Ok, so this is the code:
I am currently investigating the L1, L2 and L3 bandwidth of our latest Haswell
CPU (Xeon E5-2680 v3). The L1, L2 and L3 size of this CPU is 32 KiB, 256 KiB
and 32 MiB, respectively.
I am using a SAXPY-like kernel (i.e., X += Y) to measure the bandwidth. Please
find the benchmark results attached.
We came across an issue in our Intel (2016) compiler build where it appears separate, but identically named, classes in anonymous namespaces are not being resolved correctly. We've managed to create small example that illustrates the problem, and wondered if somebody could please investigate.
I've started a new thread, after the old threads not been resolved yet.
We develop application which does the following:
1. Encode multiple video streams
2. Deliver graphic images from multiple streams --> fpga --> display (from fpga)
3. Using media sdk server on Linux
The 2nd above need to support OpenGL.
As far as I understand to support OpenGL we must use the GPU.
I would plea like to know:
Does intel media SDK supports Intel(R) Xeon(R) CPU E3-1284L v4 @ 2.90GHz?
I can't find any info even on ark.intel.com
Which version of SDK should I need to choose, the latest?
It seems that patches from kernel 3.14.5 already in kernel version like 4.4.*
Is it true? I've just try to apply patches to newer kernels but they are can not be applied (but it's normal) and after read some code I see that the code in kernel looks close to patches :)
Am I misunderstand and kernel 4.4.* not ready to work with quicksync out of the box or not?
My application, on the command line uses these arguments:
./a.out --zones 32,32,32 --gset 1 --groups 32 --quad 96 --nest DGZ --niter 10
For the OS X gui, the Application parameters input does not accept dash-dash syntax. It appears to try to format the double dashes into a single long dash, much like Word would do.
How do I pass double-dash arguments to my application?
On a related question, in the GUI where do you set ENV variables, like OMP_NUM_THREADS=16 ?
Does the encoder need progressive as input ?
For example, with the encode demo, if we use interlaced, does it mean that the encode demo first use vpp to change from interlaced to progressive ?
Dear Intel developers,
by using a __mm128 type, what is the best and fast way to fill that type one float per time starting from an array of float? Thanks.