Hacker News Reader

Comment by mrb | Hacker News Reader

Intel has the fastest processors but it doesn't have the best value for the money, in the entry level and mid level consumer market (sub-$300 range).

Concrete example: I recently built a Linux video transcoding machine to reencode H.264 BluRay movies to smaller H.264 files [1]. A $110 AMD FX-8300 is able to reencode an average BluRay movie around 47-50 fps, while you would need to spend more than twice that amount of money to match this performance with Intel (the performance of the FX-8300 at this tasks falls between a $200 Intel i5-4590 and a $250 Xeon E3-1231 v3).

[1] Using: avconv -i input.mkv -threads auto -s hd720 -c:v libx264 -c:a libmp3lame -sn -b:v 1400k -b:a 128k -ac 2 output.mkv

View on HN

There's also the power cost, depending on how much you're running that... went from an 8350 to an i7-4790K early last year, the power cost difference is pretty significant.

Even if the AMD chip runs 100 watts hotter and you're encoding constantly all year that's still only $100, small compared to the chip price difference. In datacenters where you have to think about density, cooling, etc then power efficiency gets more important but not so much for consumers except in terms of fan noise.

But over two years, what would the extra $200 get you in terms of CPU options?

> But over two years, what would the extra $200 get you in terms of CPU options?

It gets you to do the math.

The $100 estimate was based on a scenario where the AMD product ran 100W hotter and ran all year long with a heavy workload.

In reality, the comparable offer from Intel runs only 10W hotter, and a normal workload, at best, only costs some cents per year.

After all those years saving money going with the best and cheaper offering from AMD, the money you save gets you more cash to upgrade your computer.

Because economic matter.

But with quicksync you can probably get a multiple of that.

Quicksync is fine for real-time communication or streaming, not for files where you want maximal visual quality with minimal bits used.

It has this bad reputation but in my experience it gives a very decent quality. I may not have stretched the compression ratio too. But at the age of 8TB consumer hard drives...

It gives a completely and quantifiably awful quality for storage, archival, and typical media center purposes. Transcoding at high efficiency is not what Quick Sync is for. It is for streaming, where it is quite fine. This is known.

I'm not repeating cargo cult science; I benchmarked it because I didn't believe it either (which is good! be suspicious!) and it would have been quite handy for a startup I was developing. But alas. Looks like dogshit or costs too many bits.

Be careful not to confuse "looks fine to me" with quantifiable.

I guess people have different degrees of tolerance. But I am looking at anandtech's samples:

http://www.anandtech.com/show/7007/intels-haswell-an-htpc-pe...

and I honestly can't see the difference between the QuickSync and regular H264 samples. I am sure there are some tiny subtleties that I miss and perhaps my screen is not good enough to render the difference, but if we are talking about that level of degradation, and unless you are really a purist, the 10x performance gain is really worth it.

That clip started muddy looking from all that dust. So, it's less obvious.

Reply on news.ycombinator.com