Analyzers
Talk to fellow users of Intel Analyzer tools (Intel VTune™ Profiler, Intel Advisor)
5056 Discussions

Why "Contested Accesses" metric in microarchitecture exploration is N/A?

g199209
Beginner
1,356 Views

g199209_0-1681192968389.png

 

Sampling interval to long? I use 5ms now.

0 Kudos
10 Replies
g199209
Beginner
1,350 Views

And in this report, L1 Bound is 52.4%, but all the children of it is 0.0%, what does this means?

0 Kudos
DiyaN_Intel
Moderator
1,326 Views

Hi, 

 

Thank you for posting in Intel Communities.

Could you please share the following details- 

 

1. Sample reproducer code.

2. VTune version.

3. Exact steps and the commands used.

4. OS details.

5. Processor details.

6. Self-checker logs.

 

To get the self-checker logs you have to make sure that the VTune self-checker script is running fine. This is to validate that appropriate drivers are installed and the system is set up properly to collect the performance data. 

The "vtune-self-checker.sh" script is available from <vtune-install-dir>/bin64 on the Windows or Linux system on which you installed VTune Profiler.

On running this script, it will create a log file and display the location of the log file.

Kindly share that file with us so that we can verify it from our side.

 

Thanks and Regards,

Diya

 

0 Kudos
g199209
Beginner
1,317 Views

1. 

```

#include <iostream>
#include <cstdint>
#include <mutex>
#include <thread>
#include <vector>
#include <atomic>

 

struct {
volatile uint64_t sum0;
volatile uint64_t sum1;
volatile uint64_t sum2;
volatile uint64_t sum3;
}__attribute__((aligned(64))) sum;

constexpr uint64_t N = 100000000000ULL;

void threadSum(uint64_t start, uint64_t end, volatile uint64_t* sum) {
for (uint64_t i = start; i < end; ++i) {
*sum = *sum + i;
}
}

int main() {
std::vector<std::thread> threads;
constexpr uint64_t STEP = N / 4;
uint64_t start = 0;
volatile uint64_t* pSum[] = {&sum.sum0, &sum.sum1, &sum.sum2, &sum.sum3};
for (int i = 0; i < 4; ++i) {
threads.emplace_back(threadSum, start, start + STEP, pSum[i]);
start += STEP;
}

for (auto &t : threads) {
t.join();
}

std::cout << sum.sum0 + sum.sum1 + sum.sum2 + sum.sum3 << std::endl;
return 0;
}

```

 

```

g++ --std=c++17 -O3 -Wall -g -o sum -lpthread sum.cpp

```

 

gcc version 8.3.1 20190311 (Red Hat 8.3.1-3) (GCC)

 

2.

VTune 2023.0.0.624757

 

3. 

g199209_0-1681264082555.png

 

4.

Local: macOS Ventura 13.2.1

Remote: Linux 3.10.0

 

5.

Intel(R) Core(TM) i9-10980XE

 

6. 

There's no `vtune-self-checker.sh` on remote linux.

0 Kudos
DiyaN_Intel
Moderator
1,279 Views

Hi, 

 

Thank you for providing the details.

As we can see you are using Vtune 2023.0.0, can you please update it to Vtune 2023.1.0 and check once? 

 

If the issue still persists, please follow the below steps.

We assume that you need the GUI experience so can you please try the port forwarding steps given below:

 

Port Forwarding in Servers:

If you have a web browser on your host machine, you can follow the below process.

Note: VPN must be disconnected while doing port forwarding in servers

 

Step 1: To connect to a remote server, we use "ssh user_name@<IP_address of that server>"

Since we need to port forward from our local host to the remote server, we use the below command.

 

ssh -L localhost:<port_number>:localhost:<port_number> <server_username@server_ipaddress>
Eg: ssh -L localhost:8080:localhost:8080 sdp@10.165.154.190

 

Step 2: Source the variables.

 

 source /opt/intel/oneapi/setvars.sh

 

Step 3: Let us try to open the GUI of Vtune

 

vtune-backend --web-port 8080 --allow-remote-access --enable-server-profiling

 

Step 4: After running the above command, you can see two links are generated. Copy any one of the GUI links (Note: do not press ctrl+C) and open it in any of your browsers.

DiyaN_Intel_0-1681373470364.png

Step 5: Click on “Advanced” and proceed and set the passphrase. 

  The passphrase should contain:

   *  A special character

   *  Any capital alphabet

   * The length of the passphrase should be minimum of 8 characters.

 

DiyaN_Intel_3-1681373680247.png

DiyaN_Intel_4-1681373695417.png

 

Vtune can now be accessed from a browser and you can do the respective analysis. 

NoteIntel® VTune Profiler for macOS is now deprecated and will be discontinued in a future release, so you might not be getting the desired results, try this workaround, hope it helps you get the desired results.

 

If this resolves your issue, make sure to accept this as a solution. This would help others with similar issues.

 

Thanks and Regards, 

Diya

 

0 Kudos
g199209
Beginner
1,248 Views

@DiyaN_Intel wrote:

NoteIntel® VTune Profiler for macOS is now deprecated and will be discontinued in a future release


What the bad news!

0 Kudos
g199209
Beginner
1,232 Views

I use the latest Vtune 2023.1.0 Linux version, the result is almost the same.

 

g199209_0-1681456440190.png

 

Self-checker result:

 

Intel(R) VTune(TM) Profiler Self Check Utility
Copyright (C) 2009 Intel Corporation. All rights reserved.
Build Number: 625246

HW event-based analysis (counting mode) (Intel driver)
Example of analysis types: Performance Snapshot
Collection: Ok
Finalization: Ok...
Report: Ok

Instrumentation based analysis check
Example of analysis types: Hotspots and Threading with user-mode sampling
Collection: Ok
Finalization: Ok...
Report: Ok

HW event-based analysis check (Intel driver)
Example of analysis types: Hotspots with HW event-based sampling, HPC Performance Characterization, etc.
Collection: Ok
vtune: Warning: To enable hardware event-based sampling, VTune Profiler has disabled the NMI watchdog timer. The watchdog timer will be re-enabled after collection completes.
Finalization: Ok...
vtune: Warning: Cannot locate debugging information for the Linux kernel. Source-level analysis will not be possible. Function-level analysis will be limited to kernel symbol tables. See the Enabling Linux Kernel Analysis topic in the product online help for instructions.

Report: Fail

HW event-based analysis check (Intel driver)
Example of analysis types: Microarchitecture Exploration
Collection: Ok
vtune: Warning: To enable hardware event-based sampling, VTune Profiler has disabled the NMI watchdog timer. The watchdog timer will be re-enabled after collection completes.
Finalization: Ok...
vtune: Warning: Cannot locate debugging information for the Linux kernel. Source-level analysis will not be possible. Function-level analysis will be limited to kernel symbol tables. See the Enabling Linux Kernel Analysis topic in the product online help for instructions.

Report: Fail

HW event-based analysis with uncore events (Intel driver)
Example of analysis types: Memory Access
Collection: Ok
vtune: Warning: To enable hardware event-based sampling, VTune Profiler has disabled the NMI watchdog timer. The watchdog timer will be re-enabled after collection completes.
Finalization: Ok...
vtune: Warning: Cannot locate debugging information for the Linux kernel. Source-level analysis will not be possible. Function-level analysis will be limited to kernel symbol tables. See the Enabling Linux Kernel Analysis topic in the product online help for instructions.

Report: Fail

HW event-based analysis with stacks (Perf)
Example of analysis types: Hotspots with HW event-based sampling and call stacks
Collection: Ok
vtune: Warning: To profile kernel modules during the session, make sure they are available in the /lib/modules/kernel_version/ location.
Finalization: Ok...
vtune: Warning: Cannot locate debugging information for the Linux kernel. Source-level analysis will not be possible. Function-level analysis will be limited to kernel symbol tables. See the Enabling Linux Kernel Analysis topic in the product online help for instructions.

Report: Ok

HW event-based analysis with context switches (Intel driver)
Example of analysis types: Threading with HW event-based sampling
Collection: Ok
vtune: Warning: To enable hardware event-based sampling, VTune Profiler has disabled the NMI watchdog timer. The watchdog timer will be re-enabled after collection completes.
Finalization: Ok...
vtune: Warning: Cannot locate debugging information for the Linux kernel. Source-level analysis will not be possible. Function-level analysis will be limited to kernel symbol tables. See the Enabling Linux Kernel Analysis topic in the product online help for instructions.

Report: Ok

Checking DPC++ application as prerequisite for GPU analyses: Fail
Unable to run DPC++ application on GPU connected to this system. If you are using an Intel GPU and want to verify profiling support for DPC++ applications, check these requirements:
* Install Intel(R) GPU driver.
* Install Intel(R) Level Zero GPU runtime.
* Install Intel(R) oneAPI DPC++ Runtime and set the environment.

The check observed a product failure on your system.
Review errors in the output above to fix a problem or contact Intel technical support.

The system is ready for the following analyses:
* Performance Snapshot
* Hotspots and Threading with user-mode sampling
* Hotspots with HW event-based sampling and call stacks
* Threading with HW event-based sampling

The following analyses have failed on the system:
* Hotspots with HW event-based sampling, HPC Performance Characterization, etc.
* Microarchitecture Exploration
* Memory Access
* GPU Compute/Media Hotspots (characterization mode)
* GPU Compute/Media Hotspots (source analysis mode)

0 Kudos
DiyaN_Intel
Moderator
1,222 Views

Hi, 


Thank you for providing all the details. 


We are working on this internally and will get back to you with an update soon.


Thanks and regards,

Diya


0 Kudos
DiyaN_Intel
Moderator
1,158 Views

Hi,


Can you please share the result directory that you get while running the microarchitecture exploration? 


This will help us to investigate further on the issue.


Thanks and Regards,

Diya





0 Kudos
DiyaN_Intel
Moderator
1,090 Views

Hi,


We have not heard back from you.

Can you please share the result directory that you get while running the microarchitecture exploration? 

This will help us to investigate further the issue.

Unless and until you share the result directory, we won't be able to do anything on our part.

If you face any issues please get back to us.


Thanks and Regards,

Diya



0 Kudos
DiyaN_Intel
Moderator
989 Views

Hi, 


We have not heard back from you. This thread will no longer be monitored by Intel. If you need further assistance, please post a new question.


Thanks and Regards, 

Diya



0 Kudos
Reply