From: "Teddy Gage" Received: from mail-ot1-f49.google.com ([209.85.210.49] verified) by media-motion.tv (CommuniGate Pro SMTP 6.1.0) with ESMTPS id 7242688 for AE-List@media-motion.tv; Mon, 11 Feb 2019 15:25:08 +0100 Received: by mail-ot1-f49.google.com with SMTP id m1so6020363otf.5 for ; Mon, 11 Feb 2019 06:35:28 -0800 (PST) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=gmail.com; s=20161025; h=mime-version:references:in-reply-to:from:date:message-id:subject:to; bh=GQGollL5B4EhAz7vX6HraYgFt8QbPBL2SQaxsMTbV8E=; b=ri8ykVnHhuUwCl/VapoEnN78hnynGqNZl0QTCZda8uQsS3EBI3W5GQTWoeZle6vC4e E3yFka71Edn+ZkElBlFL3/OeHj+aZZhzryPX8gFfybuOY4AGzyqxuMl3rIBI9IwuQrOE QA+EUYCF7n4CbeZzcqnCar18fQSXLETtxn4rHOp6Rm74NjMUElRqb1UU3ZRevG0AZ6vU STui+B065skQisE9qfEiynUva/3W0+PFfdF4QGmP7XsAOt6YW+M0Ouuc5jAuz4sceT12 DYveDl8sRkNrju4UjZLbwKEhJ83+OaLG5oecnXL6+E8cWjyTgp0XrlifsE2Uc5jlEV4Z fYSg== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20161025; h=x-gm-message-state:mime-version:references:in-reply-to:from:date :message-id:subject:to; bh=GQGollL5B4EhAz7vX6HraYgFt8QbPBL2SQaxsMTbV8E=; b=lINawQ+sZSSzM2ZzWzBV33zk/nJvLhrycF8KoURhcl8iVFzyI+uGMk9a7Pq/RotJBX DjgQBQJExhOBktSY1XjCWwgAafDMoQq5kWI/2ypWgY1GQcQwTaz1JlLmkGdrbPJHqw8C s87q6wHxhxUkwHBPoZ0WT0QBrH8G3h3iGrd5b39gNeRquCUBb8+/kRk+wd8AmKZUTEmp 7t+bS7Ga8ZhHvroU2BZ5W7oFViGtRq/3dHKcMZQK9aCrn6dEKayq4B3LBxemVQQniYzJ L9wYKKTo+lMGprq6O1rLwDUwakuFAg3dw6ZyHyrrZZHEM92lJfrB6U2FnXHPNuHRj2hu UXWg== X-Gm-Message-State: AHQUAuZHETsrfmskb68vTRE/2Szsu1kHiUddyZg/MB3rbdtjC+UsHQse Ex4PZYPKKPjruWbhEv8g6mwz6ZRiVGs47NkrABd9Lg== X-Google-Smtp-Source: AHgI3IY4EbTozHqRKMHmrXIINluQyDUOS7x5D2Rb4MgQ07ZGlHW6uAU+cU0UcrSUFo0ESU+GBPc5o2SX9UkWGng81zY= X-Received: by 2002:a9d:4505:: with SMTP id w5mr2340524ote.314.1549895725721; Mon, 11 Feb 2019 06:35:25 -0800 (PST) MIME-Version: 1.0 References: In-Reply-To: Date: Mon, 11 Feb 2019 09:35:13 -0500 Message-ID: Subject: Re: [AE] Feedback on Workstation components To: After Effects Mail List Content-Type: multipart/alternative; boundary="0000000000002bad1305819f3815" --0000000000002bad1305819f3815 Content-Type: text/plain; charset="UTF-8" Content-Transfer-Encoding: quoted-printable Excellent breakdown Carey- While I absolutely applaud Puget=E2=80=99s attem= pts to provide benchmarks for real world tasks, I have issues with their methodology and don=E2=80=99t agree with all their results. Still better to= study toms hardware or Anandtech benches directly for various components, and take the Pudget results with a grain of salt. They are still trying to sell you something. - Teddy On Mon, Feb 11, 2019 at 9:28 AM Carey Dissmore wrote: > A few comments: > > Puget Systems REALLY gets it when it comes to addressing the needs of our > market. They have a model where they do a ton of research and provide > excellent support for their tested, configured system designs. They share= a > lot of their data and findings on their website, which benefits the > community as a whole. They are uniquely positioned to do this type of > testing and I=E2=80=99m very thankful they share. For the vast majority o= f people > who do NOT have interest in building their own systems, but nevertheless > are realizing that a custom workstation is truly the best path forward fo= r > their needs, I can=E2=80=99t think of a better choice. > > Now onto the geek stuff, (which Puget already knows): > > A - Building PC: I=E2=80=99m inclined towards PC building too, and have b= uilt many > workstation machines. I built my first PC around 2001 mostly for gaming a= nd > offloaded encoding/rendering but I largely stuck with Mac for post > production work until about 2012 when I made the complete jump to primary > post production on Windows workstations. Teddy and I had a good chat on t= he > phone back then just to gut check my system architecture before pulling t= he > trigger on a $6k box build. I=E2=80=99ve built and rebuilt many since. > > B-Performance: It used to be that synthetic benchmarks were quite useful > in determining performance as it applied to video post work. You pretty > much always needed the highest performance hardware you could get and the= re > wasn=E2=80=99t a whole lot of differentiation about tasks, as it all sort= of lifted > with the general performance of the machine. You could look at relative > performance and extrapolate your value or expected performance for the ap= ps > and work you did and make a decision about value for money/ROI/Cost-benef= it > with some simple math. > > Those days are over. That won=E2=80=99t work anymore. The trend for nearl= y a > decade is that PC performance gains have forked in multiple directions: > CPUs have added cores, and clock speed increases have crept higher but ve= ry > slowly. Due to thermals, you are often left with a choice of more cores v= s > more clock speed. GPUs have introduced massive parallelism tuned for > specific compute loads but data has to be moved from memory/fast CPU acce= ss > over the bus to the GPU and back so that is a factor in determining which > types of =E2=80=9Cloads=E2=80=9D are best suited to GPU compute. The prob= lem is that > synthetic benchmarks can be designed to torture test a specific type of > computing load that shows maximized performance of a particular component > but it has never been less relevant to apply those to your expectations i= n > the apps you ACTUALLY USE to do post production work. > > Therefore, the ONLY benchmarks that are relevant are the ones that are > done by testing the apps you ACTUALLY USE to determine a particular > component or configuration=E2=80=99s value to you. Here=E2=80=99s some ap= ps that I use and > what I have found they need from a machine. > > For example here=E2=80=99s some differing compute loads and what types of= PC > hardware/config will serve it well. > > Storage (applies to all): Everything wants fast storage and IO. Splitting > up storage between media and boot and cache, each optimized is good. > Storage/cacheing has become an issue too and managing the flow of data an= d > using high performance caching devices separate from both system drive an= d > media storage is key to optimizing. As you look at the applications below > > AE: In daily use, largely still single threaded operations with dashes of > multithreading throughout, and a splash of GPU compute here and there, > particularly on plugins/effects. Today, this means favoring higher clock > speeds over core count on the CPU. Get as much RAM as you can justify > because AE and RAM previews will use it. Generally AE-only artists don=E2= =80=99t > need tons of data storage, but what they have should be fast. Often they > are best served by something like a 9900k system and a modest GPU from > NVidia. > Note that one of the challenges of optimizing AE for multithreading is > that not every type of compute task lends itself well to parallelism. > Sometimes you have to wait for the result of one computation to use the > result of that to start the next computation. This same performance > optimization issue also affects other similar apps like Fusion. > > Premiere: A little more balanced between CPU cores and clock speed and > GPU. Having said that Premiere still favors clock speed over massive core > count, can take advantage of Quicksync on newer Intel CPUs BTW, and uses > GPU but does not rely heavily on it and primarily processes color and > effects on GPU (with some exceptions). Premiere doesn=E2=80=99t require q= uite as > much RAM as AE, but I guess it doesn=E2=80=99t hurt. Having said that Pre= miere > seems to need an awful lot of compute hardware thrown at it to smoothly > handle media at 4k. By comparison FCPX (which I don=E2=80=99t use because= it=E2=80=99s Mac > only) is way more optimized and gets better performance from lesser > hardware. I think there=E2=80=99s a lot of room for improvement for Premi= ere here > but it probably means a rewrite. > > Encoding: h264 or now h265 encoding is massively parallel and really show= s > off what a lot of cores can do. Currently CPU dominated we=E2=80=99re sta= rting to > see this moving to GPU options as well. If all you did all day was encode > stuff, go for core count and fast IO. However, for most of us, encoding > occupies only a few minutes per day compared to creative work. > > Resolve: GPU dominated (for color page operations anyway). Here, CPU > performance matters but a mid-line CPU is probably adequate but what you > want are GPU compute units. The best units you can afford, plus more of > them. 2 or 3 GPUs scale really well with Resolve, but in my testing I=E2= =80=99ve > found it best to keep them matched. I did some mismatched testing with > different gen GPUs and it=E2=80=99s pretty variable and inefficient. Not = that every > little bit doesn=E2=80=99t help=E2=80=A6but one card ends up waiting on t= he other to finish > a task. PCI-E lanes required for multi-GPU support push the user to go wi= th > higher level platforms like X299. > > Cinema 4D: When rendering, massive CPU core count is your friend. But whe= n > modeling, key framing, animating etc a lot fo that is more single threade= d > but obviously previews can be run OpenGL (etc) on the GPU. Other renderer= s > are GPU-dominant. 3D is not my forte so I=E2=80=99m going to let others c= over this. > > But I said all of that as a preface to make this point: You can choose an= y > one activity or app and build a kick-butt machine tuned specifically to t= he > requirements of that activity, and walk away with maximized ROI. An AE-on= ly > machine is going to look quite different than a Resolve-only machine, and > again quite different from a C4D optimized machine. But often we need to > engage in a mix of activities and the choice about whether to try to buil= d > the ultimate =E2=80=9Cdo it all=E2=80=9D workstation ($$$$) vs separating= the tasks is > something we all have to look at in light of our own mix of work. > > Anyway, when I sat down I had no idea I=E2=80=99d write ll that but alas,= the > coffee was good and plentiful. Plus, when you really get into this stuff > there are just a lot of considerations to get your head around. By no mea= ns > did I write a definitive guide here (is there such a thing?) and understa= nd > why Chris Zwar=E2=80=99s is over 10,000 words. I look forward to it. > > Carey Dissmore > > > > On Feb 11, 2019, at 6:33 AM, Byron Nash wrote: > > Puget really impressed me with the depth and detail of their hardware > analysis for very targeted demographics. It's nice to have someone really > understand what the difference is between editing, motion graphics and GP= U > rendering. > > On Sun, Feb 10, 2019 at 11:52 PM Teddy Gage > wrote: > >> Sure thing Chris, feel free to reach me at teddygage@gmail.com. Goes for >> anyone- always happy to consult or chat hardware or software. Have also >> been doing a fair amount of VFX supervising recently for a couple TV sho= ws. >> >> On Sun, Feb 10, 2019 at 11:14 PM Chris Zwar >> wrote: >> >>> Hi all, >>> >>> I started researching performance last year, with a view to writing up >>> an article and some thoughts for the ProVideo Coalition. Currently I'm >>> about 50% of the way through my plan and I'm over 10,000 words so it's = a >>> bit slow going. Even so, I was wanting to reach out to list members wh= o >>> are especially interested and knowledgable on performance related matte= rs. >>> >>> Teddy & Stephen were both on my list of people to approach for input an= d >>> this thread just reinforces that! Are you two able let me know the best= way >>> to contact you off-list please? >>> And if there's anyone else who has strong knowledge or strong opinions >>> on performance/ components then let me know too. >>> >>> Cheers, >>> >>> -Chris >>> >>> >>> On Sat, 9 Feb 2019 at 07:18, Byron Nash wrote= : >>> >>>> Those are good thoughts Stephen and I'll keep that in mind going >>>> forward. I am planning to use a Blackmagic card for video output but t= he >>>> monitors I have to use right now are 8bit HD. I'm starting to get into= 4K >>>> VFX for films on my machine but most of the rest of the team is buildi= ng HD >>>> promos/mograph content. My only pet peeve with the BM cards is when yo= u >>>> have a 4K screen it won't downscale through the card and just goes bla= ck. >>>> >>>> On Fri, Feb 8, 2019 at 11:10 AM Stephen van Vuuren < >>>> AE-List@media-motion.tv> wrote: >>>> >>>>> > In regards to 10bit output an option not being discussed is using a >>>>> Blackmagic or Aja PCI card or device. These will give you the option = of >>>>> 10bit YUV or RGB output (monitor dependent). 4K capable versions of t= hese >>>>> are available for as little as $200. >>>>> >>>>> >>>>> >>>>> I mentioned external card but you are not budgeting for the 4K 10-bit >>>>> display that you will need to see the output. >>>>> >>>>> >>>>> >>>>> I have a $199 BM MiniMonitor 4K and it=E2=80=99s great going out to a= 50=E2=80=9D HDR >>>>> TV). *But that does not get you 4K 10-bit on your desktop display or >>>>> apps that don=E2=80=99t support that card*. You have to have a second= monitor >>>>> where as Quadro allows you to have 4K 10-bit on your main desktop mon= itor. >>>>> >>>>> >>>>> >>>>> If you only have room for on 4K 10-bit, a card plus additional monito= r >>>>> is not an option. And even if you do, my setup I would rather always = see >>>>> 10-bit as much as possible in all my work and catch issues. >>>>> >>>>> >>>>> >>>>> As video moves to 10-bit and HDR in 4K, even for low-end, 10-bit is >>>>> going to be a bigger deal. >>>>> >>>>> >>>>> >>>>> *From:* After Effects Mail List >>>>> *Sent:* Friday, February 8, 2019 12:25 AM >>>>> *To:* After Effects Mail List >>>>> *Subject:* Re: [AE] Feedback on Workstation components >>>>> >>>>> >>>>> >>>>> If you=E2=80=99re talking about a budget system Quadro cards aren=E2= =80=99t even in >>>>> the conversation (5k was mentioned). If one of the main uses of the m= achine >>>>> will be Redshift you=E2=80=99re better off buying the fastest GTX or = RTX card(s) >>>>> you can afford. The RTX cards may end up being a better choice in the= near >>>>> future, if Nvidia=E2=80=99s and Redshift=E2=80=99s recent announcemen= ts are to be believed. >>>>> >>>>> >>>>> >>>>> In regards to 10bit output an option not being discussed is using a >>>>> Blackmagic or Aja PCI card or device. These will give you the option = of >>>>> 10bit YUV or RGB output (monitor dependent). 4K capable versions of t= hese >>>>> are available for as little as $200. >>>>> >>>>> >>>>> Dennis Wilkins >>>>> >>>>> >>>>> >>>>> >>>>> >>>>> On Feb 7, 2019, at 8:49 PM, Stephen van Vuuren < >>>>> AE-List@media-motion.tv> wrote: >>>>> >>>>> >>>>> >>>>> Steve, >>>>> >>>>> >>>>> >>>>> That link is for 10-bit DirectX output which is for games and Direct = X >>>>> apps =E2=80=93 not relevant to this discussion. Adobe and most other = pro post apps >>>>> don=E2=80=99t use or support DirectX. And just because a driver or mo= nitor flags >>>>> 10-bit does not mean there is actually 10-bits happening. >>>>> >>>>> >>>>> Better info is here: >>>>> https://imagescience.com.au/knowledge/10-bit-output-support >>>>> >>>>> >>>>> >>>>> *From:* After Effects Mail List >>>>> *Sent:* Thursday, February 7, 2019 11:40 PM >>>>> *To:* After Effects Mail List >>>>> *Subject:* Re: [AE] Feedback on Workstation components >>>>> >>>>> >>>>> >>>>> 10bit display support is all over the place on mac. MBP 2016 is >>>>> driving a 24=E2=80=9D external HDMI monitor in 10bit will driving the= native >>>>> display in 8bit. recent model imacs with 5K displays running in 10bit= . >>>>> >>>>> >>>>> >>>>> you can get nV cards on win to run in 10bit. here is how to run nV or >>>>> AMD on win10 in 10bit >>>>> https://www.pugetsystems.com/labs/articles/Setting-Graphics-Card-Soft= ware-to-Display-10-bit-Output-1221/ >>>>> >>>>> >>>>> >>>>> >>>>> >>>>> S >>>>> >>>>> >>>>> >>>>> >>>>> On Feb 7, 2019, at 10:25 PM, Stephen van Vuuren < >>>>> AE-List@media-motion.tv> wrote: >>>>> >>>>> >>>>> >>>>> >>> Unless you NEED 10 bit color output (unlikely) do not get a >>>>> Quadro. They are intended for industrial CAD / CAM workflows and are = 1/10th >>>>> as fast for rendering CUDA Redshift. They are intended for use cases = where >>>>> you need extreme floating point precision, like designing critical >>>>> infrastructure. Nobody is using Quadros for Redshift / Octane. >>>>> >>>>> >>>>> >>>>> I disagree that 10-bit is =E2=80=9Cunlikely=E2=80=9D. It=E2=80=99s mo= re of a ignorance of >>>>> benefits that an unlikely scenario. I get a lot of pro colorist and p= ost >>>>> house delivering stuff with banding, too much shadow noise and color/= gamma >>>>> problems that have obviously been graded in 8-bit on typical 1000:1 p= anels. >>>>> See it all the time. >>>>> >>>>> >>>>> Same as I see people mixing on headphones or uncalibrated speakers an= d >>>>> charge pro prices for the service. The defense is always the same >>>>> =E2=80=9Clooks/sounds great here!=E2=80=9D. >>>>> >>>>> >>>>> >>>>> GeForce cards cannot display video in Premiere or images in Photoshop >>>>> in 10-bit, period or most any other pro app. You have to have a monit= or >>>>> card going to a separate, second monitor. That card or box + monitor = for a >>>>> 4K setup is not cheap. >>>>> >>>>> >>>>> >>>>> Quadro cards offer native 10-bit OS level support with Quadros and Ma= c >>>>> is now fully supported. >>>>> >>>>> >>>>> >>>>> Most people think they do don=E2=80=99t need 10-bit output but end up >>>>> delivering work that looks bad in 10-bit, 12-bit and HDR environments= or >>>>> displays with wider than the 8-bit (and 6-bit+dithering panels out th= ere). >>>>> >>>>> >>>>> >>>>> DCP is 12-bit. Rec2020 is 10-bit as is all Netflix, Amazon etc. HDR> >>>>> Hell, YouTube HDR is 10-bit. >>>>> https://support.google.com/youtube/answer/7126552?hl=3Den >>>>> >>>>> >>>>> >>>>> If you=E2=80=99re delivering 10-bit and while it used to be rare, the= se days >>>>> almost everything I deliver is 10 bit or higher. >>>>> >>>>> >>>>> So I just don=E2=80=99t take people mastering final projects for prof= essional >>>>> deliver very seriously that don=E2=80=99t have a fully 10-bit or high= er workflow. >>>>> >>>>> >>>>> Quadro plus a good affordable HDR 4K 10-bit monitor is the same or >>>>> less money than a GeForce + 10-bit video output plus dedicated 4K dis= play >>>>> and offers better drivers, more stability and more flexibility (like >>>>> Photoshop work). >>>>> >>>>> >>>>> >>>>> >>>>> >>>>> *From:* After Effects Mail List >>>>> *Sent:* Thursday, February 7, 2019 2:04 PM >>>>> *To:* After Effects Mail List >>>>> *Subject:* Re: [AE] Feedback on Workstation components >>>>> >>>>> >>>>> >>>>> Thanks for the input. It has confirmed my intuition to stay away from >>>>> the Quadro and Xeon for this build. Looking at either 2x 2080's or 1x >>>>> 2080ti now. Don't have funds for 2 of those. :-( >>>>> >>>>> No need for 10bit output here. Mostly making HD TV promos on this >>>>> machine. >>>>> >>>>> >>>>> >>>>> On Thu, Feb 7, 2019 at 10:58 AM Teddy Gage >>>>> wrote: >>>>> >>>>> okay I'm seeing a lot of mixed advice. The problem here is that there >>>>> is a major real-world consequence of getting the wrong type of proces= sor >>>>> these days, and it is REALLY important to buy the right system for YO= UR >>>>> workflow >>>>> >>>>> >>>>> >>>>> The big points - >>>>> >>>>> >>>>> >>>>> - Unless you NEED 10 bit color output (unlikely) do not get a quadro. >>>>> They are intended for industrial CAD / CAM workflows and are 1/10th a= s fast >>>>> for rendering CUDA Redshift. They are intended for use cases where yo= u need >>>>> extreme floating point precision, like designing critical infrastruct= ure. >>>>> Nobody is using Quadros for Redshift / Octane. >>>>> >>>>> - The same goes for Xeons. Do not get a xeon system. Many builders tr= y >>>>> to sell you on xeons because they can make more markup on it. They ar= e >>>>> extremely expensive and offer little to no performance benefits vs. i= 7 or >>>>> i9. This is basically due to A) lack of overclocking ability and B) w= ay >>>>> more expensive ECC RAM required (not necessary). The large number of = cores >>>>> means each core is slower than the equivalent i7/i9. AE and most task= s in >>>>> Cinema are still single-threaded. The exception is if you plan on doi= ng a >>>>> lot of CPU rendering in Arnold (which is going GPU soon anyway), or h= igh >>>>> end simulation work in houdini etc. >>>>> >>>>> - Basically - the more cores you have, the slower each core is, this >>>>> is based on thermal and manufacturing limitations. AMD TR chips are g= ood >>>>> but the same applies - only necessary for pure multithreaded CPU task= s. >>>>> >>>>> - Currently, for future-proofing, the only GPU I'd look at right now >>>>> is the GTX 2080ti - but try to find the "blower" models and not the "= open >>>>> fan" designs, as those tend to just push hot air around inside your c= ase. >>>>> >>>>> >>>>> >>>>> I'm not going into a specific parts list but from what you're saying >>>>> I'd be looking at a system with an i9 chip with 8-12 cores, high mega= hertz >>>>> boost / OC abvility (4.2-5 Ghz) very good all in one liquid cooling (= a >>>>> 360mm radiator is essential), and one or two GTX 2080ti's. And yes, R= AM >>>>> prices have blown up and stayed there due to manufacturing shortages, >>>>> supposedly. >>>>> >>>>> >>>>> >>>>> I am currently rocking a machine I built with the core i7 7900X, 4x >>>>> 1080ti, 128 GB RAM for similar workflows and it does great. I will sa= y >>>>> avoid Asus motherboards for now, they can have issues >>>>> >>>>> -Teddy >>>>> >>>>> >>>>> >>>> -- >> _____________________________ >> Teddy Gage >> President >> Shotgun Inc. >> 360 VR | MOGRAPH | VFX >> SHOTGUNPOST.COM >> BKLYN NY 11215 >> >> > -- _____________________________ Teddy Gage President Shotgun Inc. 360 VR | MOGRAPH | VFX SHOTGUNPOST.COM BKLYN NY 11215 --0000000000002bad1305819f3815 Content-Type: text/html; charset="UTF-8" Content-Transfer-Encoding: quoted-printable
Excellent breakdown Carey- While I absolutely applau= d Puget=E2=80=99s attempts to provide benchmarks for real world tasks, I ha= ve issues with their methodology and don=E2=80=99t agree with all their res= ults. Still better to study toms hardware or Anandtech benches directly for= various components, and take the Pudget results with a grain of salt. They= are still trying to sell you something.=C2=A0
- Teddy

On Mon, Feb 11, 2019 at 9:28 AM Carey Dissmore <AE-List@media-motion.tv> wrote:
<= /div>
A few comments:=C2=A0

Puget Sy= stems REALLY gets it when it comes to addressing the needs of our market. T= hey have a model where they do a ton of research and provide excellent supp= ort for their tested, configured system designs. They share a lot of their = data and findings on their website, which benefits the community as a whole= . They are uniquely positioned to do this type of testing and I=E2=80=99m v= ery thankful they share. For the vast majority of people who do NOT have in= terest in building their own systems, but nevertheless are realizing that a= custom workstation is truly the best path forward for their needs, I can= =E2=80=99t think of a better choice.=C2=A0

Now ont= o the geek stuff, (which Puget already knows):

A - Building PC: I=E2=80=99m inclined towards PC building too, and have b= uilt many workstation machines. I built my first PC around 2001 mostly for = gaming and offloaded encoding/rendering but I largely stuck with Mac for po= st production work until about 2012 when I made the complete jump to primar= y post production on Windows workstations. Teddy and I had a good chat on t= he phone back then just to gut check my system architecture before pulling = the trigger on a $6k box build. I=E2=80=99ve built and rebuilt many since.= =C2=A0

B-Performance: It used to be that synthetic= benchmarks were quite useful in determining performance as it applied to v= ideo post work. You pretty much always needed the highest performance hardw= are you could get and there wasn=E2=80=99t a whole lot of differentiation a= bout tasks, as it all sort of lifted with the general performance of the ma= chine. You could look at relative performance and extrapolate your value or= expected performance for the apps and work you did and make a decision abo= ut value for money/ROI/Cost-benefit with some simple math. =C2=A0

Those days are over. That won=E2=80=99t work anymore. The t= rend for nearly a decade is that PC performance gains have forked in multip= le directions: CPUs have added cores, and clock speed increases have crept = higher but very slowly. Due to thermals, you are often left with a choice o= f more cores vs more clock speed. GPUs have introduced massive parallelism = tuned for specific compute loads but data has to be moved from memory/fast = CPU access over the bus to the GPU and back so that is a factor in determin= ing which types of =E2=80=9Cloads=E2=80=9D are best suited to GPU compute. = The problem is that synthetic benchmarks can be designed to torture test a = specific type of computing load that shows maximized performance of a parti= cular component but it has never been less relevant to apply those to your = expectations in the apps you ACTUALLY USE to do post production work.=C2=A0=

Therefore, the ONLY benchmarks that are relevant = are the ones that are done by testing the apps you ACTUALLY USE to determin= e a particular component or configuration=E2=80=99s value to you. Here=E2= =80=99s some apps that I use and what I have found they need from a machine= .=C2=A0

For example here=E2=80=99s some differing = compute loads and what types of PC hardware/config will serve it well.=C2= =A0

Storage (applies to all): Everything wants fas= t storage and IO. Splitting up storage between media and boot and cache, ea= ch optimized is good. Storage/cacheing has become an issue too and managing= the flow of data and using high performance caching devices separate from = both system drive and media storage is key to optimizing. As you look at th= e applications below

AE: In daily use, largely sti= ll single threaded operations with dashes of multithreading throughout, and= a splash of GPU compute here and there, particularly on plugins/effects. T= oday, this means favoring higher clock speeds over core count on the CPU. G= et as much RAM as you can justify because AE and RAM previews will use it. = Generally AE-only artists don=E2=80=99t need tons of data storage, but what= they have should be fast. Often they are best served by something like a 9= 900k system and a modest GPU from NVidia.=C2=A0
Note that one of = the challenges of optimizing AE for multithreading is that not every type o= f compute task lends itself well to parallelism. Sometimes you have to wait= for the result of one computation to use the result of that to start the n= ext computation. This same performance optimization issue also affects othe= r similar apps like Fusion.=C2=A0

Premiere: A litt= le more balanced between CPU cores and clock speed and GPU. Having said tha= t Premiere still favors clock speed over massive core count, can take advan= tage of Quicksync on newer Intel CPUs BTW, and uses GPU but does not rely h= eavily on it and primarily processes color and effects on GPU (with some ex= ceptions). Premiere doesn=E2=80=99t require quite as much RAM as AE, but I = guess it doesn=E2=80=99t hurt. Having said that Premiere seems to need an a= wful lot of compute hardware thrown at it to smoothly handle media at 4k. B= y comparison FCPX (which I don=E2=80=99t use because it=E2=80=99s Mac only)= is way more optimized and gets better performance from lesser hardware. I = think there=E2=80=99s a lot of room for improvement for Premiere here but i= t probably means a rewrite.

Encoding: h264 or now = h265 encoding is massively parallel and really shows off what a lot of core= s can do. Currently CPU dominated we=E2=80=99re starting to see this moving= to GPU options as well. If all you did all day was encode stuff, go for co= re count and fast IO. However, for most of us, encoding occupies only a few= minutes per day compared to creative work.

Resolv= e: GPU dominated (for color page operations anyway). Here, CPU performance = matters but a mid-line CPU is probably adequate but what you want are GPU c= ompute units. The best units you can afford, plus more of them. 2 or 3 GPUs= scale really well with Resolve, but in my testing I=E2=80=99ve found it be= st to keep them matched. I did some mismatched testing with different gen G= PUs and it=E2=80=99s pretty variable and inefficient. Not that every little= bit doesn=E2=80=99t help=E2=80=A6but one card ends up waiting on the other= to finish a task. PCI-E lanes required for multi-GPU support push the user= to go with higher level platforms like X299.=C2=A0

Cinema 4D: When rendering, massive CPU core count is your friend. But whe= n modeling, key framing, animating etc a lot fo that is more single threade= d but obviously previews can be run OpenGL (etc) on the GPU. Other renderer= s are GPU-dominant. 3D is not my forte so I=E2=80=99m going to let others c= over this.

But I said all of that as a preface to = make this point: You can choose any one activity or app and build a kick-bu= tt machine tuned specifically to the requirements of that activity, and wal= k away with maximized ROI. An AE-only machine is going to look quite differ= ent than a Resolve-only machine, and again quite different from a C4D optim= ized machine. But often we need to engage in a mix of activities and the ch= oice about whether to try to build the ultimate =E2=80=9Cdo it all=E2=80=9D= workstation ($$$$) vs separating the tasks is something we all have to loo= k at in light of our own mix of work.

Anyway, when= I sat down I had no idea I=E2=80=99d write ll that but alas, the coffee wa= s good and plentiful. Plus, when you really get into this stuff there are j= ust a lot of considerations to get your head around. By no means did I writ= e a definitive guide here (is there such a thing?) and understand why Chris= Zwar=E2=80=99s is over 10,000 words. I look forward to it.

Carey Dissmore



On Feb 11, 2019, at 6:33 AM, Byron Nas= h <AE-List@= media-motion.tv> wrote:

Puget really impressed me with= the depth and detail of their hardware analysis for very targeted demograp= hics. It's nice to have someone really understand what the difference i= s between editing, motion graphics and GPU rendering.=C2=A0

On Sun, Feb 10, = 2019 at 11:52 PM Teddy Gage <AE-List@media-motion.tv> wrote:
Sure thing = Chris, feel free to reach me at teddygage@gmail.com. Goes for anyone- always happy to con= sult or chat hardware or software. Have also been doing a fair amount of VF= X supervising recently for a couple TV shows.=C2=A0

On Sun, Feb 1= 0, 2019 at 11:14 PM Chris Zwar <AE-List@media-motion.tv> wrote:
Hi all,
I started researching performance last year, with a view to wr= iting up an article and some thoughts for the ProVideo Coalition.=C2=A0 Cur= rently I'm about 50% of the way through my plan and I'm over 10,000= words so it's a bit slow going.=C2=A0 Even so, I was wanting to reach = out to list members who are especially interested and knowledgable on perfo= rmance related matters.

Teddy & Stephen were b= oth on my list of people to approach for input and this thread just reinfor= ces that! Are you two able let me know the best way to contact you off-list= please?
And if there's anyone else who has strong knowledge = or strong opinions on performance/ components then let me know too.

Cheers,

-Chris


On Sat, 9 Feb 2019 at 07:18, Byron Nash <AE-List@media-motion.tv> wrote:
=
Th= ose are good thoughts Stephen and I'll keep that=C2=A0in mind going for= ward. I am planning to use a Blackmagic card for video output but the monit= ors I have to use right now are 8bit HD. I'm starting to get into 4K VF= X for films on my machine but most of the rest of the team is building HD p= romos/mograph content. My only pet peeve with the BM cards is when you have= a 4K screen it won't downscale through the card and just goes black.= =C2=A0

On Fri, Feb 8, 2019 at 11:10 AM Stephen van Vuuren <AE-List@media-motion.tv&= gt; wrote:

> In regards to 10bit output an option not b= eing discussed is using a Blackmagic or Aja PCI card or device. These will = give you the option of 10bit YUV or RGB output (monitor dependent). 4K capa= ble versions of these are available for as little as $200.

=C2=A0

I mentioned external card but you are not budg= eting for the 4K 10-bit display that you will need to see the output.

=C2=A0

I have a $199 BM MiniMonitor 4K and it=E2=80=99s great going out to = a 50=E2=80=9D HDR TV). But that does not get you 4K 10-bit on your desktop display or apps that= don=E2=80=99t support that card. You have to have a second monitor whe= re as Quadro allows you to have 4K 10-bit on your main desktop monitor.<= /u>

=C2=A0

If you only have room for on 4K 10-bit, a card plus additional mon= itor is not an option. And even if you do, my setup I would rather always s= ee 10-bit as much as possible in all my work and catch issues.

=C2=A0

= As video moves to 10-bit and HDR in 4K, even for low-end, 10-bit is going t= o be a bigger deal.

=C2=A0

= From: After Effects Mail List <AE-List@media-motion.tv>
Sent: Friday, February 8, 2019 12:25 AM
To: After Effects Mail List <AE-List@media-motion.tv>
Subject: Re: [AE] Feedback on Workstation components

=C2=A0

If you=E2=80=99re talking about a budget system= Quadro cards aren=E2=80=99t even in the conversation (5k was mentioned). I= f one of the main uses of the machine will be Redshift you=E2=80=99re bette= r off buying the fastest GTX or RTX card(s) you can afford. The RTX cards may end up being a better choice in the near future, if Nvidia= =E2=80=99s and Redshift=E2=80=99s recent announcements are to be believed.<= u>

=C2=A0

In regards to 10bit output an option not being = discussed is using a Blackmagic or Aja PCI card or device. These will give = you the option of 10bit YUV or RGB output (monitor dependent). 4K capable v= ersions of these are available for as little as $200.


Dennis Wilkins

=C2=A0



On Feb 7, 2019, at 8:49 PM, Stephen van Vuuren = <AE-List@me= dia-motion.tv> wrote:

=C2=A0

Steve,

=C2=A0

That link is for 10-bit DirectX output which is= for games and Direct X apps =E2=80=93 not relevant to this discussion. Ado= be and most other pro post apps don=E2=80=99t use or support DirectX. And j= ust because a driver or monitor flags 10-bit does not mean there is actually 10-bits happening.

=C2=A0

From:=C2=A0After E= ffects Mail List <AE-List@media-motion.tv><= span class=3D"m_-8524901683949332893gmail-m_-8825890941941454335m_302595607= 9890968975gmail-m_-8121024094433066792gmail-m_-4363005164514653576apple-con= verted-space">=C2=A0
Sent:=C2=A0Thursday, February 7, 2019 11:40 PM=
To:=C2=A0After Effects Mail List <AE-List@media-motion.tv>
Subject:=C2=A0Re: [AE] Feedback on Workstation= components

=C2=A0

10bit display support is all over the place on = mac. MBP 2016 is driving a 24=E2=80=9D external HDMI monitor in 10bit will = driving the native display in 8bit. recent model imacs with 5K displays run= ning in 10bit.=C2=A0

=C2=A0

you can get nV cards on win to run in 10bit. he= re is how to run nV or AMD on win10 in 10bit =C2=A0https://ww= w.pugetsystems.com/labs/articles/Setting-Graphics-Card-Software-to-Display-= 10-bit-Output-1221/

=C2=A0

=C2=A0

S




On Feb 7, 2019, at 10:25 PM, Stephen van Vuuren= <AE-List@media-motion.tv> wrote:=

=C2=A0

>>> Unless you NEED 10 bit color outpu= t (unlikely) do not get a Quadro. They are intended for industrial CAD / CA= M workflows and are 1/10th as fast for rendering CUDA Redshift. They are in= tended for use cases where you need extreme floating point precision, like designing critical infrastructure. Nobody is using Q= uadros for Redshift / Octane.=C2=A0

=C2=A0

I disagree that 10-bit is =E2=80=9Cunlikely=E2= =80=9D. It=E2=80=99s more of a ignorance of benefits that an unlikely scena= rio. I get a lot of pro colorist and post house delivering stuff with bandi= ng, too much shadow noise and color/gamma problems that have obviously been graded in 8-bit on typical 1000:1 panels. See it all the time.=C2=A0


Same as I see people mixing on headphones or uncalibrated speakers and char= ge pro prices for the service. The defense is always the same =E2=80=9Clook= s/sounds great here!=E2=80=9D.

=C2=A0

GeForce cards cannot display video in Premiere = or images in Photoshop in 10-bit, period or most any other pro app. You hav= e to have a monitor card going to a separate, second monitor. That card or = box + monitor for a 4K setup is not cheap.

=C2=A0

Quadro cards offer native 10-bit OS level suppo= rt with Quadros and Mac is now fully supported.

=C2=A0

Most people think they do don=E2=80=99t need 10= -bit output but end up delivering work that looks bad in 10-bit, 12-bit and= HDR environments or displays with wider than the 8-bit (and 6-bit+ditherin= g panels out there).

=C2=A0

DCP is 12-bit. Rec2020 is 10-bit as is all Netf= lix, Amazon etc. HDR> Hell, YouTube HDR is 10-bit.https://support.google.com/youtube/answer/7126552?hl=3D= en

=C2=A0

If you=E2=80=99re delivering 10-bit and while i= t used to be rare, these days almost everything I deliver is 10 bit or high= er.


So I just don=E2=80=99t take people mastering final projects for profession= al deliver very seriously that don=E2=80=99t have a fully 10-bit or higher = workflow.


Quadro plus a good affordable HDR 4K 10-bit monitor is the same or less mon= ey than a GeForce + 10-bit video output plus dedicated 4K display and offer= s better drivers, more stability and more flexibility (like Photoshop work)= .

=C2=A0

=C2=A0

From:=C2=A0After E= ffects Mail List <AE-List@media-motion.tv><= span class=3D"m_-8524901683949332893gmail-m_-8825890941941454335m_302595607= 9890968975gmail-m_-8121024094433066792gmail-m_-4363005164514653576apple-con= verted-space">=C2=A0
Sent:=C2=A0Thursday, February 7, 2019 2:04 PM<= br> To:=C2=A0After Effects Mail List <AE-List@media-motion.tv>
Subject:=C2=A0Re: [AE] Feedback on Workstation= components

=C2=A0

Thanks for the input. It has confirmed my intui= tion to stay away from the Quadro and Xeon for this build. Looking at eithe= r 2x 2080's or 1x 2080ti now. Don't have funds for 2 of those. :-(= =C2=A0

No need for 10bit output here. Mostly making HD= TV promos on this machine.

=C2=A0

On Thu, Feb 7, 2019 at 10:58 AM Teddy Gage <= AE-List@media-motion.tv> wrote:<= /p>

okay I'm seeing a lot of mixed advice. The = problem here is that there is a major real-world consequence of getting the= wrong type of processor these days, and it is REALLY important to buy the = right system for YOUR workflow

=C2=A0

The big points -=C2=A0

=C2=A0

- Unless you NEED 10 bit color output (unlikely= ) do not get a quadro. They are intended for industrial CAD / CAM workflows= and are 1/10th as fast for rendering CUDA Redshift. They are intended for = use cases where you need extreme floating point precision, like designing critical infrastructure. Nobody is using Q= uadros for Redshift / Octane.=C2=A0

- The same goes for Xeons. Do not get a xeon sy= stem. Many builders try to sell you on xeons because they can make more mar= kup on it. They are extremely expensive and offer little to no performance = benefits vs. i7 or i9. This is basically due to A) lack of overclocking ability and B) way more expensive ECC RAM r= equired (not necessary). The large number of cores means each core is slowe= r than the equivalent i7/i9. AE and most tasks in Cinema are still single-t= hreaded. The exception is if you plan on doing a lot of CPU rendering in Arnold (which is going GPU soon an= yway), or high end simulation work in houdini etc.=C2=A0

- Basically - the more cores you have, the slow= er each core is, this is based on thermal and manufacturing limitations. AM= D TR chips are good but the same applies - only necessary for pure multithr= eaded CPU tasks.=C2=A0

- Currently, for future-proofing, the only GPU = I'd look at right now is the GTX 2080ti - but try to find the "blo= wer" models and not the "open fan" designs, as those tend to= just push hot air around inside your case.=C2=A0

=C2=A0

I'm not going into a specific parts list bu= t from what you're saying I'd be looking at a system with an i9 chi= p with 8-12 cores, high megahertz boost / OC abvility (4.2-5 Ghz) very good= all in one liquid cooling (a 360mm radiator is essential), and one or two GTX 2080ti's. And yes, RAM prices have blown up and sta= yed there due to manufacturing shortages, supposedly.=C2=A0

=C2=A0

I am currently rocking a machine I built with t= he core i7 7900X, 4x 1080ti, 128 GB RAM for similar workflows and it does g= reat. I will say avoid Asus motherboards for now, they can have issues

-Teddy=C2=A0

=C2=A0

--
_____= ________________________
Teddy Gage
President=
Shotgun Inc.
360 VR | MOGRAPH | VFX<= /div>
BKLYN NY 11215


-- =
_____________________________
<= font color=3D"#cc0000" face=3D"arial, helvetica, sans-serif" size=3D"2">Ted= dy Gage
President
Shotgun Inc.
360 VR | MOGRAPH | VFX
BKLYN NY 11215

--0000000000002bad1305819f3815--