LOL! Sorry...my post got "doubled" and this first one only part of it...

Ahem..."How many Cores/Threads can v15 realistically use?"

Wow. First impression must be off the chart now... (shyly sneaks into a corner...)

^_^

Paul L. Ming

I'm using 16 threads on my Ryzen 7 1800x. Vegas 15 will vary on CPU utilization based on the task you are doing at the time... For instance, Sony XAVC encoding is ridiculously slow and has about 30% CPU utilization, though other tasks come much closer to maxing it out. I know the default thread count in the settings is 32, and if that is the case I'd imagine the max would be even higher, though I'm not sure how much real world benefit there would be. I don't know where the point of diminishing returns would be at.

I think it is more important to have a well balanced system. There is also no point having 20 cores running at 2.5GHz as even a 6 core at 4.5GHz and higher will outperform it. Similar, putting a low-end, just meet the specs GPU into a system with a powerful CPU doesn't make sense. RAM that is clocked higher than the CPU's memory controller is rated for, is also a waste. You are better off running higher spec RAM at a the max memory controller speed but with a lower latency.

Hiya!

Yeah, I was more specific in my original post that got "eaten and puked out as 'Hiya!'"...so I'll add a bit more. :)

I mostly do 3D stuff (mostly using Silo, MODO, Softimage|XSI, ZBrush, Substance, and Lightwave). With those, having a lot of Cores/Threads is generally better than having a higher speed (the Xeon 2699's I have are at a whopping 2.2ghz with 3.6ghz turbo).

That said...where I live there isn't a lot of call for 3D artists...so all I do for that is freelance stuff, mostly for indie games and whatnot. There is a surprisingly big video/film call for up here (Yukon)...with documentaries, small indie films, short-films, and "flyby renderings of potential architecture" things (like visualization for placement of a mining camp, for example). So...I'm adding "video editing and compositing" to my list of skills. The video stuff is very much "seasonal", as in summer, which is about 3 months long...so pretty much go Go GO!

And here we are. :) After doing more research and reading the EXCELLENT replies herein, I think I'll most likely do all my Video stuff on the Ryzen system. I'll leave the heavy-rendering (3D) to the Dual Xeon system. In reality, I'll likely try it one both systems and see which one performs better...I suspect the Ryzen.

Thanks again everyone! :)

^_^

Paul L. Ming

I think it is more important to have a well balanced system. There is also no point having 20 cores running at 2.5GHz as even a 6 core at 4.5GHz and higher will outperform it. Similar, putting a low-end, just meet the specs GPU into a system with a powerful CPU doesn't make sense. RAM that is clocked higher than the CPU's memory controller is rated for, is also a waste. You are better off running higher spec RAM at a the max memory controller speed but with a lower latency.

X2

Check out what Intel has to offer as well, as Vegas 15 makes good use of Intel's QSV, which is currently the fastest way to encode AVC video in Vegas 15. Intel chips also do well in Vegas benchmarks that don't utilize QSV, though for the price the Ryzens come close enough. I'd say for the money, a Ryzen 7 CPU with a powerful Nvidia GPU would be the best bet, but get intel if you can afford it. Ryzen 1800x does great in C4D, by the way. Its 16 threads chew through renders pretty quick. Way faster than comparable intel chips. It won't replace the Xeon, but you'll find it useful for network renders.

"which is currently the fastest way to encode AVC video in Vegas 15."

This is just not true. It may be true for specific setups.

There are a lot of variables ... the Intel cpu generation, the Nvidia graphics card generation, and most importantly the Magix avc codec you use to render to, for quality, for performance, speed, etc.

I have found that Nvidia timeline HW acceleration + Nvenc render acceleration is much faster than QSV, but i'd stop short of claiming it the best combination, because on my system the cpu is 3rd generation QSV but the Nvidia card is a more modern generation.

The only way to come to a more definitive conclusion about all of this is to produce a chart/spreadsheet input from as many users as possible, with different graphic cards, cpus etc. It may well not even give a single winner but more a pointer of good choices. It would require a SINGLE standard test similar to the Red Car test (but one thats not saturated by the more powerful systems) and most importantly a very limited number of Magix avc HW accelerated codecs used, 2 or ideally 1.

Different people at the moment are using different systems with different test codecs, so while not useless, its far from definitive. For example on my system, using Magix Qsv and NV HW acceleration, QSV has 7 possible settings, Nvenc has 36 (6x6). Then the next variable is one of 2 for HW acceleration, QSV or Nvidia.

So 7x36x2 = 504 possible different results. And thats not including the CPU only option.

"which is currently the fastest way to encode AVC video in Vegas 15."

This is just not true. It may be true for specific setups.

There are a lot of variables ... the Intel cpu generation, the Nvidia graphics card generation, and most importantly the Magix avc codec you use to render to, for quality, for performance, speed, etc.

I have found that Nvidia timeline HW acceleration + Nvenc render acceleration is much faster than QSV, but i'd stop short of claiming it the best combination, because on my system the cpu is 3rd generation QSV but the Nvidia card is a more modern generation.

The only way to come to a more definitive conclusion about all of this is to produce a chart/spreadsheet input from as many users as possible, with different graphic cards, cpus etc. It may well not even give a single winner but more a pointer of good choices. It would require a SINGLE standard test similar to the Red Car test (but one thats not saturated by the more powerful systems) and most importantly a very limited number of Magix avc HW accelerated codecs used, 2 or ideally 1.

Different people at the moment are using different systems with different test codecs, so while not useless, its far from definitive. For example on my system, using Magix Qsv and NV HW acceleration, QSV has 7 possible settings, Nvenc has 36 (6x6). Then the next variable is one of 2 for HW acceleration, QSV or Nvidia.

So 7x36x2 = 504 possible different results. And thats not including the CPU only option.

very true.

Every test I've seen done so far, QSV beat NVENC even on the 10 series GTX cards. As soon as I have the time I'll organize a spreadsheet for benchmarks, but I'm just reporting my non-scientific observations.