Dell Precision 7540 and 7740 Owner's Thread

Thanks. At what percentage capacity will the Intel GPU be taxed for such videos?

 

Not sure what you're asking. Newer Intel integrated GPUs have hardware decoding for HEVC video (in addition to the older H.264) and should be able to play back 4K/60fps videos without breaking a sweat, as long as your player software supports hardware-accelerated playback.

 

I was wondering how much it would sweat. Similar to 70% cpu utilization, 80% etc,

 

Well, not much, because there is a separate hardware circuit for decoding videos. That is what is meant by 'hardware codec', because the circuitry and the silicon space is separate from the main pipelined circuit of the CPU that contains the ALU and FPU.

Of course, software codecs (x264, x265) exist, too, but many video programs directly support the various hardware encoders (Intel QuickSync, NVIDIA CUVID/NVENC, AMD VCE/VCN), because they are so much faster, with a minimal loss in quality.

 

Yes, but the seperate hardware circuit is also going to be operating at some operations per second, just like a CPU. So what level it is being taxed should also be measurable...

 

Certainly, and you have the Video Decode graph in Task Manager (not sure of a Linux/OS X equivalent) to measure this:

 

I'll do a measurement when I get back to my work machine on Monday but as I recall it is barely noticeable (<10%). 4K video playback isn't an intensive operation anymore for newer machines.

 

So what IS an intensive application? For video games, isn't the graphics load basically the same? Display things on the screen?
Not a gamer here, so I'm a bit clueless.

 

Not really.

With 3D video games, the graphics card is reading vertices and texture data from VRAM, and generating a 3D scene by rasterising said vertices, applying textures to them and then putting them through a fragment shader. Here, lighting, anti-aliasing and all other measures of post-processing (like ambient occlusion, volumetric lighting, bloom, motion blur, etc) are applied, and finally pushed to the screen, and you see your nice 3D graphics. This is highly computationally intensive, but only in a certain way: it's basically plenty of vector and matrix mathematics.

Each general-purpose core in a GPU is designed in a way to maximise performance of this operation (it's called single-instruction, multiple data: SIMD). Take this transformation matrix, pre-multiply it to that fragment position vector, take that fragment normal, multiply it by some normal map colour vector, etc etc. Repeat ad nauseum for every vertex and every fragment normal (this is a very, very simplified example).

With video decoding, the computational load is somewhat different: it tends towards compression and decompression rather than pure vector mathematics. I don't claim to be an expert on compression, so I can't go into any more detail, save the fact that the codecs on board Intel/NVIDIA/AMD GPUs are designed in a way to maximise performance of said compression/decompression algorithms.

 

Hey I just picked up a Dell 7740 refurbished from the Dell outlet .The price was right. It seems like a nice laptop. I put in 2 sticks of Samsung 32 gig memory for a total of 64 gig. I ordered 2 WD_BLACK SN750 NVMe SSD to rebuild my laptop with the new release Windows 10 to be released in a few days. My unit came with the 1920x1080 display.

The latest Intel AX200 wireless driver from Intel is quicker for me. I don't know if anybody else has installed it.