I am actually trying to use Intel Trace Analyzer and Collector (ITAC) to profile my MPI code written in Fortran.

The code does execute MPI_init at first and MPI_finalize at the end.

Following the thread of [ https://software.intel.com/en-us/forums/intel-clusters-and-hpc-technolog... ],

I was able to obtain .stf and .prot files yet the size of .stf file is zero, thus unable to open it with ITAC.

I'm using "Intel® Parallel Studio XE 2018 Update 3" for compilers, mvapich2.2.2-qlc-intel18 for MPI library, which is compatible with ITAC

and CentOS 6.4 for the OS.

Any help will be deeply appreciated.

Hi,

I am using Intel screen capture plugin from [Intel(R)_Media_SDK_2016.0.1] and sample code from [Intel Media SDK Samples 2016 6.0.0.142]. I am facing a strange issue on 6th gen laptop that if I configure capture plugin to output in RGB32 format then it works well, but if I set to output in NV12 then m_mfxSession.SyncOperation() is always failing with MFX_ERR_DEVICE_FAILED error.

This issue is observed on few laptops and all those were having 6th gen CPU and Windows7 64 bit OS. One of that environment is as following:

OS: Microsoft Windows 7 (6.1) Professional 64-bit  Service Pack 1 (Build 7601)
CPU: Intel(R) Core(TM) i7-6820HQ CPU @ 2.70GHz (Skylake)
GPU: Intel(R) HD Graphics 530
DriverVersion: 20.19.15.4331
DriverDate: 11/20/2015

As per windows update, there is no further updates available for this graphics driver.

My screen capture plugin configuration is as following:

Output format :NV12

Input Resolution :1920x1088

Crop X,Y,W,H :0,0,1920,1080

Output Resolution :1920x1080

Frame rate :24.00

Memory type :system

MediaSDK impl :hw

MediaSDK version :1.17

So, can you please help me out to find out the root cause and its fix? Is it any known issue or limitation with 6th gen CPU + Windows7 OS combination?  

What does PARTIAL_ACCELERATION mean in described configuration ?
Will be HW encoding used or all pipeline will be executed in the CPU ?

Encoding pipeline consists of two operations:
1. VPP transcoding    YUV12 --> NV12
2. Encoding      NV12 --> h.264

Direct3D11 Acceleration device is used.
MFXVideoENCODE_Init()    returns MFX_ERR_NONE
MFXVideoVPP_Init()      returns MFX_WRN_PARTIAL_ACCELERATION

Dear Intel technician

1. Hardware-acceleration on the Intel Media SDK 2018 R1, Does it support HEVC codec decode, encode, 8-bit and 10-bit From Linux and Windows?

2. Can I Software(CPU) HEVC 8- and 10-bit decode, encode Intel Server Studio 2018 R1 Professinal Edition and Intel Xeon processor E5 v4 or v5 buying it?

3. Can I SW, HW accerlated encode, decode, 8-bit and 10bit a free download of Intel Media SDK 2018 R1 from Windows Server 2016?

Thank you

This question is appeared as a result of the discussion:
https://software.intel.com/en-us/forums/intel-media-sdk/topic/780765

vitm: Let's focus on asynchronous pipeline.
It consists of two operations:

VPP(f1) --> ENCODE(f2) --> SYNC(f1) --> SYNC(f2)

Liu, Mark (Intel):
there is mistake in your post. MSDK is a hardware based library, so all the key operations are asynchnized, your pipe line description should be:

VPP(f1) --> SYNC(f1) --> ENCODE(f2) -->SYNC(f2)

This is critical since without the first sync operation, the second enc operation might wait forever for the surface be available.

> I do not agree with this answer.

What does documentation say

mediasdk-man.pdf
chapter: Asynchronous Pipeline ( page 21-22 )

To construct an asynchronous pipeline:

The application passes the output of an upstream SDK function to the input of the downstream SDK function
…
The application only needs to synchronize after the last SDK function.
Explicit synchronization of intermediate results is not required and in fact can slow performance.

For performance considerations, the application must submit multiple operations and delays synchronization as much as possible.

The operation sequence,
ENCODE(f1) → ENCODE(f2) → SYNC(f1) → SYNC(f2) is recommended,

compared with:
ENCODE(f1) → SYNC(f1) → ENCODE(f2) → SYNC(f2)

------------------------------

> Where is the mistake ?

The reference manual advises using EndOfStream​ with MFXVideoENCODE_Reset but EndOfStream​ is deprecated according to the mfxExtCodingOption documentation. Who's wrong?

Good day,

Please help in clarifying following matter.

As I understand following table describes HEVC decode/encode capability with Media Server Studio 2017 on windows/linux.

Given for Skylake chipset.

                         decode encode

8 bit HW HW

10 bit CPU HW/CPU

With Media SDK 2018 (Windows) this table changes to following

                         decode encode

8 bit HW HW

10 bit GPU(gacc) HW/GPU(gacc)

Question 1 : Am I correct in this assumption ?

Question 2 : If I have following stack: 

Windows Server 2016

Intel Media SDK 2018

Intel Xeon 6th generation (SkyLake)

It can support server level quality of HEVC 10-bit decoding(gpu) and encoding(hw/gpu) ?

Good Day

After downloading the mediasdk-tutorials-0.0.3, I modified the simple_1_session source as shown below.

    int main()

    {

        ....

        MFXVideoSession session;

        sts = Initialize(impl, ver, &session, NULL);

        sts = Initialize(impl, ver, &session, NULL);

        sts = Initialize(impl, ver, &session, NULL);

        sts = Initialize(impl, ver, &session, NULL);

        sts = Initialize(impl, ver, &session, NULL);

        sts = Initialize(impl, ver, &session, NULL);

        ....

    }

iHD_drv_video.so is loaded from root account, but why is it failing to find i965 in video group account?

my environment CentOS, MSS SDK 2015 R4..

// root

# ./simple_session

libva info: VA-API version 0.35.0

libva info: va_getDriverName() returns 0

libva info: User requested driver 'iHD'

libva info: Trying to open /opt/intel/mediasdk/lib64/iHD_drv_video.so

libva info: Found init function __vaDriverInit_0_32

libva info: va_openDriver() returns 0

libva info: VA-API version 0.35.0

libva info: va_getDriverName() returns 0

libva info: User requested driver 'iHD'

libva info: Trying to open /opt/intel/mediasdk/lib64/iHD_drv_video.so

libva info: Found init function __vaDriverInit_0_32

libva info: va_openDriver() returns 0

libva info: VA-API version 0.35.0

libva info: va_getDriverName() returns 0

libva info: User requested driver 'iHD'

libva info: Trying to open /opt/intel/mediasdk/lib64/iHD_drv_video.so

libva info: Found init function __vaDriverInit_0_32

libva info: va_openDriver() returns 0

libva info: VA-API version 0.35.0

libva info: va_getDriverName() returns 0

libva info: User requested driver 'iHD'

libva info: Trying to open /opt/intel/mediasdk/lib64/iHD_drv_video.so

libva info: Found init function __vaDriverInit_0_32

libva info: va_openDriver() returns 0

libva info: VA-API version 0.35.0

libva info: va_getDriverName() returns 0

libva info: User requested driver 'iHD'

libva info: Trying to open /opt/intel/mediasdk/lib64/iHD_drv_video.so

libva info: Found init function __vaDriverInit_0_32

libva info: va_openDriver() returns 0

libva info: VA-API version 0.35.0

libva info: va_getDriverName() returns 0

libva info: User requested driver 'iHD'

libva info: Trying to open /opt/intel/mediasdk/lib64/iHD_drv_video.so

libva info: Found init function __vaDriverInit_0_32

libva info: va_openDriver() returns 0

// video group user

# ./simple_session

libva info: VA-API version 0.35.0

libva info: va_getDriverName() returns 0

libva info: User requested driver 'iHD'

libva info: Trying to open /opt/intel/mediasdk/lib64/iHD_drv_video.so

libva info: Found init function __vaDriverInit_0_32

libva info: va_openDriver() returns 0

libva info: VA-API version 0.35.0

libva info: va_getDriverName() returns 1

libva error: va_getDriverName() failed with operation failed,driver_name=i965

terminate called after throwing an instance of 'std::bad_alloc'

  what():  std::bad_alloc

Abort (core dumped)

Good Day.

1. H.264(AVC) and H.265(HEVC) encoding Linux/Windows performance banchmark comparison result?

2. H.264(AVC) and H.265(HEVC) encoding Linux/Windows performance and quality test results by media server studio version?

Thank you..

hey ,nice to meet to:

when i am testing the sample sample decode with screen capture plu-in :error belows:

sample_decode.exe h264-hw -p 22d62c07e672408fbb4cc20ed7a053e4 -scr:w 1500 -scr:h 700 -d3d11  -o output.yuv

sts=MFX_ERR_OINVALID_VIDEO_PARAM(-15),CDecodingPipeline::init,m_pmfxDEX->Init failed at pipeline_decode.cpp:447

sts=MFX_ERR_OINVALID_VIDEO_PARAM(-15),wmain,Pipeline.Init failed at sample_decode.cppL681

plugin_loader.h:199[INFO]MFXBaseUSER_Unload(session=0x000001c9c34A20),sts=0

how can i resolve this issue?

Hi, i'm trying to decode and encode 10 bit HEVC using hardware on Kabylake i7 7700 processor, command line used

sample_multi_transcode -i::h265 "D:\Intel Media SDK Plugins\4K Demos\Samsung Journey Of Color 4K Demo.265" -o::h265 D:\TESTE.hevc -hw -async 10 -pe hevce_gacc -pd hevcd_hw -cqp -qpi 20 -qpp 22 -qpb 24 -u balanced -l 1 -gop_size 24 -dist 4 -num_ref 6 -bref -bpyr -ec::p010 -w 1920 -h 1080 -WeightedPred::implicit -WeightedBiPred::implicit -amfs:on -extbrc::implicit -pp

I'm on Windows 10, tryed the first media sdk driver release and the latest, both produce same problems, if i use hevc decode plugin from Media SDK 2018R1 it works perfectly but are too slow for decode > encode, i have uploaded the result hevc stream in attachment

Regards

First of all,i am glad to recive to reply,I have tried two Intel media SDK (Intel(R) _media_sdk_2016.0.2 and Intel(R) _media_sdk_2017.0.1)

First:

INTELMEDIASDK_WINSDK_PATH     C:\Program Files (x86)\Windows Kits\10

INTELMEDIASDKROOT                      C:\Program Files (x86)\IntelSWTools\Intel(R)_Media_SDK_2016.0.2

file: ..x64/Debug/

                          22d62c07e672408fbb4cc20ed7a053e4/mfxplugin64_screen_capture.dll

                          22d62c07e672408fbb4cc20ed7a053e4/plugin.cfg

                          54542616243341e693ae899942ce7355/mfxplugin_camera_pipe64.dll

                          54542616243341e693ae899942ce7355/plugin.cfg

                          lib/igfx_s3dcontrol.lib

                          lib/libmfx.lib

                          lib/libmfx_d.lib

                          lib/sample_common.lib

                          libmfxaudiosw64.dll

                          libmfxsw64.dll

                          sample_decode.exe

                          sample_decode.ilk

                          sample_decode.pdb

commondline :

sample_decode.exe capture -d3d11  -hw -f 25 -p 22d62c07e672408fbb4cc20ed7a053e4 -w 1920 -h 1080 -fourcc nv12  -o output.yuv

yes,work well:

decoding Sample Version 6.0

Decoding Started

Frame number:550 ,fps 4.920 ,fread_fps:0.000,fwrite_fps:5.388

    MY QUESTION IS:one:Why the use of CPU is very high,how can i cut down the cpu?，two:  ,how can i get the 25 fps??

second:
INTELMEDIASDK_WINSDK_PATH            C:\Program Files (x86)\Windows Kits\10
INTELMEDIASDKROOT                               C:\Program Files (x86)\IntelSWTools\Intel(R)_Media_SDK_2017.0.1
x64/Debug/        22d62c07e672408fbb4cc20ed7a053e4/mfxplugin64_screen_capture.dll
                          22d62c07e672408fbb4cc20ed7a053e4/plugin.cfg
                          lib/igfx_s3dcontrol.lib
                          lib/libmfx.lib
                          lib/libmfx_d.lib
                         lib/sample_common.lib
                         libmfxaudiosw64.dll
                         libmfxsw64.dll
                         sample_decode.exe

commondline:sample_decode.exe capture -d3d11 -p 22d62c07e672408fbb4cc20ed7a053e4 -scr:w 1920 -scr:h 1080 -fourcc nv12
or:::sample_decode.exe capture -d3d11  -f 25 -p 22d62c07e672408fbb4cc20ed7a053e4 -w 1920 -h 1080 -fourcc nv12
Decoding Sample Version  8.0
Features:Press 1 to toggle fullscreen rending on/off
Why is this version going wrong?

hello,nice to meet you:

i'm download the ZIP:https://software.intel.com/zh-cn/articles/intel-media-sdk-screen-capture,(simple_screen_capture.zip)
x64/Debug/
         22d62c07e672408fbb4cc20ed7a053e4/mfxplugin64_screen_capture.dll
        22d62c07e672408fbb4cc20ed7a053e4/plugin.cfg
        lib/igfx_s3dcontrol
        lib/libmfx.lib
        lib/libmfx_d.lib
        lib/sample_common.lib
        libmfxaudiosw64.dll
        libmfxsw64.dll
        tutorial_screen_capture.exe
my input commondline:
tutorial_screen_capture.exe -n 500 -w 1920 -h 1080 -enc -enc_codec 'avc' -enc_file test1.avc

ERROR:
Started
Failed to Query DEcoder
Assertion failed:mfxSts==MFX_ERR_NONE,file tutorial_screen_capture.cpp,line 692
MY QUESTION:how can i successful use the tutorial_screen_capture.exe????

Hello,

I have code which use sdk hardware implementation with IO.Patern as system memory. 

My application rendered video from multiple camera simultaneously. so it perform decode+render operation for multiple camera simultaneously.

I observed SDK not able to decode correctly after 10 FullHD stream and 18 HD streams. It is throwing randomly MFX_ERR_UNSUPPORTED return code.

cpu usage and gpu usage  52% ans 65%(approx.) respectively for 10 FullHd streams.

cpu usage and gpu usage  70% ans 78%(approx) respectively for 18 FullHd streams. 

System Information :

cpu : intel core i-3770 @3.40ghz 3.40ghz

graphics : intel HD graphics 4000

my questions is is there any upper limitation for sdk ?

Hi there,

I am trying the sample_encode and sample_decode programs from the Media SDK  2018 R1 API version 1.26 on windows 10

the input file for encoding can be played by a raw yuv player correctly. 

however I have two questions regarding to the generated encoding output file:

1.  when trying to play the encoding output file on VLC, it plays only part of the video contents. i.e. the video finishes before the full content as compared to the input yuv file.  However,  the sample_decode -r option does  render the full content  given the same encoding output file.

2. the sample_decode renders the video at a much higher frame rate compared to  raw yuv input.  i.e. if the input yuv file plays for 10 seconds, the sample_decode would complete the full video content in 3 seconds.  Is this expected? 

 Any advice is greatly appreciated. thank you.

here is the command prompt message for sample_encode and sample_decode programs

D:\MyProjects\IntelMediaSDK\samples\samples\_build\x64\Debug> .\sample_encode.exe -nv12 -tff -sw h264 -u speed -i "d:

\intelsdk_test\nv12.yuv" -o "d:\intelsdk_test\sample_encode.h264" -w 720 -h 576 -dump "d:\intelsdk_test\config.txt"

Encoding Sample Version 8.3.26.0

Input file format       NV12

Output video            AVC

Source picture:

        Resolution      720x576

        Crop X,Y,W,H    0,0,720,576

Destination picture:

        Resolution      720x576

        Crop X,Y,W,H    0,0,720,576

Frame rate      30.00

Bit rate(Kbps)  3000

Gop size        60

Ref dist        4

Ref number      0

Idr Interval    0

Target usage    speed

Memory type     system

Media SDK impl          sw

Media SDK version       1.26

Processing started

Frame number: 241

Encoding fps: 74

Processing finished

PS D:\MyProjects\IntelMediaSDK\samples\samples\_build\x64\Debug> .\sample_decode.exe h264 -sw -i "D:\intelsdk_test\sampl

e_encode.h264" -r

Decoding Sample Version 8.3.26.0

Input video     AVC

Output format   NV12

Input:

  Resolution    720x576

  Crop X,Y,W,H  0,0,0,0

Output:

  Resolution    720x576

Frame rate      30.00

Memory type             d3d

MediaSDK impl           sw

MediaSDK version        1.26

Decoding started

Frame number:  241, fps: 195.018, fread_fps: 0.000, fwrite_fps: 1723.559

Decoding finished

I have a media server studio 2017 pro.

I can not find mfxplugin32_hevcd.sw.dll.

Is not support software decode for HEVC with Win32 apps?

I have been running Media Server Studio 2017R3 for quite a while but have been experiencing crashes with the nimble transcoder.  Therefore, I have been attempting to install Media Server Studio 2018R1 to see if this fixes the problem.

Linux: CentOS Linux release 7.4.1708 (Core)

proc/cpuinfo:

processor : 0

vendor_id : GenuineIntel

cpu family : 6

model : 94

model name : Intel(R) Core(TM) i7-6700K CPU @ 4.00GHz

stepping : 3

microcode : 0xba

cpu MHz : 1599.843

cache size : 8192 KB

physical id : 0

siblings : 8

core id : 0

cpu cores : 4

apicid : 0

initial apicid : 0

fpu : yes

fpu_exception : yes

cpuid level : 22

wp : yes

flags : fpu vme de pse tsc msr pae mce cx8 apic sep mtrr pge mca cmov pat pse36 clflush d

ts acpi mmx fxsr sse sse2 ss ht tm pbe syscall nx pdpe1gb rdtscp lm constant_tsc art arch_perfmon p

ebs bts rep_good nopl xtopology nonstop_tsc aperfmperf eagerfpu pni pclmulqdq dtes64 monitor ds_cpl

 vmx est tm2 ssse3 fma cx16 xtpr pdcm pcid sse4_1 sse4_2 x2apic movbe popcnt tsc_deadline_timer aes

 xsave avx f16c rdrand lahf_lm abm 3dnowprefetch epb invpcid_single intel_pt tpr_shadow vnmi flexpr

iority ept vpid fsgsbase tsc_adjust bmi1 hle avx2 smep bmi2 erms invpcid rtm mpx rdseed adx smap cl

flushopt xsaveopt xsavec xgetbv1 dtherm ida arat pln pts hwp hwp_notify hwp_act_window hwp_epp

bogomips : 8016.00

clflush size : 64

cache_alignment : 64

address sizes : 39 bits physical, 48 bits virtual

power management:

lspci -nn -s 0:02.0

00:02.0 VGA compatible controller [0300]: Intel Corporation HD Graphics 530 [8086:1912] (rev 06)

I have completed the install using the install scripts, but I am getting an error when I verify the installation using vainfo

$ vainfo

libva info: VA-API version 1.0.0

libva info: va_getDriverName() returns -1

libva info: User requested driver 'iHD'

libva info: Trying to open /opt/intel/mediasdk/lib64/iHD_drv_video.so

libva info: Found init function __vaDriverInit_1_0

get chip id failed: -1 [13]

param: 4, val: 0

libva error: /opt/intel/mediasdk/lib64/iHD_drv_video.so init failed

libva info: va_openDriver() returns 18

vaInitialize failed with error code 18 (invalid parameter),exit

Any suggestions appreciated.

Cheers

Canukhd

The simple_encode_d3d_async sample does

EncRequest.NumFrameSuggested = EncRequest.NumFrameSuggested + mfxEncParams.AsyncDepth;

to determine how many surfaces to create. But if you call QueryIOSurf with a larger AsyncDepth, it returns a larger NumFrameSuggested​, suggesting it already accounts for AsyncDepth internally. In my testing, the original value of NumFrameSuggested is insufficient and makes my program run out of surfaces. Adding a single surface seems to fix this.

How many surfaces should I create?
Bruno

Good Day..

Which one of Intel Media SDK 2018 or Intel Media Server Studio 2017 is better for H.264(AVC) transcoding on Windows Server?

When transcoding from Linux to H.264(AVC), Intel Media Server Studio 2018 has better performance between

 Intel Media Server Studio 2018 and Intel Media Server Studio 2017, right?

How good is performance if benchmarking results?

Thank you..

Hello,

Can you give me some advice on a number of memory leaks I am seeing with the latest Intel® Media SDK 2018 R1 - Media Samples 8.3.26.352\sample_decode

If you were to loop the contents of  'main()' 5 times using the latest decode sample, adding the call to Pipeline.Close(); on each iteration (also add  UnregisterClass(m_sWindowParams.lpClassName, GetModuleHandle(NULL)); in CDecodeD3DRender::Close()) you will see the attached memory leaks.

It is very common for a window to host a number of different streams for decode.

This is a simple test to illustrate the leaks, in our released code the structure is more defined for purpose but we see the same leaks.

Our window is never actually closed.

Thank you in advance for any advice.

Rich.