At long last, you are clear it’s time for a new workstation, or, that it’s time the whole team got new workstations.
You’re dreaming of what life might be like rendering in minutes not hours. Or even, that you might cross over into the magical realm known as Real Time Rendering. Imagine real time rendering in Maya, real time rendering in AutoCAD, real time rendering in C4D….
Well, maybe not dreaming. But isn’t this close to the truth? Today I want to look at how to buy a workstation, and also, how you might determine the value of this purchase.
Your budget is the first stop. Being clear about your budget will allow you to focus on building a machine that will be optimized for the work you do. Having a number that’s bouncing around, or having no number at all, makes the process, well, chaotic and inefficient. Usually.
So what goes into the budget? There are probably several ways of looking at this, but I’m going to look at two: Cost – Benefit and Percentage of Budget.
I. Cost – Benefit
I’m no mathematician, so feel free to point out what I miss here. That said I do think some math is in order here. You are either a media professional or a technical design professional (CAD, CAE, CAM). And let’s say you work full time, conservatively 50 hours a week. You know there’s got to be something better than what you’re working with, but what really is the value of a faster machine?
Let’s say you have three year old workstation with an 8-core CPU currently running at 3.1 GHz, 16GB RAM and a 2GB Quadro K2000 GPU. Taking 10 minutes to breakdown your workday and week, you see you spend 16 hours per week (about two days weekly) on average processing / rendering something – fluid simulation, modeling, compositing, transcoding, whatever.
You have been shopping. You like the new i-X2, and you like the new HP Z840. You like the sound of 64GB of RAM and the Dual 8-core Intel Xeon E5-2667 v3, 3.2 GHz, 135W (16 Cores) option, with either the Quadro K4200 or Quadro K5200 for GPU is appealing. Leaving the GPU out of the equation, your 16 core i-X2 will have processing power at least 100% greater than your current machine. In other words, you will have 8 hours you didn’t, or be 8+ hours more productive.
If you earn $50 per hour, that’s $400 week. $1600 per month. or $19,200 per year…so this $10,635 i-X2 (with Quadro K4200) will pay for itself in 6 months. If you earn $100 per hour, it’s 3 months. But there’s more, because the lion’s share of media and design applications do offer GPU acceleration as well, which will, conservatively, reduce rendering time another 25% or more. And, depending on your workflow, configuration of your workstation to be optimized for renderers like Octane Render can virtually eliminate render / processing time altogether – meaning the hallowed real time rendering. This is available with the i7-X 5960X and two GTX Titan Black or the forthcoming 9-series Titan replacement, meaning your investment could pay for itself in … 4-6 weeks.
II. Percentage of Budget
The other approach is budgeting for your purchase, or viewing the purchase as a percentage of budget. Let’s say you have 3 new projects for the upcoming quarter, two must be completed in 30 days and the third in 90 days. What percentage of this budget does it make sense to spend on new hardware?
Answer: The amount equal to what will allow your team to produce excellent work, on time and under budget. This means the right hardware, configured correctly, for the work you do. You know yourself and/or your team. You know the bottlenecks in your current hardware (and software! Let’s be honest if you’re still on Adobe Creative Suite 5.5 your parking brake is on, so to speak). I could break this down in much more detail, but I trust you’re aware of those factors in your work / business, and so can determine how best to achieve your objectives and the relative value of the right tools for the job.
Which leads us to something the value of which is hard to quantify: The Right Tools.
About a year and a half ago a famous maker of sports clothing and gear contacted us. They’d ordered eight E5-2660 dual xeon workstations with Quadro 4000 GPUs the year before from a well-known custom PC maker from Texas. A friend of ours referred us because his friend on the design team said “We need to get more done faster, plain and simple.”
They told us their workflow – Maya and 1-2 other applications along with their i/o. We built them an i7-X with dual Geforce GPUs – a no-brainer for Maya and Octane Render as their render engine.
The i7-X outperformed their Dual Xeon boxes by a 100% margin, and cost some $5,000 less each (For the money, the i7-X is the fastest PC for Maya / Octane, hands down).
This is key – and it points to something very important. If you ask HP, Dell, or Apple
“My core applications are Premiere, After Effects, Cinema 4D and Photoshop – what’s the best configuration for my needs?”
“What’s the best GPU for Solidworks and AutoCAD?”
you will have a hard time getting a straight answer. Or someone who can answer it. Or, you’ll go to the HP website to look at workstations and see nothing but Quadro or FirePro GPUs – good GPUs, but for many applications they are not the best, and they’re almost never the best value.
Do not take my word for it. Do not. Call them and ask. There are people at those companies who can answer these questions, but you may spend a very long time trying to find them.
So what do you do? You are left wondering
“Should I get two GPUs?
“How much RAM do I need for After Effects CC? What difference will it make?” and
“What should I do about media drive to not drop frames in Premiere Pro CC?”
You probably have questions on other aspects of your configuration also. Why? Because you’re not a hardware expert, you’re a visual effects artist. Or an architect. Or an engineer.
So, how to buy a workstation?
You can do lots and lots of research, look for benchmarks for different apps for weeks, or ask us. Ask others too – the right answers to your questions are a worthwhile endeavor than can make a difference in your work every single day. This is what we do every day, for media pros and technical design professionals and some scientific types. It’s what our small community of users is all about, and you are who we are here for and who we support.