective memory writes."></metric>
diff --git a/projects/rocprofiler/src/core/counters/metrics/basic_counters.xml b/projects/rocprofiler/src/core/counters/metrics/basic_counters.xml
index 0cb7ed4b50..7e34a2d0f1 100755
--- a/projects/rocprofiler/src/core/counters/metrics/basic_counters.xml
+++ b/projects/rocprofiler/src/core/counters/metrics/basic_counters.xml
@@ -615,6 +615,7 @@
   <metric name="TCC_EA0_RD_UNCACHED_32B" block=TCC event=40 descr="Number of 32-byte TCC/EA read due to uncached traffic. A 64-byte request will be counted as 2"></metric>
   <metric name="TCC_EA0_RDREQ_LEVEL" block=TCC event=44 descr="The sum of the number of TCC/EA read requests in flight. This is primarily meant for measure average EA read latency. Average read latency = TCC_PERF_SEL_EA_RDREQ_LEVEL/TCC_PERF_SEL_EA_RDREQ."></metric>
   <metric name="TCC_TAG_STALL" block=TCC event=45 descr="Number of cycles the normal request pipeline in the tag was stalled for any reason. Normally, stalls of this nature are measured exactly from one point the pipeline, but that is not the case for this counter. Probes can stall the pipeline at a variety of places, and there is no single point that can reasonably measure the total stalls accurately."></metric>
+  <metric name="TCC_BUBBLE" block="TCC" event=56 descr="Number of 128-byte read requests sent to EA."></metric>
   <metric name="TCC_NORMAL_WRITEBACK" block=TCC event=68 descr="Number of writebacks due to requests that are not writeback requests."></metric>
   <metric name="TCC_ALL_TC_OP_WB_WRITEBACK" block=TCC event=73 descr="Number of writebacks due to all TC_OP writeback requests."></metric>
   <metric name="TCC_NORMAL_EVICT" block=TCC event=74 descr="Number of evictions due to requests that are not invalidate or probe requests."></metric>
diff --git a/projects/rocprofiler/src/core/counters/metrics/derived_counters.xml b/projects/rocprofiler/src/core/counters/metrics/derived_counters.xml
index ce21d93bcf..0ba12b37da 100755
--- a/projects/rocprofiler/src/core/counters/metrics/derived_counters.xml
+++ b/projects/rocprofiler/src/core/counters/metrics/derived_counters.xml
@@ -173,6 +173,7 @@
   <metric name="TCP_TCC_RW_WRITE_REQ_sum"                               expr=sum(TCP_TCC_RW_WRITE_REQ,16) descr="Total write requests with RW mtype from this TCP to all TCCs. Sum over TCP instances."></metric>
   <metric name="TCP_TCC_RW_ATOMIC_REQ_sum"                              expr=sum(TCP_TCC_RW_ATOMIC_REQ,16) descr="Total atomic requests with RW mtype from this TCP to all TCCs. Sum over TCP instances."></metric>
   <metric name="TCP_PENDING_STALL_CYCLES_sum"                           expr=sum(TCP_PENDING_STALL_CYCLES,16) descr="Stall due to data pending from L2. Sum over TCP instances."></metric>
+  <metric name="TCP_TCR_TCP_STALL_CYCLES_PERCENT"                       expr=100*TCP_TCR_TCP_STALL_CYCLES_sum/TCP_GATE_EN1_sum descr="Percentage of time TCP is stalled by TCR."></metric>
   <metric name="TCA_CYCLE_sum"                                          expr=sum(TCA_CYCLE,16)  descr="Number of cycles. Sum over all TCA instances "></metric>
   <metric name="TCA_BUSY_sum"                                           expr=sum(TCA_BUSY,16)   descr="Number of cycles we have a request pending. Sum over all TCA instances."></metric>
   <metric name="TCC_BUSY_avr"                                           expr=avr(TCC_BUSY,32) descr="TCC_BUSY avr over all memory channels."></metric>
@@ -212,6 +213,7 @@
   <metric name="TCC_EA_RDREQ_GMI_CREDIT_STALL_sum"                      expr=sum(TCC_EA_RDREQ_GMI_CREDIT_STALL,32) descr="Number of cycles there was a stall because the read request interface was out of GMI credits. Stalls occur regardless of whether a read needed to be performed or not. Sum over TCC instances."></metric>
   <metric name="TCC_EA_RDREQ_DRAM_CREDIT_STALL_sum"                     expr=sum(TCC_EA_RDREQ_DRAM_CREDIT_STALL,32) descr="Number of cycles there was a stall because the read request interface was out of DRAM credits. Stalls occur regardless of whether a read needed to be performed or not. Sum over TCC instances."></metric>
   <metric name="TCC_TAG_STALL_sum"                                      expr=sum(TCC_TAG_STALL,32) descr="."></metric>
+  <metric name="TCC_TAG_STALL_PERCENT"                                  expr=100*TCC_TAG_STALL_sum/TCC_CYCLE_sum descr="Percentage of time the TCC tag lookup pipeline is stalled."></metric>
   <metric name="TCC_NORMAL_WRITEBACK_sum"                               expr=sum(TCC_NORMAL_WRITEBACK,32) descr="Number of writebacks due to requests that are not writeback requests. Sum over TCC instances."></metric>
   <metric name="TCC_ALL_TC_OP_WB_WRITEBACK_sum"                         expr=sum(TCC_ALL_TC_OP_WB_WRITEBACK,32) descr="Number of writebacks due to all TC_OP writeback requests. Sum over TCC instances."></metric>
   <metric name="TCC_NORMAL_EVICT_sum"                                   expr=sum(TCC_NORMAL_EVICT,32) descr="Number of evictions due to requests that are not invalidate or probe requests. Sum over TCC instances."></metric>
@@ -372,6 +374,7 @@
   <metric name="TCP_TCC_RW_WRITE_REQ_sum"                               expr=sum(TCP_TCC_RW_WRITE_REQ,16) descr="Total write requests with RW mtype from this TCP to all TCCs. Sum over TCP instances."></metric>
   <metric name="TCP_TCC_RW_ATOMIC_REQ_sum"                              expr=sum(TCP_TCC_RW_ATOMIC_REQ,16) descr="Total atomic requests with RW mtype from this TCP to all TCCs. Sum over TCP instances."></metric>
   <metric name="TCP_PENDING_STALL_CYCLES_sum"                           expr=sum(TCP_PENDING_STALL_CYCLES,16) descr="Stall due to data pending from L2. Sum over TCP instances."></metric>
+  <metric name="TCP_TCR_TCP_STALL_CYCLES_PERCENT"                       expr=100*TCP_TCR_TCP_STALL_CYCLES_sum/TCP_GATE_EN1_sum descr="Percentage of time TCP is stalled by TCR."></metric>
   <metric name="TCA_CYCLE_sum"                                          expr=sum(TCA_CYCLE,16)  descr="Number of cycles. Sum over all TCA instances "></metric>
   <metric name="TCA_BUSY_sum"                                           expr=sum(TCA_BUSY,16)   descr="Number of cycles we have a request pending. Sum over all TCA instances."></metric>
   <metric name="TCC_BUSY_avr"                                           expr=avr(TCC_BUSY,16) descr="TCC_BUSY avr over all memory channels."></metric>
@@ -405,12 +408,16 @@
   <metric name="TCC_EA0_RDREQ_32B_sum"                                   expr=sum(TCC_EA0_RDREQ_32B,16) descr="Number of 32-byte TCC/EA read requests Sum over TCC instances."></metric>
   <metric name="TCC_EA0_RD_UNCACHED_32B_sum"                             expr=sum(TCC_EA0_RD_UNCACHED_32B,16) descr="Number of 32-byte TCC/EA read due to uncached traffic. A 64-byte request will be counted as 2 Sum over TCC instances."></metric>
   <metric name="TCC_TAG_STALL_sum"                                      expr=sum(TCC_TAG_STALL,16) descr="."></metric>
+  <metric name="TCC_TAG_STALL_PERCENT"                                  expr=100*TCC_TAG_STALL_sum/TCC_CYCLE_sum descr="Percentage of time the TCC tag lookup pipeline is stalled."></metric>
+  <metric name="TCC_BUBBLE_sum"                                         expr=sum(TCC_BUBBLE,16) descr="Number of 128-byte read requests sent to EA. Sum over TCC instances."></metric>
   <metric name="TCC_NORMAL_WRITEBACK_sum"                               expr=sum(TCC_NORMAL_WRITEBACK,16) descr="Number of writebacks due to requests that are not writeback requests. Sum over TCC instances."></metric>
   <metric name="TCC_ALL_TC_OP_WB_WRITEBACK_sum"                         expr=sum(TCC_ALL_TC_OP_WB_WRITEBACK,16) descr="Number of writebacks due to all TC_OP writeback requests. Sum over TCC instances."></metric>
   <metric name="TCC_NORMAL_EVICT_sum"                                   expr=sum(TCC_NORMAL_EVICT,16) descr="Number of evictions due to requests that are not invalidate or probe requests. Sum over TCC instances."></metric>
   <metric name="TCC_ALL_TC_OP_INV_EVICT_sum"                            expr=sum(TCC_ALL_TC_OP_INV_EVICT,16) descr="Number of evictions due to all TC_OP invalidate requests. Sum over TCC instances."></metric>
   <metric name="TCC_EA0_RDREQ_DRAM_sum"                                  expr=sum(TCC_EA0_RDREQ_DRAM,16) descr="Number of TCC/EA read requests (either 32-byte or 64-byte) destined for DRAM (MC). Sum over TCC instances."></metric>
   <metric name="TCC_EA0_WRREQ_DRAM_sum"                                  expr=sum(TCC_EA0_WRREQ_DRAM,16) descr="Number of TCC/EA write requests (either 32-byte of 64-byte) destined for DRAM (MC). Sum over TCC instances."></metric>
+  <metric name="TCC_EA_READBW"                                          expr=(128*TCC_BUBBLE_sum)+64*(TCC_EA0_RDREQ_sum-TCC_BUBBLE_sum-TCC_EA0_RDREQ_32B_sum)+32*TCC_EA0_RDREQ_32B_sum descr="EA read bandwidth."></metric>
+  <metric name="TCC_EA_READ_LATENCY"                                    expr=TCC_EA0_RDREQ_LEVEL_sum/(TCC_BUBBLE_sum+TCC_EA0_RDREQ_sum) descr="Lantency of an EA read."></metric>
   <metric name="FETCH_SIZE"                                             expr=(TCC_EA0_RDREQ_32B_sum*32+(TCC_EA0_RDREQ_sum-TCC_EA0_RDREQ_32B_sum)*64)/1024 descr="The total kilobytes fetched from the video memory. This is measured with all extra fetches and any cache or memory effects taken into account."></metric>
   <metric name="WRITE_SIZE"                                             expr=((TCC_EA0_WRREQ_sum-TCC_EA0_WRREQ_64B_sum)*32+TCC_EA0_WRREQ_64B_sum*64)/1024 descr="The total kilobytes written to the video memory. This is measured with all extra fetches and any cache or memory effects taken into account."></metric>
   <metric name="WRITE_REQ_32B"                                          expr=TCC_EA0_WRREQ_64B_sum*2+(TCC_EA0_WRREQ_sum-TCC_EA0_WRREQ_64B_sum) descr="The total number of 32-byte effective memory writes."></metric>
diff --git a/projects/rocprofiler/test/tool/gfx_metrics.xml b/projects/rocprofiler/test/tool/gfx_metrics.xml
index 6e9e688eb2..81dc9dde74 100755
--- a/projects/rocprofiler/test/tool/gfx_metrics.xml
+++ b/projects/rocprofiler/test/tool/gfx_metrics.xml
@@ -616,6 +616,7 @@
   <metric name="TCC_EA0_RD_UNCACHED_32B" block=TCC event=40 descr="Number of 32-byte TCC/EA read due to uncached traffic. A 64-byte request will be counted as 2"></metric>
   <metric name="TCC_EA0_RDREQ_LEVEL" block=TCC event=44 descr="The sum of the number of TCC/EA read requests in flight. This is primarily meant for measure average EA read latency. Average read latency = TCC_PERF_SEL_EA_RDREQ_LEVEL/TCC_PERF_SEL_EA_RDREQ."></metric>
   <metric name="TCC_TAG_STALL" block=TCC event=45 descr="Number of cycles the normal request pipeline in the tag was stalled for any reason. Normally, stalls of this nature are measured exactly from one point the pipeline, but that is not the case for this counter. Probes can stall the pipeline at a variety of places, and there is no single point that can reasonably measure the total stalls accurately."></metric>
+  <metric name="TCC_BUBBLE" block="TCC" event=56 descr="Number of 128-byte read requests sent to EA."></metric>
   <metric name="TCC_NORMAL_WRITEBACK" block=TCC event=68 descr="Number of writebacks due to requests that are not writeback requests."></metric>
   <metric name="TCC_ALL_TC_OP_WB_WRITEBACK" block=TCC event=73 descr="Number of writebacks due to all TC_OP writeback requests."></metric>
   <metric name="TCC_NORMAL_EVICT" block=TCC event=74 descr="Number of evictions due to requests that are not invalidate or probe requests."></metric>
diff --git a/projects/rocprofiler/test/tool/metrics.xml b/projects/rocprofiler/test/tool/metrics.xml
index 3750dcd338..fab7522810 100755
--- a/projects/rocprofiler/test/tool/metrics.xml
+++ b/projects/rocprofiler/test/tool/metrics.xml
@@ -166,6 +166,7 @@
   <metric name="TCP_TCC_RW_WRITE_REQ_sum"                               expr=sum(TCP_TCC_RW_WRITE_REQ,16) descr="Total write requests with RW mtype from this TCP to all TCCs. Sum over TCP instances."></metric>
   <metric name="TCP_TCC_RW_ATOMIC_REQ_sum"                              expr=sum(TCP_TCC_RW_ATOMIC_REQ,16) descr="Total atomic requests with RW mtype from this TCP to all TCCs. Sum over TCP instances."></metric>
   <metric name="TCP_PENDING_STALL_CYCLES_sum"                           expr=sum(TCP_PENDING_STALL_CYCLES,16) descr="Stall due to data pending from L2. Sum over TCP instances."></metric>
+  <metric name="TCP_TCR_TCP_STALL_CYCLES_PERCENT"                       expr=100*TCP_TCR_TCP_STALL_CYCLES_sum/TCP_GATE_EN1_sum descr="Percentage of time TCP is stalled by TCR."></metric>
   <metric name="TCA_CYCLE_sum"                                          expr=sum(TCA_CYCLE,16)  descr="Number of cycles. Sum over all TCA instances "></metric>
   <metric name="TCA_BUSY_sum"                                           expr=sum(TCA_BUSY,16)   descr="Number of cycles we have a request pending. Sum over all TCA instances."></metric>
   <metric name="TCC_BUSY_avr"                                           expr=avr(TCC_BUSY,32) descr="TCC_BUSY avr over all memory channels."></metric>
@@ -205,6 +206,7 @@
   <metric name="TCC_EA_RDREQ_GMI_CREDIT_STALL_sum"                      expr=sum(TCC_EA_RDREQ_GMI_CREDIT_STALL,32) descr="Number of cycles there was a stall because the read request interface was out of GMI credits. Stalls occur regardless of whether a read needed to be performed or not. Sum over TCC instances."></metric>
   <metric name="TCC_EA_RDREQ_DRAM_CREDIT_STALL_sum"                     expr=sum(TCC_EA_RDREQ_DRAM_CREDIT_STALL,32) descr="Number of cycles there was a stall because the read request interface was out of DRAM credits. Stalls occur regardless of whether a read needed to be performed or not. Sum over TCC instances."></metric>
   <metric name="TCC_TAG_STALL_sum"                                      expr=sum(TCC_TAG_STALL,32) descr="."></metric>
+  <metric name="TCC_TAG_STALL_PERCENT"                                  expr=100*TCC_TAG_STALL_sum/TCC_CYCLE_sum descr="Percentage of time the TCC tag lookup pipeline is stalled."></metric>
   <metric name="TCC_NORMAL_WRITEBACK_sum"                               expr=sum(TCC_NORMAL_WRITEBACK,32) descr="Number of writebacks due to requests that are not writeback requests. Sum over TCC instances."></metric>
   <metric name="TCC_ALL_TC_OP_WB_WRITEBACK_sum"                         expr=sum(TCC_ALL_TC_OP_WB_WRITEBACK,32) descr="Number of writebacks due to all TC_OP writeback requests. Sum over TCC instances."></metric>
   <metric name="TCC_NORMAL_EVICT_sum"                                   expr=sum(TCC_NORMAL_EVICT,32) descr="Number of evictions due to requests that are not invalidate or probe requests. Sum over TCC instances."></metric>
@@ -292,6 +294,7 @@
   <metric name="TCP_TCC_RW_WRITE_REQ_sum"                               expr=sum(TCP_TCC_RW_WRITE_REQ,16) descr="Total write requests with RW mtype from this TCP to all TCCs. Sum over TCP instances."></metric>
   <metric name="TCP_TCC_RW_ATOMIC_REQ_sum"                              expr=sum(TCP_TCC_RW_ATOMIC_REQ,16) descr="Total atomic requests with RW mtype from this TCP to all TCCs. Sum over TCP instances."></metric>
   <metric name="TCP_PENDING_STALL_CYCLES_sum"                           expr=sum(TCP_PENDING_STALL_CYCLES,16) descr="Stall due to data pending from L2. Sum over TCP instances."></metric>
+  <metric name="TCP_TCR_TCP_STALL_CYCLES_PERCENT"                       expr=100*TCP_TCR_TCP_STALL_CYCLES_sum/TCP_GATE_EN1_sum descr="Percentage of time TCP is stalled by TCR."></metric>
   <metric name="TCA_CYCLE_sum"                                          expr=sum(TCA_CYCLE,16)  descr="Number of cycles. Sum over all TCA instances "></metric>
   <metric name="TCA_BUSY_sum"                                           expr=sum(TCA_BUSY,16)   descr="Number of cycles we have a request pending. Sum over all TCA instances."></metric>
   <metric name="TCC_BUSY_avr"                                           expr=avr(TCC_BUSY,16) descr="TCC_BUSY avr over all memory channels."></metric>
@@ -325,12 +328,16 @@
   <metric name="TCC_EA0_RDREQ_32B_sum"                                   expr=sum(TCC_EA0_RDREQ_32B,16) descr="Number of 32-byte TCC/EA read requests Sum over TCC instances."></metric>
   <metric name="TCC_EA0_RD_UNCACHED_32B_sum"                             expr=sum(TCC_EA0_RD_UNCACHED_32B,16) descr="Number of 32-byte TCC/EA read due to uncached traffic. A 64-byte request will be counted as 2 Sum over TCC instances."></metric>
   <metric name="TCC_TAG_STALL_sum"                                      expr=sum(TCC_TAG_STALL,16) descr="."></metric>
+  <metric name="TCC_TAG_STALL_PERCENT"                                  expr=100*TCC_TAG_STALL_sum/TCC_CYCLE_sum descr="Percentage of time the TCC tag lookup pipeline is stalled."></metric>
+  <metric name="TCC_BUBBLE_sum"                                         expr=sum(TCC_BUBBLE,16) descr="Number of 128-byte read requests sent to EA. Sum over TCC instances."></metric>
   <metric name="TCC_NORMAL_WRITEBACK_sum"                               expr=sum(TCC_NORMAL_WRITEBACK,16) descr="Number of writebacks due to requests that are not writeback requests. Sum over TCC instances."></metric>
   <metric name="TCC_ALL_TC_OP_WB_WRITEBACK_sum"                         expr=sum(TCC_ALL_TC_OP_WB_WRITEBACK,16) descr="Number of writebacks due to all TC_OP writeback requests. Sum over TCC instances."></metric>
   <metric name="TCC_NORMAL_EVICT_sum"                                   expr=sum(TCC_NORMAL_EVICT,16) descr="Number of evictions due to requests that are not invalidate or probe requests. Sum over TCC instances."></metric>
   <metric name="TCC_ALL_TC_OP_INV_EVICT_sum"                            expr=sum(TCC_ALL_TC_OP_INV_EVICT,16) descr="Number of evictions due to all TC_OP invalidate requests. Sum over TCC instances."></metric>
   <metric name="TCC_EA0_RDREQ_DRAM_sum"                                  expr=sum(TCC_EA0_RDREQ_DRAM,16) descr="Number of TCC/EA read requests (either 32-byte or 64-byte) destined for DRAM (MC). Sum over TCC instances."></metric>
   <metric name="TCC_EA0_WRREQ_DRAM_sum"                                  expr=sum(TCC_EA0_WRREQ_DRAM,16) descr="Number of TCC/EA write requests (either 32-byte of 64-byte) destined for DRAM (MC). Sum over TCC instances."></metric>
+  <metric name="TCC_EA_READBW"                                          expr=(128*TCC_BUBBLE_sum)+64*(TCC_EA0_RDREQ_sum-TCC_BUBBLE_sum-TCC_EA0_RDREQ_32B_sum)+32*TCC_EA0_RDREQ_32B_sum descr="EA read bandwidth."></metric>
+  <metric name="TCC_EA_READ_LATENCY"                                    expr=TCC_EA0_RDREQ_LEVEL_sum/(TCC_BUBBLE_sum+TCC_EA0_RDREQ_sum) descr="Lantency of an EA read."></metric>  
   <metric name="FETCH_SIZE"                                             expr=(TCC_EA0_RDREQ_32B_sum*32+(TCC_EA0_RDREQ_sum-TCC_EA0_RDREQ_32B_sum)*64)/1024 descr="The total kilobytes fetched from the video memory. This is measured with all extra fetches and any cache or memory effects taken into account."></metric>
   <metric name="WRITE_SIZE"                                             expr=((TCC_EA0_WRREQ_sum-TCC_EA0_WRREQ_64B_sum)*32+TCC_EA0_WRREQ_64B_sum*64)/1024 descr="The total kilobytes written to the video memory. This is measured with all extra fetches and any cache or memory effects taken into account."></metric>
   <metric name="WRITE_REQ_32B"                                          expr=TCC_EA0_WRREQ_64B_sum*2+(TCC_EA0_WRREQ_sum-TCC_EA0_WRREQ_64B_sum) descr="The total number of 32-byte effective memory writes."></metric>