What computer are you using for serious CAD/Graphic work at anchor.

[jasn]

So I'm in the camp of choosing the most powerful laptop that I can, when I finally move on to my boat. These tend to be gaming laptops and customization is key. I also don't like 17 inch laptops, because of their size and weight, so I focus on finding the best performing 15 inch laptop model. IMO the two most popular customizable brands today are MSI and Clevo. You can find many resellers who will customize models from both companies with a variety of options that you won't typically find with the bigger name brand companies. (Or if you're a build it yourself type person, you can get these laptops without components, and install your own)

I tend to favor the Clevo over the MSI because I like their industrial, (admittedly bland), build. One of the top 15" models today is the Clevo P751TM-G, which can be fitted with the desktop Intel® Core™ i7-8700K 6 Core Processor. It can also be installed with 4 - 16GB memory modules for a total of 64GB, and there are currently two resellers, (that I'm aware of), that will fit the nVidia mobile 1080 GPU into the laptop. (Normally it's either the 1060 or 1070 GPU). The European reseller is Eurocom, and they call their version, the Sky X4C. (The other reselller is HIDEvolution, in the US) Finally the laptop has 2 - M2 drive slots, and 2 - SATA drive slots, so you can add a lot of storage to it as well. The down side is that battery life is limited, and the power draw can be anywhere from 30-38w at idle, on up to 256w. This is a good review of the Eurocom model of this laptop.

Anyway, you can customize almost every aspect of these laptops at many resellers, and if you want to get the most power into the smallest form factor, it's the way to go.

Good Luck..

P.S. I am a current Clevo laptop customer of HIDEvolution, but not Eurocom.

[leftbrainstuff]

ProEngineer Wildfire 4.0

I have 10,000+ hours on 3D CAD, CFD, FEA. ProE v18 since 1997. ProE was more than a decade ahead of anything else for a long time.

I run ProE on several laptops. Nothing fancy. 16GB ram is fine. I also run ProE on AWS workspacea thin client. I've also used Unigraphics, Catia and Solidworks but always go back to ProE. Freecad looks interesting but I dont have the time to contribute. They dont evwn support assembly modelling yet.

I design most boat fabrication updates in 3D and then build to print. I don't deal with very large assemblies anymore and my laptops would be unsuitable.

I also use circuitikz (Latex) for all my electeical and mechanical schematics.

[jasn]

FWIW...

Around 2001, the computer industry began producing CPUs with more than one processing core. IBM Power4 CPUs first appeared in 2001, AMD introduced the Opteron and Athlon CPUs in 2003, and Intel revised the Xeon and Pentium with a dual core design in 2005. All of these CPUs featured, for the first time, 2 CPU "cores". They operate just as they sound, each core being able to independently process workloads, and increase the processing power of the overall CPU. (They also typically share some things as well, like I/O, memory channels, and some level of CPU cache)

Around 2004, CPU manufacturers also started to include a Simultaneous multithreading, (SMT), feature, which allows an individual CPU/core to work on more than one workload at a time, each being a unique "thread". With this feature enabled, a modern operating system will report the number of CPUs as the number of CPU cores, times the number of threads per core, as the total number of CPUs available. Today Intel's SMT feature, called Hyper-Threading, typically defines 2 threads per core. So using an Intel CPU, like the 4 core, i3/5/7 Core CPU, the OS will report 8 available CPUs, (4 physical cores times 2 threads each).

Intel Xeon CPUs are primarily designed for server and workstation workloads. These are workloads where multiple applications are running at the same time, and where the applications themselves are also multi-threaded. These workloads tend to look like ones which utilize the CPU at a high rate, for a consistent period of time. In this case CPU speed isn't valued as much as things like overall core count, and other pipeline features. Intel Core CPUs, are primarily designed for client computing workloads, where the workloads tend to look more "spiky" in terms of CPU utilization, and which work, (and have to wait on), other parts of the computer, particularly the graphics processing unit (GPU). This is typical of applications like high end graphics applications and games. So core count isn't valued as much as having a higher speed of processing.

If you use the Intel product comparison site, you'll notice differences between the Xeon and Core CPU families that include, the Xeon CPUs being typically more expensive, featuring more cores, slower speeds, and support for higher end "server" features, like the more expensive error correcting, (ECC), memory. However, the lines are getting pretty blurred between these two primary use cases and CPU product families. It's true that while the highest end Xeon server CPU today has 28 cores, and wouldn't ever be used in a workstation, Intel has also introduced the Xeon E3 and D CPUs, which target workstation applications, and include things like Intel HD Graphics support, normally only available on the consumer CPUs. At the same time the Core CPUs are advancing in functionality towards supporting higher end workloads, such as increasing core count and overall performance.

(Please pardon the generalizations)

[micheln]

In a similar situation. Best option i decided on was to simply remote desktop into the computer at home. So basically you have the decked out machine setup on land, and from the boat I work the same way as at the office. My work machine cant be replicated on a boat anyway. Now I actually use a widows phone with continum , a large screen keyboard an mouse, works flawless for over a year.

Only thing is I need cellular internet access and even that does not need to supper, but 4g works great, no real lag. If I where to ditch the house I would just host it in a data center, it's cheap enough.

Food for thought

[cyan]

Originally Posted by jasn

FWIW...

Around 2001, the computer industry began producing CPUs with more than one processing core. IBM Power4 CPUs first appeared in 2001, AMD introduced the Opteron and Athlon CPUs in 2003, and Intel revised the Xeon and Pentium with a dual core design in 2005. All of these CPUs featured, for the first time, 2 CPU "cores". They operate just as they sound, each core being able to independently process workloads, and increase the processing power of the overall CPU. (They also typically share some things as well, like I/O, memory channels, and some level of CPU cache)

Around 2004, CPU manufacturers also started to include a Simultaneous multithreading, (SMT), feature, which allows an individual CPU/core to work on more than one workload at a time, each being a unique "thread". With this feature enabled, a modern operating system will report the number of CPUs as the number of CPU cores, times the number of threads per core, as the total number of CPUs available. Today Intel's SMT feature, called Hyper-Threading, typically defines 2 threads per core. So using an Intel CPU, like the 4 core, i3/5/7 Core CPU, the OS will report 8 available CPUs, (4 physical cores times 2 threads each).

Intel Xeon CPUs are primarily designed for server and workstation workloads. These are workloads where multiple applications are running at the same time, and where the applications themselves are also multi-threaded. These workloads tend to look like ones which utilize the CPU at a high rate, for a consistent period of time. In this case CPU speed isn't valued as much as things like overall core count, and other pipeline features. Intel Core CPUs, are primarily designed for client computing workloads, where the workloads tend to look more "spiky" in terms of CPU utilization, and which work, (and have to wait on), other parts of the computer, particularly the graphics processing unit (GPU). This is typical of applications like high end graphics applications and games. So core count isn't valued as much as having a higher speed of processing.

If you use the Intel product comparison site, you'll notice differences between the Xeon and Core CPU families that include, the Xeon CPUs being typically more expensive, featuring more cores, slower speeds, and support for higher end "server" features, like the more expensive error correcting, (ECC), memory. However, the lines are getting pretty blurred between these two primary use cases and CPU product families. It's true that while the highest end Xeon server CPU today has 28 cores, and wouldn't ever be used in a workstation, Intel has also introduced the Xeon E3 and D CPUs, which target workstation applications, and include things like Intel HD Graphics support, normally only available on the consumer CPUs. At the same time the Core CPUs are advancing in functionality towards supporting higher end workloads, such as increasing core count and overall performance.

(Please pardon the generalizations)

This CPU history is quite interesting to me. Thanks!
However, in graphics-intensive applications, the GPU is super important, IMO. In my line of work, I have seen graphics processing speeds increase by more than 100 times when we port a multi-threaded CPU application over to a GPU. Thank goodness for the gamer kids (and recently blockchain currency) that have propelled the GPU industry!