Intel® VTune™ Amplifier
Similar to the Summary window, available in GUI, the summary report provides overall performance data of your target. Intel® VTune™ Amplifier automatically generates the summary report when data collection completes. To disable this report, use the no-summary option in your command when performing a collect or collect-with action.
Use the following syntax to generate the Summary report from a preexisting result:
$ amplxe-cl -report summary -result-dir <result_path>
Example 1: Basic Hotspots Summary
The following example generates the summary report for the r000hs Basic Hotspots analysis result. For User-Mode Sampling and Tracing Collection results, the summary report includes Collection and Platform information, CPU information and summary per the basic metrics.
$ amplxe-cl -report summary -r r000hs
Collection and Platform Info
----------------------------
Parameter r000hs
------------------------ -------------------------------------------------------------
Application Command Line /home/tachyon/find_hotspots
Operating System Ubuntu 11.04
Computer Name My Computer
Result Size 2926817
Collection start time 11:17:06 13/10/2015 UTC
Collection stop time 11:17:20 13/10/2015 UTC
CPU
---
Parameter r000hs
---------------------- -------------------------------------------------
Name 4th generation Intel® Core™ Processor family
Frequency 2494226458
Logical CPU Core Count 4
Summary
-------
Elapsed Time: 14.094
CPU Time: 11.319
Average CPU Usage: 0.749
amplxe: Executing actions 100 % doneExample 2: Advanced Hotspots Summary
This example generates the summary report for the r001ah Advanced Hotspots analysis result. For Hardware Event-based Sampling Collection results, the summary report includes Collection and Platform information, CPU information, summary per the basic metrics, and an event summary.
$ amplxe-cl -report summary -r r001ah
Collection and Platform Info
----------------------------
Parameter r001ah
------------------------ --------------------------------------------------------------------------------------------
Application Command Line /home/tachyon/find_hotspots
Operating System Ubuntu 11.04
Computer Name My Computer
Result Size 37188680
Collection start time 09:59:01 11/09/2015 UTC
Collection stop time 09:59:28 11/09/2015 UTC
CPU
---
Parameter 001ah
---------------------- -------------------------------------------------
Name 4th generation Intel® Core™ Processor family
Frequency 2494232562
Logical CPU Core Count 4
Summary
-------
Elapsed Time: 26.785
CPU Time: 16.394
Average CPU Usage: 0.610
CPI Rate: 0.413
Event summary
-------------
Hardware Event Type Hardware Event Count:Self Hardware Event Sample Count:Self Events Per Sample
------------------------ ------------------------- -------------------------------- -----------------
INST_RETIRED.ANY 110633200000 58228 1900000
CPU_CLK_UNHALTED.THREAD 45653200000 24028 1900000
CPU_CLK_UNHALTED.REF_TSC 40889900000 21521 1900000Use the Elapsed Time metric as your performance baseline to estimate your optimizations. The CPU Usage metric is the total CPU time divided by the Elapsed time, which demonstrates an average value of CPU utilization. For example, 'Average CPU Usage: 5.907' means near 6 cores were running on the average for your program overall. If you have, for example, 8 core system, there is some potential to parallelize the code.
Example 3: Concurrency Summary
This example generates the summary report for a concurrency analysis result r002cc. The CPU portion of the report shows the number of logical CPU cores, which is the target concurrency level. The summary portion of the report shows that elapsed time for the target is 14.077s and the average concurrency is 1.120, which means that a single thread was running during most of the elapsed time. The Average CPU Usage metric is demonstrates that only 0.743 cores were running on the average during the analysis.
$ amplxe-cl -report summary -r r002cc
CPU
---
Parameter r002cc
--------------------- -------------------------------------------------
Name 4th generation Intel® Core™ Processor family
Frequency 2494227397
Logical CPU Core Count 4
Summary
-------
Average Concurrency: 1.120
Elapsed Time: 14.077
CPU Time: 11.121
Wait Time: 64.556
Average CPU Usage: 0.743Example 4: Locks and Waits Summary
This example generates a summary report for the Locks and Waits analysis result r003lw. The summary portion of the report shows that the multithreaded target spent 64 seconds waiting, with an average concurrency of only 1.073. To identify the cause of the wait, view the result in the GUI performance pane, or generate a performance report.
$ amplxe-cl -report summary -r r003lw
Summary
-------
Average Concurrency: 1.073
Elapsed Time: 13.911
CPU Time: 11.031
Wait Time: 64.468
Average CPU Usage: 0.768Example 5: General Exploration Metrics Summary
The General Exploration analysis metrics provide a high-level overview of hardware performance problems. The General Exploration Metrics section appears in a summary report for Hardware Event-based Sampling Collection results. The set of metrics displayed in the summary depends on the profiled CPU type and list of events.
Information is displayed by execution categories to make it easy to identify which part of the pipeline to focus on. In this example, the Clockticks per Instructions Retired (CPI) is over 1, indicating there is a problem with application performance. In the Bad Speculation category, Branch Mispredict indicates that a significant proportion of branches are mispredicted, leading to excessive wasted work or Backend stalls due to the machine need to recover its state from a speculative path. To get more information about detected hardware issues, view the result in the GUI Summary pane.
$ amplxe-cl -report summary -r r001ge
General Exploration Metrics
---------------------------
Parameter r001ge
------------------------------------ ---------------
Clockticks 28772002814.000
Instructions Retired 25182681309
CPI Rate 1.143
MUX Reliability 0.876
Front-End Bound 0.083
Front-End Latency 0.100
ICache Misses 0.038
ITLB Overhead 0.005
Branch Resteers 0.156
DSB Switches 0.056
Length Changing Prefixes 0.0
MS Switches 0.086
Front-End Bandwidth 0.0
Front-End Bandwidth DSB 0.089
Front-End Bandwidth MITE 0.053
Front-End Bandwidth LSD 0.001
Bad Speculation 0.324
Branch Mispredict 0.324
Machine Clears 0.0
Back-End Bound 0.451
Memory Bound 0.345
L1 Bound 0.204
DTLB Overhead 0.324
Loads Blocked by Store Forwarding 0.006
Lock Latency 0.0
Split Loads 0.001
4K Aliasing 0.013
FB Full 0.283
L3 Bound 0.0
Contested Accesses 0.000
Data Sharing 0.0
LLC Hit 0.116
SQ Full 0.009
DRAM Bound 0.0
Memory Latency 0.0
LLC Miss 0.055
Store Bound 0.0
Store Latency 0.000
False Sharing 0.0
Split Stores 0.012
DTLB Store Overhead 0.019
Core Bound 0.106
Divider 0.277
Port Utilization 0.291
Cycles of 0 Ports Utilized 0.276
Cycles of 1 Port Utilized 0.129
Cycles of 2 Ports Utilized 0.138
Cycles of 3+ Ports Utilized 0.395
Port 0 0.144
Port 1 0.182
Port 2 0.243
Port 3 0.284
Port 4 0.170
Port 5 0.119
Port 6 0.169
Port 7 0.134
Retiring 0.141
General Retirement 0.132
Microcode Sequencer 0.009
Assists 0.0