topic Re: VTune's self-check fails on GPU Compute/Media Hotspots in Analyzers

VTune's self-check fails on GPU Compute/Media Hotspots

Hrishi2407 — Sun, 16 Mar 2025 16:00:45 GMT

Hi,

I'm using

OS: Ubuntu 24.10 VTune: 2025.0.1 CPU: Intel(R) Core(TM) Ultra 7 GPU: Intel(R) Arc(TM) Graphics

After running "vtune-self-checker.sh"

Intel(R) VTune(TM) Profiler Self Check Utility Copyright (C) 2009 Intel Corporation. All rights reserved. Build Number: 629235 HW event-based analysis (counting mode) (Perf) Example of analysis types: Performance Snapshot Collection: Ok vtune: Error: Cannot stop collection of GPU events 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 (Perf) Example of analysis types: Hotspots with HW event-based sampling, HPC Performance Characterization, etc. Collection: Ok vtune: Warning: To profile kernel modules during the session, make sure they are available in the /lib/modules/kernel_version/ location. vtune: Warning: Precise events are not available on this platform in the driverless mode and will be collected as non-precise. Some metrics may be less accurately attributed to the program code. Please consider loading VTune Profiler driver. Updating Linux kernel to newer version may also help. 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 check (Perf) Example of analysis types: Microarchitecture Exploration Collection: Ok vtune: Warning: To profile kernel modules during the session, make sure they are available in the /lib/modules/kernel_version/ location. vtune: Warning: Precise events are not available on this platform in the driverless mode and will be collected as non-precise. Some metrics may be less accurately attributed to the program code. Please consider loading VTune Profiler driver. Updating Linux kernel to newer version may also help. 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 uncore events (Perf) Example of analysis types: Memory Access Collection: Ok vtune: Warning: To profile kernel modules during the session, make sure they are available in the /lib/modules/kernel_version/ location. vtune: Warning: Precise events are not available on this platform in the driverless mode and will be collected as non-precise. Some metrics may be less accurately attributed to the program code. Please consider loading VTune Profiler driver. Updating Linux kernel to newer version may also help. 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 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. vtune: Warning: Precise events are not available on this platform in the driverless mode and will be collected as non-precise. Some metrics may be less accurately attributed to the program code. Please consider loading VTune Profiler driver. Updating Linux kernel to newer version may also help. 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 (Perf) Example of analysis types: Threading with HW event-based sampling Collection: Ok vtune: Warning: To profile kernel modules during the session, make sure they are available in the /lib/modules/kernel_version/ location. vtune: Warning: Precise events are not available on this platform in the driverless mode and will be collected as non-precise. Some metrics may be less accurately attributed to the program code. Please consider loading VTune Profiler driver. Updating Linux kernel to newer version may also help. 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 system is ready to be used for performance analysis with Intel VTune Profiler. Review warnings in the output above to find product limitations, if any. The system is ready for the following analyses: * Performance Snapshot * Hotspots and Threading with user-mode sampling * Hotspots with HW event-based sampling, HPC Performance Characterization, etc. * Microarchitecture Exploration * Memory Access * Hotspots with HW event-based sampling and call stacks * Threading with HW event-based sampling The following analyses have failed on the system: * GPU Compute/Media Hotspots (characterization mode) * GPU Compute/Media Hotspots (source analysis mode) Log location: /tmp/vtune-tmp-user/self-checker-2025.03.07_05.36.40/log.txt

The drivers are loaded properly and SYCL programs run on the GPU, but VTune's self-check fails on GPU Compute/Media Hotspots. Any solutions to this?

Re: VTune's self-check fails on GPU Compute/Media Hotspots

yuzhang3_intel — Mon, 17 Mar 2025 05:39:06 GMT

Can you check the log file and see if the sycl8 library is missing? If yes, it means you need to install the Intel oneAPI DPC++/C++ compiler because the current self-check utility depends on the library located in the compiler.

/tmp/vtune-tmp-user/self-checker-2025.03.07_05.36.40/log.txt

When the sycl environment is ready, you need to install the GPU driver and compute runtime:

https://dgpu-docs.intel.com/driver/installation.html

https://github.com/intel/compute-runtime/releases

Re: VTune's self-check fails on GPU Compute/Media Hotspots

Hrishi2407 — Mon, 17 Mar 2025 09:40:56 GMT

I previously mentioned that the drivers were loaded properly, but after running "sudo ./insmod-sep -q",I realized that some required VTune drivers are not loaded. Only pax and socperf3 are present while sep5, vtssp and socwatch2_15 are not loaded.

pax driver is loaded and owned by group "user" with file permissions "660". socperf3 driver is loaded and owned by group "user" with file permissions "660". sep5 driver is not loaded. socwatch2_15 driver is not loaded. vtsspp driver is not loaded.

I tried running "sudo ./build-driver"

C compiler to use: [ /bin/gcc ] C compiler version: 14.2.0 Make command to use: [ /bin/make ] Make version: 4.3 Kernel source directory: [ /lib/modules/6.11.0-19-generic/build ] Kernel version: 6.11.0-19-generic Cleaning workspaces ... Done Building socperf driver ... warning: the compiler differs from the one used to build the kernel The kernel was built by: x86_64-linux-gnu-gcc-14 (Ubuntu 14.2.0-4ubuntu2) 14.2.0 You are using: gcc (Ubuntu 14.2.0-4ubuntu2) 14.2.0 Skipping BTF generation for /opt/intel/oneapi/vtune/2025.0/sepdk/src/socperf/src/socperf3.ko due to unavailability of vmlinux Done Building sep driver ... warning: the compiler differs from the one used to build the kernel The kernel was built by: x86_64-linux-gnu-gcc-14 (Ubuntu 14.2.0-4ubuntu2) 14.2.0 You are using: gcc (Ubuntu 14.2.0-4ubuntu2) 14.2.0 /opt/intel/oneapi/vtune/2025.0/sepdk/src/sys64.o: warning: objtool: .text+0x35: unreachable instruction /opt/intel/oneapi/vtune/2025.0/sepdk/src/sys64.o: warning: objtool: .text+0x36: unreachable instruction /opt/intel/oneapi/vtune/2025.0/sepdk/src/sys64.o: warning: objtool: .text+0x37: unreachable instruction /opt/intel/oneapi/vtune/2025.0/sepdk/src/sys64.o: warning: objtool: .text+0x3b: unreachable instruction /opt/intel/oneapi/vtune/2025.0/sepdk/src/sys64.o: warning: objtool: .text+0x40: unreachable instruction /opt/intel/oneapi/vtune/2025.0/sepdk/src/sys64.o: warning: objtool: .text+0x45: unreachable instruction /opt/intel/oneapi/vtune/2025.0/sepdk/src/sys64.o: warning: objtool: .text+0x4a: unreachable instruction /opt/intel/oneapi/vtune/2025.0/sepdk/src/sys64.o: warning: objtool: .text+0x4f: unreachable instruction /opt/intel/oneapi/vtune/2025.0/sepdk/src/sys64.o: warning: objtool: .text+0x54: unreachable instruction /opt/intel/oneapi/vtune/2025.0/sepdk/src/sys64.o: warning: objtool: .text+0x59: unreachable instruction /opt/intel/oneapi/vtune/2025.0/sepdk/src/sys64.o: warning: objtool: .text+0x5e: unreachable instruction /opt/intel/oneapi/vtune/2025.0/sepdk/src/sys64.o: warning: objtool: .text+0x63: unreachable instruction /opt/intel/oneapi/vtune/2025.0/sepdk/src/sys64.o: warning: objtool: .text+0x68: unreachable instruction /opt/intel/oneapi/vtune/2025.0/sepdk/src/sys64.o: warning: objtool: .text+0x6c: unreachable instruction /opt/intel/oneapi/vtune/2025.0/sepdk/src/sys64.o: warning: objtool: .text+0x71: unreachable instruction /opt/intel/oneapi/vtune/2025.0/sepdk/src/sys64.o: warning: objtool: .text+0x76: unreachable instruction /opt/intel/oneapi/vtune/2025.0/sepdk/src/sys64.o: warning: objtool: .text+0x7b: unreachable instruction /opt/intel/oneapi/vtune/2025.0/sepdk/src/sys64.o: warning: objtool: .text+0x80: unreachable instruction /opt/intel/oneapi/vtune/2025.0/sepdk/src/sys64.o: warning: objtool: .text+0x85: unreachable instruction /opt/intel/oneapi/vtune/2025.0/sepdk/src/sys64.o: warning: objtool: .text+0x8a: unreachable instruction /opt/intel/oneapi/vtune/2025.0/sepdk/src/sys64.o: warning: objtool: .text+0x8c: unreachable instruction /opt/intel/oneapi/vtune/2025.0/sepdk/src/sys64.o: warning: objtool: .text+0x8e: unreachable instruction /opt/intel/oneapi/vtune/2025.0/sepdk/src/sys64.o: warning: objtool: .text+0x90: unreachable instruction /opt/intel/oneapi/vtune/2025.0/sepdk/src/sys64.o: warning: objtool: .text+0x92: unreachable instruction /opt/intel/oneapi/vtune/2025.0/sepdk/src/sys64.o: warning: objtool: .text+0x95: unreachable instruction /opt/intel/oneapi/vtune/2025.0/sepdk/src/sys64.o: warning: objtool: .text+0x9a: unreachable instruction /opt/intel/oneapi/vtune/2025.0/sepdk/src/sys64.o: warning: objtool: .text+0x9d: unreachable instruction /opt/intel/oneapi/vtune/2025.0/sepdk/src/sys64.o: warning: objtool: .text+0xa2: unreachable instruction /opt/intel/oneapi/vtune/2025.0/sepdk/src/sys64.o: warning: objtool: .text+0xa3: unreachable instruction /opt/intel/oneapi/vtune/2025.0/sepdk/src/sys64.o: warning: objtool: .text+0xa5: unreachable instruction /opt/intel/oneapi/vtune/2025.0/sepdk/src/sys64.o: warning: objtool: .text+0xa7: unreachable instruction /opt/intel/oneapi/vtune/2025.0/sepdk/src/sys64.o: warning: objtool: .text+0xaa: unreachable instruction /opt/intel/oneapi/vtune/2025.0/sepdk/src/sys64.o: warning: objtool: .text+0xae: unreachable instruction /opt/intel/oneapi/vtune/2025.0/sepdk/src/sys64.o: warning: objtool: .text+0xb3: unreachable instruction /opt/intel/oneapi/vtune/2025.0/sepdk/src/sys64.o: warning: objtool: .text+0xb8: unreachable instruction /opt/intel/oneapi/vtune/2025.0/sepdk/src/sys64.o: warning: objtool: .text+0xbd: unreachable instruction /opt/intel/oneapi/vtune/2025.0/sepdk/src/sys64.o: warning: objtool: .text+0xc2: unreachable instruction /opt/intel/oneapi/vtune/2025.0/sepdk/src/sys64.o: warning: objtool: .text+0xc7: unreachable instruction /opt/intel/oneapi/vtune/2025.0/sepdk/src/sys64.o: warning: objtool: .text+0xcc: unreachable instruction /opt/intel/oneapi/vtune/2025.0/sepdk/src/sys64.o: warning: objtool: .text+0xd1: unreachable instruction /opt/intel/oneapi/vtune/2025.0/sepdk/src/sys64.o: warning: objtool: .text+0xd6: unreachable instruction /opt/intel/oneapi/vtune/2025.0/sepdk/src/sys64.o: warning: objtool: .text+0xdb: unreachable instruction /opt/intel/oneapi/vtune/2025.0/sepdk/src/sys64.o: warning: objtool: .text+0xe0: unreachable instruction /opt/intel/oneapi/vtune/2025.0/sepdk/src/sys64.o: warning: objtool: .text+0xe5: unreachable instruction /opt/intel/oneapi/vtune/2025.0/sepdk/src/sys64.o: warning: objtool: .text+0xea: unreachable instruction /opt/intel/oneapi/vtune/2025.0/sepdk/src/sys64.o: warning: objtool: .text+0xef: unreachable instruction /opt/intel/oneapi/vtune/2025.0/sepdk/src/sys64.o: warning: objtool: .text+0xf4: unreachable instruction /opt/intel/oneapi/vtune/2025.0/sepdk/src/sys64.o: warning: objtool: .text+0xf8: unreachable instruction /opt/intel/oneapi/vtune/2025.0/sepdk/src/sys64.o: warning: objtool: .text+0xf9: unreachable instruction Skipping BTF generation for /opt/intel/oneapi/vtune/2025.0/sepdk/src/sep5.ko due to unavailability of vmlinux Done Building pax driver ... warning: the compiler differs from the one used to build the kernel The kernel was built by: x86_64-linux-gnu-gcc-14 (Ubuntu 14.2.0-4ubuntu2) 14.2.0 You are using: gcc (Ubuntu 14.2.0-4ubuntu2) 14.2.0 Skipping BTF generation for /opt/intel/oneapi/vtune/2025.0/sepdk/src/pax/pax.ko due to unavailability of vmlinux Done Building vtsspp driver ... make[1]: Entering directory '/opt/intel/oneapi/vtune/2025.0/sepdk/src/vtsspp' warning: the compiler differs from the one used to build the kernel The kernel was built by: x86_64-linux-gnu-gcc-14 (Ubuntu 14.2.0-4ubuntu2) 14.2.0 You are using: gcc (Ubuntu 14.2.0-4ubuntu2) 14.2.0 Skipping BTF generation for /opt/intel/oneapi/vtune/2025.0/sepdk/src/vtsspp/vtsspp.ko due to unavailability of vmlinux make[1]: Leaving directory '/opt/intel/oneapi/vtune/2025.0/sepdk/src/vtsspp' Done Building socwatch driver ... Building drivers for Linux target /lib/modules/6.11.0-19-generic/build will be used as the kernel build directory Using alternate compiler '/bin/gcc' Removing previously built driver and scripts from /opt/intel/oneapi/vtune/2025.0/sepdk/src/socwatch/drivers mkdir: cannot create directory ‘/opt/intel/oneapi/vtune/2025.0/sepdk/src/socwatch/drivers’: File exists '-c /bin/gcc -l -k /lib/modules/6.11.0-19-generic/build' options will be used to build the SoCWatch driver -c /bin/gcc -l -k /lib/modules/6.11.0-19-generic/build ************ Building socwatch2_15-x32_64-6.11.0-19-genericsmp.ko driver ************ Using C compiler = /bin/gcc Using kernel build dir = /lib/modules/6.11.0-19-generic/build No module symvers file found Using common inc dir = /opt/intel/oneapi/vtune/2025.0/sepdk/src/socwatch/socwatch_driver/../common/include "/lib/modules/6.11.0-19-generic/build/include/linux/cpufreq.h" exists! "/lib/modules/6.11.0-19-generic/build/include/linux/device/class.h" exists! Using file name sw_driver Make args = KERNEL_SRC_DIR=/lib/modules/6.11.0-19-generic/build CPUFREQ_FIX_BACKPORTED=1 CLASSCREATE_CHANGE_BACKPORTED=1 DO_DEBUG_BUILD=0 DO_PROFILING=0 COMMON_INC_DIR=/opt/intel/oneapi/vtune/2025.0/sepdk/src/socwatch/socwatch_driver/../common/include MODULE_SYMVERS_FILE= FILE_NAME=sw_driver make[1]: Entering directory '/opt/intel/oneapi/vtune/2025.0/sepdk/src/socwatch/socwatch_driver' make -C /lib/modules/6.11.0-19-generic/build M=/opt/intel/oneapi/vtune/2025.0/sepdk/src/socwatch/socwatch_driver clean make[2]: Entering directory '/usr/src/linux-headers-6.11.0-19-generic' CLEAN /opt/intel/oneapi/vtune/2025.0/sepdk/src/socwatch/socwatch_driver/Module.symvers make[2]: Leaving directory '/usr/src/linux-headers-6.11.0-19-generic' make[1]: Leaving directory '/opt/intel/oneapi/vtune/2025.0/sepdk/src/socwatch/socwatch_driver' Using C compiler = /bin/gcc Using kernel build dir = /lib/modules/6.11.0-19-generic/build No module symvers file found Using common inc dir = /opt/intel/oneapi/vtune/2025.0/sepdk/src/socwatch/socwatch_driver/../common/include "/lib/modules/6.11.0-19-generic/build/include/linux/cpufreq.h" exists! "/lib/modules/6.11.0-19-generic/build/include/linux/device/class.h" exists! Using file name sw_driver Make args = KERNEL_SRC_DIR=/lib/modules/6.11.0-19-generic/build CPUFREQ_FIX_BACKPORTED=1 CLASSCREATE_CHANGE_BACKPORTED=1 DO_DEBUG_BUILD=0 DO_PROFILING=0 COMMON_INC_DIR=/opt/intel/oneapi/vtune/2025.0/sepdk/src/socwatch/socwatch_driver/../common/include MODULE_SYMVERS_FILE= FILE_NAME=sw_driver make[1]: Entering directory '/opt/intel/oneapi/vtune/2025.0/sepdk/src/socwatch/socwatch_driver' make -C /lib/modules/6.11.0-19-generic/build M=/opt/intel/oneapi/vtune/2025.0/sepdk/src/socwatch/socwatch_driver clean make[2]: Entering directory '/usr/src/linux-headers-6.11.0-19-generic' make[2]: Leaving directory '/usr/src/linux-headers-6.11.0-19-generic' ************************************************************ KERNEL_SRC_DIR=/lib/modules/6.11.0-19-generic/build CPUFREQ_FIX_BACKPORTED=1 CLASSCREATE_CHANGE_BACKPORTED=1 MODULE_SYMVERS_FILE= DO_DRIVER_PROFILING=0 INCDIR_1=/opt/intel/oneapi/vtune/2025.0/sepdk/src/socwatch/socwatch_driver/../common/include ************************************************************ make -C /lib/modules/6.11.0-19-generic/build M=/opt/intel/oneapi/vtune/2025.0/sepdk/src/socwatch/socwatch_driver modules PWD=/opt/intel/oneapi/vtune/2025.0/sepdk/src/socwatch/socwatch_driver make[2]: Entering directory '/usr/src/linux-headers-6.11.0-19-generic' warning: the compiler differs from the one used to build the kernel The kernel was built by: x86_64-linux-gnu-gcc-14 (Ubuntu 14.2.0-4ubuntu2) 14.2.0 You are using: gcc (Ubuntu 14.2.0-4ubuntu2) 14.2.0 CC [M] /opt/intel/oneapi/vtune/2025.0/sepdk/src/socwatch/socwatch_driver/./src/sw_driver.o CC [M] /opt/intel/oneapi/vtune/2025.0/sepdk/src/socwatch/socwatch_driver/./src/sw_hardware_io.o CC [M] /opt/intel/oneapi/vtune/2025.0/sepdk/src/socwatch/socwatch_driver/./src/sw_output_buffer.o /opt/intel/oneapi/vtune/2025.0/sepdk/src/socwatch/socwatch_driver/./src/sw_output_buffer.c:200:9: warning: "MIN" redefined 200 | #define MIN(x, y) ((x) <= (y) ? (x) : (y)) | ^~~ In file included from ./include/linux/kernel.h:28, from ./include/linux/cpumask.h:11, from ./arch/x86/include/asm/paravirt.h:21, from ./arch/x86/include/asm/cpuid.h:62, from ./arch/x86/include/asm/processor.h:19, from ./arch/x86/include/asm/timex.h:5, from ./include/linux/timex.h:67, from ./include/linux/time32.h:13, from ./include/linux/time.h:60, from ./include/linux/stat.h:19, from ./include/linux/module.h:13, from /opt/intel/oneapi/vtune/2025.0/sepdk/src/socwatch/socwatch_driver/inc/sw_internal.h:70, from /opt/intel/oneapi/vtune/2025.0/sepdk/src/socwatch/socwatch_driver/./src/sw_output_buffer.c:56: ./include/linux/minmax.h:329:9: note: this is the location of the previous definition 329 | #define MIN(a,b) __cmp(min,a,b) | ^~~ CC [M] /opt/intel/oneapi/vtune/2025.0/sepdk/src/socwatch/socwatch_driver/./src/sw_tracepoint_handlers.o CC [M] /opt/intel/oneapi/vtune/2025.0/sepdk/src/socwatch/socwatch_driver/./src/sw_collector.o CC [M] /opt/intel/oneapi/vtune/2025.0/sepdk/src/socwatch/socwatch_driver/./src/sw_mem.o CC [M] /opt/intel/oneapi/vtune/2025.0/sepdk/src/socwatch/socwatch_driver/./src/sw_internal.o CC [M] /opt/intel/oneapi/vtune/2025.0/sepdk/src/socwatch/socwatch_driver/./src/sw_file_ops.o CC [M] /opt/intel/oneapi/vtune/2025.0/sepdk/src/socwatch/socwatch_driver/./src/sw_ops_provider.o CC [M] /opt/intel/oneapi/vtune/2025.0/sepdk/src/socwatch/socwatch_driver/./src/sw_trace_notifier_provider.o CC [M] /opt/intel/oneapi/vtune/2025.0/sepdk/src/socwatch/socwatch_driver/./src/sw_reader.o CC [M] /opt/intel/oneapi/vtune/2025.0/sepdk/src/socwatch/socwatch_driver/./src/sw_telem.o CC [M] /opt/intel/oneapi/vtune/2025.0/sepdk/src/socwatch/socwatch_driver/./src/sw_pmt.o CC [M] /opt/intel/oneapi/vtune/2025.0/sepdk/src/socwatch/socwatch_driver/./src/sw_counter_list.o CC [M] /opt/intel/oneapi/vtune/2025.0/sepdk/src/socwatch/socwatch_driver/./src/sw_pci.o LD [M] /opt/intel/oneapi/vtune/2025.0/sepdk/src/socwatch/socwatch_driver/socwatch2_15.o MODPOST /opt/intel/oneapi/vtune/2025.0/sepdk/src/socwatch/socwatch_driver/Module.symvers CC [M] /opt/intel/oneapi/vtune/2025.0/sepdk/src/socwatch/socwatch_driver/socwatch2_15.mod.o LD [M] /opt/intel/oneapi/vtune/2025.0/sepdk/src/socwatch/socwatch_driver/socwatch2_15.ko BTF [M] /opt/intel/oneapi/vtune/2025.0/sepdk/src/socwatch/socwatch_driver/socwatch2_15.ko Skipping BTF generation for /opt/intel/oneapi/vtune/2025.0/sepdk/src/socwatch/socwatch_driver/socwatch2_15.ko due to unavailability of vmlinux make[2]: Leaving directory '/usr/src/linux-headers-6.11.0-19-generic' make[1]: Leaving directory '/opt/intel/oneapi/vtune/2025.0/sepdk/src/socwatch/socwatch_driver' ************ Built drivers are copied to /opt/intel/oneapi/vtune/2025.0/sepdk/src/socwatch/drivers directory ************ Done Done building the drivers

and "sudo ./insmod-sep -r -g user"

PYTHON is set to python3 Warning: the following driver(s) were not found loaded in the kernel: sep5. Warning: no vtsspp driver was found loaded in the kernel. Warning: no socwatch driver was found loaded in the kernel. Removing socperf3 driver from the kernel ... done. The socperf3 driver has been successfully unloaded. Attempting to stop PAX service ... Removing pax driver from the kernel ... done. The pax driver has been successfully unloaded. PAX service has been stopped. Detecting Secure Boot status... Secure Boot is disabled --------------------- Loading PAX driver --------------------- Detecting Secure Boot status... Secure Boot is disabled Checking for PMU arbitration service (PAX) ... not detected. Attempting to start PAX service ... Executing: insmod ./pax/pax-x32_64-6.11.0-19-genericsmp.ko Setting group ownership of devices to group "user" ... done. Setting file permissions on devices to "660" ... done. The pax driver has been successfully loaded. --------------------- Loading SOCPERF driver --------------------- Detecting Secure Boot status... Secure Boot is disabled Checking for socperf driver ... not detected. Attempting to start socperf service ... Executing: insmod ./socperf/src/socperf3-x32_64-6.11.0-19-genericsmp.ko Setting group ownership of devices to group "user" ... done. Setting file permissions on devices to "660" ... done. The socperf3 driver has been successfully loaded. --------------------- Loading SEP driver --------------------- Executing: insmod ./sep5-x32_64-6.11.0-19-genericsmp.ko sym_lookup_func_addr="ffffffff9762bb90" insmod: ERROR: could not insert module ./sep5-x32_64-6.11.0-19-genericsmp.ko: Invalid parameters ERROR: sep5 driver failed to load! You may need to build sep5 driver for your kernel. Please see the sep5 driver README for instructions.

How can I resolve this issue?

Re: VTune's self-check fails on GPU Compute/Media Hotspots

yuzhang3_intel — Tue, 18 Mar 2025 02:11:56 GMT

I'm not sure what happened in the current driver load process. Did you capture the dmesg log file? Maybe we can see something wrong from the file.

Re: VTune's self-check fails on GPU Compute/Media Hotspots

Hrishi2407 — Tue, 18 Mar 2025 07:49:32 GMT

This is what the logs showed

[ 163.614849] PAX: PMU arbitration service v1.0.2 has been started. [ 164.667142] socperf3_0: SocPerf Driver loading... [ 164.667146] socperf3_0: SocPerf Driver about to register chrdev... [ 164.667147] socperf3_0: SocPerf Driver: result of alloc_chrdev_region is 0 [ 164.667148] socperf3_0: SocPerf Driver: major number is 234 [ 164.667149] socperf3_0: SocPerf Driver: detected 8 CPUs in lwpmudrv_Load [ 164.667150] socperf3_0: SocPerf Driver: creating device socperf3!c... [ 164.667339] socperf3_0: PMU check enabled! F6.Mbd.S1 index=-1 [ 164.667341] socperf3_0: No MMIO list information detected! [ 164.667342] socperf3_0: SocPerf Driver v3.0.0 has been loaded. [ 165.765716] sep5: disagrees about version of symbol pmt_telem_get_next_endpoint [ 165.765721] sep5: Unknown symbol pmt_telem_get_next_endpoint (err -22) [ 165.765744] sep5: disagrees about version of symbol pmt_telem_register_endpoint [ 165.765744] sep5: Unknown symbol pmt_telem_register_endpoint (err -22) [ 165.765811] sep5: disagrees about version of symbol pmt_telem_get_endpoint_info [ 165.765812] sep5: Unknown symbol pmt_telem_get_endpoint_info (err -22) [ 165.765836] sep5: disagrees about version of symbol pmt_telem_read [ 165.765837] sep5: Unknown symbol pmt_telem_read (err -22)

Re: VTune's self-check fails on GPU Compute/Media Hotspots

yuzhang3_intel — Tue, 18 Mar 2025 08:04:27 GMT

Please follow the steps below to disable PMT and rebuild the driver:

Steps to disable PMT:

Open /opt/intel/sep/sepdk/src/build-driver
Updated line 122 from "BUILD_PMT="YES"" to "BUILD_PMT="NO""
Build and load the driver using the regular steps

Re: VTune's self-check fails on GPU Compute/Media Hotspots

Hrishi2407 — Tue, 18 Mar 2025 08:19:08 GMT

The sep5 and vtsspp drivers are loading now but still having issues with the socwatch driver

PYTHON is set to python3 Warning: the following driver(s) were not found loaded in the kernel: sep5. Warning: no vtsspp driver was found loaded in the kernel. Warning: no socwatch driver was found loaded in the kernel. Removing socperf3 driver from the kernel ... done. The socperf3 driver has been successfully unloaded. Attempting to stop PAX service ... Removing pax driver from the kernel ... done. The pax driver has been successfully unloaded. PAX service has been stopped. Detecting Secure Boot status... Secure Boot is disabled --------------------- Loading PAX driver --------------------- Detecting Secure Boot status... Secure Boot is disabled Checking for PMU arbitration service (PAX) ... not detected. Attempting to start PAX service ... Executing: insmod ./pax/pax-x32_64-6.11.0-19-genericsmp.ko Setting group ownership of devices to group "user" ... done. Setting file permissions on devices to "660" ... done. The pax driver has been successfully loaded. --------------------- Loading SOCPERF driver --------------------- Detecting Secure Boot status... Secure Boot is disabled Checking for socperf driver ... not detected. Attempting to start socperf service ... Executing: insmod ./socperf/src/socperf3-x32_64-6.11.0-19-genericsmp.ko Setting group ownership of devices to group "user" ... done. Setting file permissions on devices to "660" ... done. The socperf3 driver has been successfully loaded. --------------------- Loading SEP driver --------------------- Executing: insmod ./sep5-x32_64-6.11.0-19-genericsmp.ko sym_lookup_func_addr="ffffffffb7e2bb90" Setting group ownership of devices to group "user" ... done. Setting file permissions on devices to "660" ... done. The sep5 driver has been successfully loaded. --------------------- Loading VTSSPP driver --------------------- Checking for vtsspp driver ... not detected. Executing: insmod ./vtsspp/vtsspp-x32_64-6.11.0-19-genericsmp.ko gid=1000 mode=0660 ksyms="ffffffffb7e2bb90" The vtsspp driver has been successfully loaded. --------------------- Loading SOCWATCH driver --------------------- Checking for socwatch driver ... not detected. Executing: insmod ./socwatch/drivers/socwatch2_15-x32_64-6.11.0-19-genericsmp.ko insmod: ERROR: could not insert module ./socwatch/drivers/socwatch2_15-x32_64-6.11.0-19-genericsmp.ko: Invalid parameters Error: socwatch2_15-x32_64-6.11.0-19-genericsmp driver failed to load! You may need to build socwatch2_15-x32_64-6.11.0-19-genericsmp driver for your kernel. Please see the socwatch2_15-x32_64-6.11.0-19-genericsmp driver README for instructions.

Re: VTune's self-check fails on GPU Compute/Media Hotspots

yuzhang3_intel — Tue, 18 Mar 2025 08:33:44 GMT

Please post the output of the command line below:

$ lsmod | grep socwatch

Re: VTune's self-check fails on GPU Compute/Media Hotspots

Hrishi2407 — Tue, 18 Mar 2025 08:49:41 GMT

It did not return any output

Re: VTune's self-check fails on GPU Compute/Media Hotspots

yuzhang3_intel — Tue, 18 Mar 2025 08:51:35 GMT

Actually, the driver is only for package power collection, you can ignore it if you don't need it. For CPU profiling, sep/vtsspp driver are enough.

Re: VTune's self-check fails on GPU Compute/Media Hotspots

Hrishi2407 — Tue, 18 Mar 2025 09:01:15 GMT

Understood, but I also require package power monitoring alongside CPU/GPU profiling.

Re: VTune's self-check fails on GPU Compute/Media Hotspots

yuzhang3_intel — Tue, 18 Mar 2025 09:27:09 GMT

This is one known issue, which is fixed in the coming 2025.1 release.

I am not sure if you can insmod the ko file manually in the path below, and execute the command line: ./insmod-socwatch -r -g vtune (the same group as VTune)

.....sepdk/src/socwatch/drivers

Re: VTune's self-check fails on GPU Compute/Media Hotspots

Hrishi2407 — Tue, 18 Mar 2025 09:50:23 GMT

It seems I'm unable to insmod the ko file manually

Re: VTune's self-check fails on GPU Compute/Media Hotspots

Hrishi2407 — Tue, 18 Mar 2025 10:29:48 GMT

It seems even after the sep5 and vtsspp drivers loading successfully, the main issue still persists

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

Re: VTune's self-check fails on GPU Compute/Media Hotspots

yuzhang3_intel — Wed, 19 Mar 2025 02:53:21 GMT

Can you check the self check log file and see if the sycl8 library is missing? If yes, it means you need to install the Intel oneAPI DPC++/C++ compiler because the current self-check utility depends on the library located in the compiler.

Re: VTune's self-check fails on GPU Compute/Media Hotspots

Hrishi2407 — Wed, 19 Mar 2025 07:03:48 GMT

I installed the Intel oneAPI DPC++/C++ compiler and after running self check and "cat /tmp/vtune-tmp-user/self-checker-2025.03.19_06.42.09/log.txt | grep -i "missing\|error\|sycl\|sycl8" ":

i915Status: MissingDriver isFtraceAvailable: ftraceAccessError,debugfsNotAccessible vtune: Error: Error 0x4000002a (Database interface error) -- Cannot run data transformation `Conditional GPU metric'. vtune: Error: Error 0x4000002a (Database interface error) -- Cannot run data transformation `Conditional GPU metric'. sycl-ls SYCL_PI_TRACE has been removed use SYCL_UR_TRACE instead SYCL_PI_TRACE has been removed use SYCL_UR_TRACE instead SYCL_PI_TRACE has been removed use SYCL_UR_TRACE instead

Also I got VTune from the Intel oneAPI Base Toolkit so the compiler should not be an issue

Re: VTune's self-check fails on GPU Compute/Media Hotspots

yuzhang3_intel — Wed, 19 Mar 2025 07:34:19 GMT

See the output of the command line:

$ sycl-ls

If oneAPI DPC++ environment is ready, you should see something below:

yuzhang3@yuzhang3-10710:~/yuzhang3/win_bk/MyDocuments/MyProject/PrjData/project$ sycl-ls
[level_zero:gpu][level_zero:0] Intel(R) oneAPI Unified Runtime over Level-Zero, Intel(R) UHD Graphics 9.7.0 [1.3.28202]
[opencl:cpu][opencl:0] Intel(R) OpenCL, Intel(R) Core(TM) i7-10710U CPU @ 1.10GHz OpenCL 3.0 (Build 0) [2024.18.10.0.08_160000]
[opencl:gpu][opencl:1] Intel(R) OpenCL Graphics, Intel(R) UHD Graphics OpenCL 3.0 NEO [23.52.28202.52]

Re: VTune's self-check fails on GPU Compute/Media Hotspots

Hrishi2407 — Wed, 19 Mar 2025 07:39:40 GMT

Sycl is able to access the gpu but VTune isn't able to

[level_zero:gpu][level_zero:0] Intel(R) oneAPI Unified Runtime over Level-Zero, Intel(R) Arc(TM) Graphics 20.4.4 [1.6.32224+14] [opencl:cpu][opencl:0] Intel(R) OpenCL, Intel(R) Core(TM) Ultra 7 258V OpenCL 3.0 (Build 0) [2024.18.12.0.05_160000] [opencl:gpu][opencl:1] Intel(R) OpenCL Graphics, Intel(R) Arc(TM) Graphics OpenCL 3.0 NEO [24.52.32224]

Re: VTune's self-check fails on GPU Compute/Media Hotspots

yuzhang3_intel — Wed, 19 Mar 2025 07:58:06 GMT

Run the test app in VTune installation folder:

yuzhang3@yuzhang3-10710:/opt/intel/oneapi/vtune/latest/bin64/self_check_apps/matrix.dpcpp$ ./matrix.dpcpp
Address of buf1 = 0x76ec74400010
Offset of buf1 = 0x76ec74400180
Address of buf2 = 0x76ec73a00010
Offset of buf2 = 0x76ec73a001c0
Address of buf3 = 0x76ec73000010
Offset of buf3 = 0x76ec73000100
Address of buf4 = 0x76ec72600010
Offset of buf4 = 0x76ec72600140
Using multiply kernel: multiply1

Available devices:
Intel(R) UHD Graphics
Intel(R) UHD Graphics

*****
* Running on Intel(R) UHD Graphics
*****
No subdevices available, using the whole device

************
* Finished *
************
Elapsed Time: 0.531978s

Re: VTune's self-check fails on GPU Compute/Media Hotspots

Hrishi2407 — Wed, 19 Mar 2025 08:20:57 GMT

These were the results:

Address of buf1 = 0x7855b21ff010 Offset of buf1 = 0x7855b21ff180 Address of buf2 = 0x7855b19fe010 Offset of buf2 = 0x7855b19fe1c0 Address of buf3 = 0x7855b11fd010 Offset of buf3 = 0x7855b11fd100 Address of buf4 = 0x7855b09fc010 Offset of buf4 = 0x7855b09fc140 Using multiply kernel: multiply1 Available devices: Intel(R) Arc(TM) Graphics Intel(R) Arc(TM) Graphics ***** * Running on Intel(R) Arc(TM) Graphics ***** No subdevices available, using the whole device ***** * Running on Intel(R) Arc(TM) Graphics ***** No subdevices available, using the whole device ************ * Finished * ************ Elapsed Time: 0.354231s