The incessant pace of progress and innovation for workstation technology never slows.
Less than a quarter after every major workstation OEM launched a full trio of models based on Intel’s Sandy Bridge-EP (a.k.a. Xeon E5), the industry leader in CPUs has already released its follow-on processor generation, code-named Ivy Bridge. And subsequently, we are now seeing the first Ivy Bridge workstations hitting the market, including Dell’s Precision T1650 and HP’s Z220.
How Does Ivy Bridge Affect the CAD Workstation Market?
What benefits can Ivy Bridge offer to those plying their trade in CAD? Well, there’s the usual broad-based boost in performance that any good generational upgrade will provide, as Intel expects a 20 percent performance improvement for general computation from Ivy Bridge (though of course mileage will vary by application). But there’s more appeal for this upcoming product family than just the usual generation-to-generation performance bump. Because while that appeal extends across applications and usage models, there are a few special nuggets of technology in this generation that will pique the interest of workstation-wielding CAD professionals.
Intel’s lead in silicon process manufacturing continues to grow, and the benefits of Ivy Bridge should prove an ideal vehicle to showcase that lead. Just as competitors are getting their 32 nm process, with Ivy Bridge Intel’s jumping a full generation ahead with a 22 nm process that allows for millions more transistors in the same silicon area.
That’s a win for workstation buyers especially, as they represent a professional community that certainly care about CPU performance, but demand a lot more. First off, a shrink buys room for more cores, and we’ll eventually see some Ivy Bridge SKUs with eight or more cores (not at first launch, but later in the product lifecycle). Far from being one-trick-ponies, today’s MCAD professionals have to be jacks-of-all-trades — a competitive market, tight budgets and tighter schedules all demand it. Drawing is just one piece of the daily workflow, complemented by a host of other critical compute tasks, from simulation to styling. And chores like finite element analysis and computational fluid dynamics multi-thread quite well, making 50% more available cores a serious weapon in driving computation time down and achieving the ultimate goal — boosting productivity.
Improved Integrated Graphics
The extra silicon space also allowed Intel to dial up the performance and functionality of its integrated graphics hardware. For example, Ivy Bridge’s P4000 GPU populates more on-chip graphics engines and supports advanced features like hardware tessellation, a proven tool that can deliver finer, more realistic 3D surfaces in less time. With its range of upgrades, Ivy Bridge can claim full DirectX11 support that its predecessor could not. And with more of those bigger, faster graphics engines, Intel can claim a 30% increase in performance for Ivy Bridge’s graphics over Sandy Bridge’s. And that means CAD professionals on a budget can now more seriously consider choosing a low-cost CPU-integrated graphics solution like the P400.
Support for Three Monitors
But looking beyond performance, Ivy Bridge’s graphics is going to provide another big draw for the MCAD professional: native support for three monitors. While the mainstream is now just discovering the benefits of dual monitors, many mechanical designers are already using three: for example, one for drawing, one for simulation and one for visualization. Prior to Ivy Bridge, a desktop with three high-resolution monitors mandated at least one discrete add-in graphics card. But with this generation, a cost-conscious MCAD user could go three-wide and stick with base platform graphics.
MCAD Users: Same Performance, 50% Fewer Watts!
With more cores to speed CAD simulation and ultra-realistic rendering, as well as a 30 percent graphics improvement, Ivy Bridge promises to be a tide that raises all boats, as all workstations — deskside or mobile — will benefit. But there’s one unique advancement debuting in Ivy Bridge that’s a particular boon to the MCAD pro on the go. You see, Ivy Bridge’s 22 nm technology introduces a revolutionary new transistor structure called TriGate that offers the same performance at 50% fewer Watts than Sandy Bridge’s 32 nm.
And that’s allowing leading vendors HP, Lenovo, Dell and Fujitsu to introduce new mobile workstation models that dramatically extend battery life at the same performance level, or deliver far more performance, with the same battery life. Either way you look at it, it’s a win when computation demands are high. And few corners of the computing world demand more performance/Watt than mechanical designers trying to accomplish demanding design work on the road.
This post reflects industry analyst Alex Herrera’s views and does not necessarily reflect the opinions, product plans or strategy of either Dell or Intel.
Optimizing hardware for SolidWorks is essential for getting the most out of this heavy-hitting CAD application, as we’ve discussed on CADspeed previously. So we were thrilled when the SolidWorks forum addressed this very issue recently on their forums.
The key to getting the most out of SolidWorks, or any CAD application for that matter, is ensuring your hardware can handle the workload. Remember that your situation is unique. In simple terms, two users using the same software on the same system may have very different perspectives on their workload efficiency if one is using 3D rendering and the other is not. Consider your needs first and foremost.
On the flip side, if you know you need new hardware, simply buying the most expensive machine may not pay off in the long run either. Think in terms of your productivity while shopping for a new workstation to get the most for your budget, hopefully with a little room to grow for those inevitable upgrades.
That said, here’s a summary of the recommendations straight from SolidWorks themselves.
RAM (Random-Access Memory)
The amount of RAM you need depends less on SolidWorks and more on the number of applications you run at the same time, plus the size and complexity of your SolidWorks parts, assemblies and drawings. SolidWorks recommends you have enough RAM to work with your common applications (i.e., Microsoft Office, email, etc.) and load your SolidWorks documents at the same time.
Processor speed is another key factor when selecting the right hardware for you. It’s hard to sort through all the different options though, so we recommend testing a system with your actual models. SolidWorks also offers a helpful Performance Test, which offers a standardized test for determining performance of your major system components (i.e., CPU, I/O, video) when working with SolidWorks datasets. Even better, when you complete the SolidWorks Performance Test, you have an option to share your score with others. This gives you, and other community members, a sense of where a system stands relative to others. Nice!
Note that SolidWorks and some of its add-ons (PhotoView 360) have some multithreaded capabilities, so the application can use the second processor or multiple cores. But SolidWorks says that rebuilds are single threaded and therefore rebuilds generally will not be faster with multiple CPUs or cores.
The size of your hard drive or solid-state drive should be based on the disk space you need. Take a look at all your system’s components: operating system, applications and documents. If you work primarily on a network, your needs may be different than those who primarily use their local drive. Don’t forget to develop a back-up plan for your data, if you don’t already have one. (You do have one, right?)
The very nature of CAD software requires a good workstation-level graphics card and driver. You are probably going to need at least a mid-range card, if not a high-end card, depending on the type of CAD work you do. For graphics cards, we recommend starting with the SolidWorks Certified Graphics Cards and System, because SolidWorks has done the testing for you.
Can’t get enough about hardware configurations for SolidWorks? Check out this great post from SolidWorks on their forums. Or learn more about the minimum requirements for SolidWorks.
Avoiding the VHS vs. Betamax War
Arguably one of the most important elements of OpenCL is that it is an open standard, not controlled by any one vendor and not limited to one kind of graphics cards or CPU. Microsoft has DirectCompute. Nvidia has the proprietary GPU-only CUDA. But OpenCL is vendor neutral with incredible momentum and the only solution that is designed for the next generation of heterogeneous computing coming from Intel and AMD.
Heterogeneous Computing Makes OpenCL Even More Relevant
Heterogeneous computing is the new term you will hear to refer to integrated CPUs and GPUs on a single die (e.g., AMD’s Fusion APUs or Intel’s Sandy Bridge). This is the future of mobile, handheld and desktop computers. The APU design is both more power efficient and solves the problem of data transfer latencies between the CPU and GPU.
This shift in processor design makes OpenCL even more relevant and ubiquitous. Because GPU and CPU are on the same die, there is no bandwidth or bus latencies when transferring data between CPU and GPU. OpenCL code runs full throttle. For additional performance, add in a discrete workstation graphics board. Any OpenCL-savvy application will automatically and seamlessly take advantage of the additional compute power.
What’s a CAD User to Do Now?
Chances are you already have a workstation graphics card in your desktop or mobile workstation. What you want are applications that take advantage of OpenCL. The best way to accelerate this is to contact your preferred CAD/CAE software vendor (e.g. ANSYS, Autodesk, CEI, Dassault, ESI, Intelligent Light, MCS, Siemens to name a few) and ask them when they will be adding OpenCL for new features or to accelerate existing features in their application. Most of the significant players are already working on it, so your voice just helps them get their products to market faster.
I’m going to be following the upcoming AMD Fusion Developer Summit very closely. Much of this conference is focused on OpenCL, so I am expecting to see some interesting announcements and demonstration that show off OpenCL capabilities. I’ll post updates as I hear them.
Author: Tony DeYoung
The first post in this series discussed why you want OpenCL. The second post described how it works. This post discusses how OpenCL will affect your workflow.
Below are some “compute” examples of where OpenCL will impact the CAD workflow:
- Linear algebra
- Signal/Audio/Image Processing/Video
- Finite difference method app
- Finite-element solving and direct solvers
- Finite particle Method FPM and airbag simulation
- Constraint Solving
- Contact search / contact analysis for nonlinear simulation
- CAD modeling engine
- Boolean operation, interference and clearance calculation
- Model tessellations
- Hidden-line removal
- Graphics visualization and rendering
- Injection molding flow simulation
- Cloth simulation
- NC tool positioning and material removal simulation
- Robotics and plant automation with robot tool path planning
- Data sorting and database operations. See PostgreSQL with OpenCL.
The greatest impact for CAD and designer productivity will be workflows where there is a tight coupling between visualization and compute or optimization and visualization. Examples are simulation-based optimizations and design studies on full vehicles (from automobiles to aeronautics to yacht design).
The Holy Grail of Rendering: Real-Time Ray Tracing
I’m a visual guy attracted to shiny spherical balls that reflect the environment off of their surface, i.e., ray tracing. OpenCL is a formidable tool to accelerate any ray tracing application by at least an order of magnitude. To me perhaps the most interesting right now is Caustic Graphics and OpenRL (Open Ray Tracing Library), their standard for writing ray tracing applications that execute across heterogeneous compute platforms. OpenRL uses OpenCL to take advantage of any GPU in the system (add-in board or APU) to accelerate ray tracing.
As a note: Apple developed OpenCL (before submitting to the open standards Khronos Group). Apple is already a major investor in Imagination Technologies, which recently bought Caustic Graphics. My conclusion: it is only a matter of time before you see the benefits of OpenRL/OpenCL on iOS devices.
Author: Tony DeYoung
The first post in this series discussed why you want OpenCL. This post will describe how it works.
The GPUs in present day graphics cards like the AMD FirePro/Radeon and Nvidia Quadro/Geforce lines are massively parallel, multithreaded, multicore processors with enormous computational power and high bandwidth. Traditionally these multicore processors have been used for graphics processing, leaving the CPU to do everything else.
More Computing Power Using Massive Parallelism
The paradigm shift with OpenCL is a non-proprietary, standardized (and familiar) language to divide up general-purpose computational code into parallel threads so the GPU and CPU can work in tandem to deliver new functionality or tackle large processing tasks.
One of the key elements about OpenCL is its ability to allocate resources to the GPU or multicore CPU depending on how much power is needed and how data intensive any given task is. An OpenCL CPU+GPU-based solution means you can get simultaneously high performance for a design as well as its analysis and simulation.
In business terms, what OpenCL means is that responsiveness and speed from existing servers to handheld devices, will improve dramatically. When algorithms are redesigned to use OpenCL, speed-ups of 10x are common, and speed-ups of 30x are not unusual. (See, for example, EDEM Simulation Engine.)
Next I’ll discuss how OpenCL will affect your workflow.
Author: Tony DeYoung