Author Topic: I tried a Mac for video editing...  (Read 172177 times)

0 Members and 1 Guest are viewing this topic.

Offline Zad

  • Super Contributor
  • ***
  • Posts: 1013
  • Country: gb
    • Digital Wizardry, Analogue Alchemy, Software Sorcery
Re: I tried a Mac for video editing...
« Reply #175 on: April 09, 2013, 11:49:07 pm »
To those people trying to push their hobby-horse of AMD processors; They make great low budget processors. Of all the people I know who produce YouTube videos for a living, none of them use AMD. They just aren't in the game when it comes to performance.

I'm sure you will say "never heard of them" but Yogscast use i5-3570K and they kick out several videos a day. John Bain (TotalBiscuit) uses i7-3930K, mainly so his recording doesn't slow down whatever software he is reviewing, and he really doesn't mess about with his tech (he uses twin Titan video cards). Xeons are robust, but intended for corporate level machines with huge up-times and high reliability. You won't get great performance for your money. They are the 9-litre diesel truck engines of the CPU world.

If you do nothing else, then please, fit a decent heatsink/fan. The standard Intel HSF that came with my i7 ramped up to 100% on a fairly modest load, and rivals a Rolls Royce Olympus for noise. With all 4 real plus 4 virtual cores running at full capacity, it got hotter than I was comfortable with using (>80C) in a typical English summer. I imagine Aus would be 10C hotter, maybe even more in a closed office. With a decent heatpipe heatsink, I can have the fan off at 45C, or have it just moving the air through and drop the CPU temp to 40C. Really hammering the CPU for long periods gets it up to 70C or so, which is reasonable.

Intel X-suffix "extreme" processors are for those with more money than they know what to do with. A tiny bit faster for exponentially more money. Equally, using performance memory will maybe reduce render times by a small number of percent. One SSD is fine, but I would suggest using two hard drives, one for source (input) video, and another for destination (output). The YouTube pros that I know, tend to FRAPS and record onto external USB3 drives.

Motherboards. Almost all use the same standard Intel chipset, and they are mostly pretty close to the reference design. The main thing that changes is fancy little heatsinks, coloured LEDs, pretty PCB solder mask colours and so on. Yes, Xeon based servers do have different chipsets.

I was going to say that the PSU specified in your quote was somewhat under-powered, but then I realised it won't have to power a GPU, so it is probably okay. I think mine is 650W or so, but I run a GTX670OC video card, which takes a fair amount of juice.


Online mariush

  • Super Contributor
  • ***
  • Posts: 5017
  • Country: ro
  • .
Re: I tried a Mac for video editing...
« Reply #176 on: April 10, 2013, 12:33:17 am »
The fact that you say that "they make great low budget processors" and "they just aren't in the game when it comes to performance" tells me you're just an Intel fanboy that probably has no actual clue about where some architectures do better or worse, what's more suited for some things or for others.

You're comparing apples to oranges, Zad. 

Dave isn't doing the same thing those guys do with the computers.

Dave needs a processor that's good at one thing: combining several pieces of video together with minimal video effects (adding a subtitle, some arrows, whatever) and compressing the content.

For this, any processor that works good with x264 (the video encoder application) is good, and the latest AMD processors are very good at encoding, as they have eight good cores.  Where they lack on processing power per core, they make up on the number of cores.

Those Youtube guys have other problems.

Video games are still relatively optimized to run on few processor cores, 1 to 3-4 cores, but usually most only use 2 cores. Intel processors have more processing power per core and they do even more stuff when going in turbo boost mode (overclocking themselves), so for those Youtube guys it makes sense to use Intel processors because the games will run a bit better. 
This will change with the future consoles that will use lots of small cpu cores mixed with a good gpu, the game developers will have incentives to optimize their games for more cores.

Next, the videos those guys make are just different than Dave's videos: they often have picture in picture, transitions, fadin from a clip to another, people talking in front of green/blue screen (the green or blue being replaced by game footage), so we're talking here about layers of video one over another. This kind of stuff is usually passed to the video card because it's so easy to treat the video frames from all those clips as textures and play with them in the video card, and that's where rendering applications start to use Cuda or OpenCL.

Dave doesn't do anything of this. He just needs something that can decode input videos fast, add minimal stuff over the image, maybe resize it and then compress it.

The latest AMD processors are fast and cheap.

Intel processors are fast, but much more expensive to reach the same raw encoding power as the newest AMD processors.  They do offer Quicksync which speeds up video decoding and encoding, but on the other hand with the money you'd save by going with an AMD system, you can get a powerful video card that brings more performance improvements than Quicksync AND the extra processing power of the Intel processors.

If you're on a budget, it's all about making tradeoffs. 
Lose 10-15% brute processing power by giving up i7 3930k  but save 450$  (560$ intel cpu vs 200$ amd cpu , 180$ intel mb vs 130$ amd mb)
Lose a bit of decoding speed by dropping quicksync but gain much more by spending 150-200$ on a video card that's supported by the encoding application.
If you'd have the money, you'd go for both but maybe you don't have the money or it just doesn't make sense to throw that much money for minimal performance increases.

In addition, those Youtube guys don't have money problems.  They can very well spend 4000$ on a computer, getting parts as donation from companies, making money from Youtube and all that. As far as I know, Dave doesn't make as much money as those guys and he also has a family to take care of.

Quote
Intel X-suffix "extreme" processors are for those with more money than they know what to do with.


Those processors have a valid market. There are software programs out there from  SAP, Oracle, Enterprise CRMs, fluid dynamics, automobile stuff, document processing etc that are licensed per CPU SOCKET , or even worse, PER CPU CORE.

People that pay 20-40.000$ and more a year for a license to run that software on one CPU socket, won't care that the server costs 5000-8000$. Those guys want the most performance on each core, the most throughput, the most processing power, whatever you want to call it. Intel is just better at doing more things on each core compared to AMD, that's how the architecture is right now.
There were times in the past AMD was better than Intel at processing power on cores, now it just happens Intel does better. The guys save more money keeping the number of cores low or the number of servers low, compared with going with multiple servers or multiple cpus in a system.

There's uses and uses.. AMD server processors are for example quite popular among people doing virtualization and software testing.. grab 4x12 core AMD opterons, a quad socket G34 board, 64 ram sticks and you have tons of virtualized systems for about 2-3000$. You can't match that with Intel systems.
« Last Edit: April 10, 2013, 12:41:44 am by mariush »
 

Offline EEVblogTopic starter

  • Administrator
  • *****
  • Posts: 37730
  • Country: au
    • EEVblog
Re: I tried a Mac for video editing...
« Reply #177 on: April 10, 2013, 12:48:49 am »
BTW, one thing I didn't think of. Power consumption.
This machine may stay on most of the day (overnight if it's uploading etc)
I don't really want a huge video card if it's going to take an extra 200W all the time or whatever.
My current i7 notebook takes about 37W in normal operation + an external monitor at about the same that I can switch off.
 

Online mariush

  • Super Contributor
  • ***
  • Posts: 5017
  • Country: ro
  • .
Re: I tried a Mac for video editing...
« Reply #178 on: April 10, 2013, 12:55:12 am »
It's not an issue. Both processors and video cards use little power on idle nowadays.
Intel processors use less power, but the difference is something like 10-15w less when idle, 30-40w less when under load.
Video cards also use very little power when idle, they generally use 10-30w when in Windows.

The system you're targeting is generally going to be a bit over 100 watts when idle, probably 200-250w if you're encoding with video card and cpu to the maximum.
 

Offline ve7xen

  • Super Contributor
  • ***
  • Posts: 1192
  • Country: ca
    • VE7XEN Blog
Re: I tried a Mac for video editing...
« Reply #179 on: April 10, 2013, 12:59:25 am »
It won't be that significant when idle, but there's not much point in it either. I don't think it will substantially improve your rendering performance.

I still think you can get your 2x for free by spending some time figuring out frameserver and avisynth/megui :)
73 de VE7XEN
He/Him
 

Offline EEVblogTopic starter

  • Administrator
  • *****
  • Posts: 37730
  • Country: au
    • EEVblog
Re: I tried a Mac for video editing...
« Reply #180 on: April 10, 2013, 01:03:45 am »
I still think you can get your 2x for free by spending some time figuring out frameserver and avisynth/megui :)

Great, then I'll get 4 times improvement!  ;D
 

Offline ddavidebor

  • Super Contributor
  • ***
  • Posts: 1190
  • Country: gb
    • Smartbox AT
I tried a Mac for video editing...
« Reply #181 on: April 10, 2013, 05:04:25 am »


Dave, what model is your camera?
David - Professional Engineer - Medical Devices and Tablet Computers at Smartbox AT
Side businesses: Altium Industry Expert writer, http://fermium.ltd.uk (Scientific Equiment), http://chinesecleavers.co.uk (Cutlery),
 

Offline EEVblogTopic starter

  • Administrator
  • *****
  • Posts: 37730
  • Country: au
    • EEVblog
Re: I tried a Mac for video editing...
« Reply #182 on: April 10, 2013, 05:39:56 am »
Dave, what model is your camera?

Canon HF G10.
 

Offline Flávio V

  • Regular Contributor
  • *
  • Posts: 62
  • Country: pt
  • Capacitor lover
Re: I tried a Mac for video editing...
« Reply #183 on: April 10, 2013, 08:36:23 pm »
I like to think of those 2 CPU manufacturers comparing to 2 famous multimeter manufacturers....

Intel is more like fluke and AMD more like agilent(more bang per $)


I know a person than acts like a intel fanboy than some days ago  said something than i was not expecting, he said that for the price of the best AMD CPU you only get a bad quad core(i5)intel CPU....and 8 cores at near 4ghz and much more cache is better than like 4 cores at 2.2ghz and something like 4mb of cache...
 

Offline ConnorGames

  • Regular Contributor
  • *
  • Posts: 97
Re: I tried a Mac for video editing...
« Reply #184 on: April 10, 2013, 09:26:04 pm »
Also, do note that GPU encoding can't match x.264 in therms of quality/MB. Another option would be to get several last-gen CPUS used and stick them in a server board: I think handbrake is quite well threaded. I have a friend who got 4 2.2Gh quad core AMD processors on ebay for $10/ea, and a server board for about $80-90. Overall the machine with 16 2.2Ghz cores cost <$500. I'll see if he can run your benchmark on it!
 

Online mariush

  • Super Contributor
  • ***
  • Posts: 5017
  • Country: ro
  • .
Re: I tried a Mac for video editing...
« Reply #185 on: April 10, 2013, 09:42:24 pm »
I don't know what Flavio V is smoking, he doesn't make sense. Or at least I agree with him, but I have the feeling he wanted to say something else.

As for ConnorGames, you're right that in general software using video cards exclusively to encode content produce lower quality content, those software apps are more suitable for quickly producing some SD content to upload to youtube, not something super high quality.

But you don't have to be use the video cards to ENCODE content from start to finish, video editing applications can use the video card to render the actual frames in the video (do overlays, apply subtitles etc) and do some partial encoding stuff like motion analysis then sending this data to the software encoder to further process the rendered frames.

The idea to buy a server grade motherboard and 4 server processors from eBay is nice, you can indeed get stuff relatively cheap.

But, unless you're going to keep the system in a cardboard box, you'll find out you'll pay quite a lot on a server case and for a power supply that has 2 x 8 pin connectors for processors (if not more). Also, each processor has its own memory pool so you'd have to buy at least 4 memory modules which have to be registered or unregistered+ecc ddr3 memory and some motherboards won't even boot without 2 memory modules on each cpu. That's a lot of additional costs.

Such system won't do quite as well as a current desktop system using a processor like FX-8320 or FX-8350 because of memory latency and having to coordinate the work of each thread between individual processors - the caches have to be kept in sync, data has to be moved from a processor's memory bank to another processor's memory bank and so on. x264 is using the cpu caches quite well.

x264 is also heavily optimized and uses pretty much any cpu extension it can get its hands on.. sse3, sse4, xop, avx, fma, you name it, if it brings improvements it uses it.  Those opteron server cpus don't have a lot of these extensions, they're really kind of weak in that regard.

So it's quite possible (I'm almost sure) a quad cpu , quad core opteron won't encode a 1080p as fast as the eight core  FX-8350 or the equivalent Intel processor today, or if it does it will be within 10-15%, certainly too little to be worth spending so much money and electricity.
« Last Edit: April 11, 2013, 02:04:49 am by mariush »
 

Offline NiHaoMike

  • Super Contributor
  • ***
  • Posts: 9008
  • Country: us
  • "Don't turn it on - Take it apart!"
    • Facebook Page
Re: I tried a Mac for video editing...
« Reply #186 on: April 11, 2013, 01:06:14 am »
It's not an issue. Both processors and video cards use little power on idle nowadays.
Intel processors use less power, but the difference is something like 10-15w less when idle, 30-40w less when under load.
Video cards also use very little power when idle, they generally use 10-30w when in Windows.

The system you're targeting is generally going to be a bit over 100 watts when idle, probably 200-250w if you're encoding with video card and cpu to the maximum.
My GTX560 TI will downclock all the way to 50MHz when idle. And all modern x86 CPUs not only downclock but also turn off unused cores. (I'm pretty sure that's also the case for GPUs but I don't know for certain.)

I read somewhere about an i7 system that only used 6W when idle. But that was with a relatively slow mobile i7, a SSD, and integrated graphics, which makes it unsuitable for video editing or gaming.
Cryptocurrency has taught me to love math and at the same time be baffled by it.

Cryptocurrency lesson 0: Altcoins and Bitcoin are not the same thing.
 

Offline EEVblogTopic starter

  • Administrator
  • *****
  • Posts: 37730
  • Country: au
    • EEVblog
Re: I tried a Mac for video editing...
« Reply #187 on: April 11, 2013, 01:25:25 am »
As for ConnorGames, you're right that in general software using video cards exclusively to encode content produce lower quality content, those software apps are more suitable for quickly producing some SD content to upload to youtube, not something super high quality.

Yes, I'm starting to think the only thing that matters for my use is raw CPU power and number of cores.
Come to think of it, I'd rather spend money on good silent fans and a sound damped case than some trotted up GPU.
 

Offline ddavidebor

  • Super Contributor
  • ***
  • Posts: 1190
  • Country: gb
    • Smartbox AT
I tried a Mac for video editing...
« Reply #188 on: April 11, 2013, 05:10:04 am »
Nvidia sometimes ago has a s system to parallelize a lot of video card for have more power. Anyone remember it?
David - Professional Engineer - Medical Devices and Tablet Computers at Smartbox AT
Side businesses: Altium Industry Expert writer, http://fermium.ltd.uk (Scientific Equiment), http://chinesecleavers.co.uk (Cutlery),
 

Offline EEVblogTopic starter

  • Administrator
  • *****
  • Posts: 37730
  • Country: au
    • EEVblog
Re: I tried a Mac for video editing...
« Reply #189 on: April 11, 2013, 09:16:02 am »
I shot the latest video in full HD 1920x1080 at 17Mbps instead of 1440x1080 at 12Mbps
Rendering using Sony is taking about twice as long for 25% more pixels.
Don't know about Handbrake yet...
 

Offline EEVblogTopic starter

  • Administrator
  • *****
  • Posts: 37730
  • Country: au
    • EEVblog
Re: I tried a Mac for video editing...
« Reply #190 on: April 11, 2013, 09:53:15 am »
Handbrake is probably about the same speed, maybe a bit slower if anything. Hard to tell with the varying content.
So all the more reason to get a new tricked out machine if I'm going to keep shooting 17Mbps footage.
 

Offline M. András

  • Super Contributor
  • ***
  • Posts: 1014
  • Country: hu
Re: I tried a Mac for video editing...
« Reply #191 on: April 11, 2013, 04:48:11 pm »
for silent fans i highly recommend the arctic cooling f pwm series or even the co (continous op duall ball bearing) im using these for years and apart from the noise of the high airflow on certain obstackles really silent moves decent amount of air 125m3/h they dont mention static pressure but it feels high
 

Offline Nobody2

  • Contributor
  • Posts: 17
  • Country: de
Re: I tried a Mac for video editing...
« Reply #192 on: April 11, 2013, 07:01:11 pm »
Yes, I'm starting to think the only thing that matters for my use is raw CPU power and number of cores.
Come to think of it, I'd rather spend money on good silent fans and a sound damped case than some trotted up GPU.
I think in that case the FX-8350 looks like a good choice:

However those were basically the only benchmarks were the AMD shined. I wouldn't recommend them for a general purpose PC. If they were affordable a 12 or 16 core Opteron should be even better.

On a side node, do you have Raspberry Pi Dave? Why not use that to upload over night and switch the power hungry computers off?

If your Youtube upload is limited, but the one to your sever not, maybe you can get your server to upload your videos to Youtube?
 

Offline HooRide

  • Regular Contributor
  • *
  • Posts: 72
  • Country: us
Re: I tried a Mac for video editing...
« Reply #193 on: April 12, 2013, 12:23:15 am »
*this may be repeat info since I didn't read all 13 pages before this post*

What a terrible experience but I am not surprised. Most Apple store employees have no training on 'Pro' software such as Final Cut. Also, what you were told about the Mac Pro being outdated is completely true. There are very solid rumors of a new model coming within a few months time.
Please enter all complaints legibly in this space -> [  ]
 

Offline firewalker

  • Super Contributor
  • ***
  • Posts: 2450
  • Country: gr
Re: I tried a Mac for video editing...
« Reply #194 on: April 12, 2013, 08:23:41 am »
Become a realist, stay a dreamer.

 

Offline M. András

  • Super Contributor
  • ***
  • Posts: 1014
  • Country: hu
Re: I tried a Mac for video editing...
« Reply #195 on: April 12, 2013, 08:04:09 pm »
ran trought your sample file on default settings in handbrake qfactor set to 28, its done in 58sec, and only hit up to 70% cpu target was my system ssd which max out at 60mb/s write
 

Online hans

  • Super Contributor
  • ***
  • Posts: 1637
  • Country: nl
Re: I tried a Mac for video editing...
« Reply #196 on: April 12, 2013, 09:07:50 pm »
Yes, I'm starting to think the only thing that matters for my use is raw CPU power and number of cores.
Come to think of it, I'd rather spend money on good silent fans and a sound damped case than some trotted up GPU.
I think in that case the FX-8350 looks like a good choice:

However those were basically the only benchmarks were the AMD shined. I wouldn't recommend them for a general purpose PC. If they were affordable a 12 or 16 core Opteron should be even better.

On a side node, do you have Raspberry Pi Dave? Why not use that to upload over night and switch the power hungry computers off?

If your Youtube upload is limited, but the one to your sever not, maybe you can get your server to upload your videos to Youtube?
As test samples and methods vary, for example this test favors the Intel chips:
http://us.hardware.info/reviews/3314/13/amd-fx-8350--8320--6300-vishera-review-finally-good-enough-benchmarks-cpu-720p-mpeg-to-x264-video-encoding

However on pass 2 it can clearly keep up with the i7 3770K

Or:
http://us.hardware.info/reviews/3314/14/amd-fx-8350--8320--6300-vishera-review-finally-good-enough-benchmarks-cpu-cyberlink-mediashow-espresso-60

Or in daily tasks like packaging: WinRar 317MB of data; a Intel i7 3770k is 10 seconds faster in this test.. That's 30%, ouch.

Moreover: the CPU may be a lot cheaper, the motherboards with 'premium features' tend to be rarer. I searched for a top brand (in my mind Asus) motherboard with top-end chipset (intel Z77 / AMD 990FX) and4 RAM slots. The cheapest Intel board was 80 euro, the AMD's was 120 euro's. Add the cost a dedicated GPU that supports stuff like triple monitors for ~60 - 80 euro's, you're already past the price difference.

Power consumption:
http://us.hardware.info/reviews/3314/22/amd-fx-8350--8320--6300-vishera-review-finally-good-enough-energy-consumption-cpu-idle--cinebench-115
The Intel chips are just a lot more power efficient. 13W difference is a lot: 13*365*24 = 113kWh per year savings@ 24/7 
The AMD chip also has a higher TDP, which means it will run hotter @ 100% load and therefore require more cooling (thus more noise).

As to respond to Dave's config (if he hasn't ordered it already..)

Intel i7 3770K seems good.
Asrock motherboard; should work I guess ;)
RAM: any 2x8GB set. I would avoid 4x4GB, even if it's a few $ cheaper. I always find that if a machine gets older a RAM upgrade is desired. With full slots it's not cheap to do because you have to buy 2 new sets instead of 1.
Power supply: 600W sounds a lot for such a system with a MB, a big CPU and a couple of hard drives.  Typically dedicated GPU's draw a lot of power (150W+) under gamign.
Case: usually a personal taste. I personally own a Fractal Design R3, which I like quite a lot. I see Scorptec has the R4:
http://www.scorptec.com.au/computer/47245-fd-r4b
It does have 3+ USB ports (including USB3.0) on the front, with audio and such. More importantly it has sound dampening stuff on both sides/bottom/front and top and also on unpopulated fan slots (removable sound padding slots). However, it is quite a bit more expensive than the BitFenix one, and installation can be a bit fiddly (especially on the R3 model, believe it is improved on R4 though).
Optionally: dedicated CPU cooler. Benchmark (Efficiency = sound x temperature, lower = better)
I always hate boxed coolers. Unfortunately, Scorptec doens't seem to have much well known 3rd-party CPU cooler brands.
I have got 1 back,1 top and 2 front fans installed and connected to my Asus P8Z77-V PRO motherboard, which has temperature fan curves. This means on idle the CPU fan spins and front fans very slightly and when things get more hot extra (much louder) fans kick in for additional cooling. Result is very silent system with reasonable idle temps yet when things get more serious it keeps everything below 60C.
« Last Edit: April 12, 2013, 09:09:46 pm by hans »
 

Online mariush

  • Super Contributor
  • ***
  • Posts: 5017
  • Country: ro
  • .
Re: I tried a Mac for video editing...
« Reply #197 on: April 12, 2013, 10:04:48 pm »

As test samples and methods vary, for example this test favors the Intel chips:
http://us.hardware.info/reviews/3314/13/amd-fx-8350--8320--6300-vishera-review-finally-good-enough-benchmarks-cpu-720p-mpeg-to-x264-video-encoding

However on pass 2 it can clearly keep up with the i7 3770K

The devil's in the details. First pass in x264 is a very light one where very basic math is done, generally just to determine which time intervals would require more bits during the actual compression. During first pass, x264 uses around 4 threads and generally the cpu hovers at around 40-60% ... the first pass doesn't do things that can actually be parallelized efficiently.  Intel processors win because they have more processing power per core and some instructions require less cpu cycles. The faster those cores finish job, the faster they switch to decoding the next input frame and process it...

x264 and most video encoding software is integer only, and FX processors have eight real integer cores, so once the actual encoding is started in 2nd pass, you can see the Vishera processors (fx-8320, fx-8350 etc) can keep up with 3770k that's more expensive and run at higher frequency.

And we're coming back to flawed tests ... see on the graph what they tested the processors with .. a 89 MB VOB file from a MPEG-2 video.. at 10-20mbps data rate, that's basically a 2 minute video.

x264 splits each frame in several small chunks and passes those chunks to threads for motion analysis and various things... with a 720x480 video (assuming it's a NTSC mpeg-2 video), the eight cores can't work efficiently. Basically AMD processors are dragged back because again, it's not much stuff done in parallel.  If it was about a proper test, they'd encode an 1080i mpeg-2 stream from TV and see the difference.

With a 2 minute SD video, you're kind of testing the memory latency, not the cpu, x264 won't even fill the caches and memory and take advantage of the CPUs with such a source file.

Quote
Or in daily tasks like packaging: WinRar 317MB of data; a Intel i7 3770k is 10 seconds faster in this test.. That's 30%, ouch.

This is again kind of a flawed test. It just shows again the Intel's processing power per core, Intel chips just finish work faster on individual cores. Winrar 3.93 is not multithreaded -well, it kinda is, in the sense that it uses 2-3 cores - , the software basically did compression on 1 or 2 cores, with the rest sitting idle. This is especially bad for AMD because with just 1-2 cores running, the Intel processors can enable turbo mode and overclock the cores to 4Ghz+

How can you test eight core AMD processors with 4 core processors with a single threaded compressor? And more importantly, how would you consider the test valid, when the software you test with was released around March 2010, and obviously doesn't use any of the functions the modern processors use (it was probably optimized for core 2 duo / core 2 quad as that was popular around 2010)

Winrar 4 and up implemented proper threading and parallel compression but even then you have fineprint.  Unfortunately can not really reach the maximum speed with rar archives so small and can't efficiently keep 8 amd cores busy with small files (that would be present in 317 MB of data). Their test case is ridiculously small with just 300 Megs of data, for a proper test about 8 GB of data should be used and with some files larger than 2-400 MB a piece.

This is clearly written in the implementation, but nobody bothers to read it: http://www.rarlab.com/rarnew.htm

Quote
   Version 4.20
  1. Changes in RAR compression:
      a) RAR general compression algorithm is optimized for better utilization of several processor cores. While some speed gain is possible even in single processor mode, best results are achieved in multi-core environment.
         Speed gain depends on data type and average file size. Several cores are utilized more efficiently when compressing large files.
[..]    c) RAR text compression algorithm cannot utilize several CPU cores efficiently, so its performance in multiprocessor environment is much lower than for general algorithm. Also its decompression speed is much lower than in general algorithm regardless of CPU number. So we decided to disable the text algorithm by default.
         If you need maximum possible compression ratio for plain text data regardless of speed, you can enable the text compression    in "Advanced compression parameter" dialog. [...]
      d) "Fastest" (-m1) compression mode also supports several processor cores now. In 4.11 it could use only a single processor core.

If you  want to test, test with 7zip.. it's open source and uses up to 8-16 cores : http://www.anandtech.com/show/6396/the-vishera-review-amd-fx8350-fx8320-fx6300-and-fx4300-tested/2


Quote
Moreover: the CPU may be a lot cheaper, the motherboards with 'premium features' tend to be rarer. I searched for a top brand (in my mind Asus) motherboard with top-end chipset (intel Z77 / AMD 990FX) and4 RAM slots. The cheapest Intel board was 80 euro, the AMD's was 120 euro's. Add the cost a dedicated GPU that supports stuff like triple monitors for ~60 - 80 euro's, you're already past the price difference.

My Gigabyte GA-990FXA-UD3 is about 140$ or about 100-110 euro, but I assure you there's really much less expensive boards with this chipset (990FX). I chose this higher end model because it has proper heatsinks on the VRM section, 8+2 vrm, pci slots etc.

The cheapest, proper AMD boards with a good chipset and good southbridge (SB950) is about 80$, for example : http://www.newegg.com/Product/Product.aspx?Item=N82E16813128553  or  http://www.newegg.com/Product/Product.aspx?Item=N82E16813157280

If you go back to your Z77 boards, and actually check the specs of those cheap motherboards, you may find out those cheap models are really bad when it comes to VRM, so little potential to overclock your CPU  (but I admit the cheaper AMD boards would also be basic when it comes to vrm)... and the cheapest Asus with z77 chipset on Newegg is 110$ : http://www.newegg.com/Product/Product.aspx?Item=N82E16813131965

But basically even the Z75 based Asrock at 85$ would be reasonable, so there's really not much price difference: http://www.newegg.com/Product/Product.aspx?Item=N82E16813157304

You also mentioned :

Quote
The AMD chip also has a higher TDP, which means it will run hotter @ 100% load and therefore require more cooling (thus more noise).

The stock Intel cooler is awful, like i already said in this thread. It's just aluminum heatsink with copper core.  The AMD stock cooler is decent, with 2 heatpipes, it's less noisy. I would say with the stock coolers, the AMD would be less annoying. But I will admit both could use a third party fan if you care about noise.

 

Offline EEVblogTopic starter

  • Administrator
  • *****
  • Posts: 37730
  • Country: au
    • EEVblog
Re: I tried a Mac for video editing...
« Reply #198 on: April 12, 2013, 10:30:25 pm »
Thanks guys, great input.
So it seems that at best, the AMD is as good as the 3770K, but essentially not any better. But the AMD is a bit cheaper, but it's higher power.
I think the Intel is probably the best bet.

Any good alternatives to Scorptec, or should I just go with them?
I like the sound of that silent case, no pun intended. My i7 notebook is really annoyingly loud when rendering.
 

Offline EEVblogTopic starter

  • Administrator
  • *****
  • Posts: 37730
  • Country: au
    • EEVblog
Re: I tried a Mac for video editing...
« Reply #199 on: April 12, 2013, 10:48:30 pm »
On a side node, do you have Raspberry Pi Dave? Why not use that to upload over night and switch the power hungry computers off?

Interesting idea!
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf