Optimize your Visual Studio CPU performance with these hardware suggestions.
Optimize your Visual Studio CPU performance with these hardware suggestions.
Hello everyone, I'm working with two developers who use Dell OEM i9-10900 SFF systems. They seem to be struggling, so I'm considering building my own setups to replace them. These machines feature a mix of LGA 1700 and LGA 1718 sockets, which isn't ideal for me. The main challenge is optimizing Visual Studio performance—do Radeon or Intel iGPUs offer better results? Also, should I opt for an Intel CPU with an RTX 3050 (6GB) instead of going the F SKU route? I do have a CUDA package in Visual Studio, but I'm unsure if it's worth the added complexity. Right now, I'm evaluating the 14700k and Asus Prime Z790M-Plus (without triple display outputs) in a standard mATX build, possibly using a Fractal Design R5 Silent. I wouldn't push overclocking, as TDP likely caps their performance. I prefer the benefits of a K SKU CPU over a non-K one, especially if they're priced similarly. The Asus Pro CSM boards are great, but they lack VRM heatsinks, which could be an issue here. It's also a 125W TDP limit board, which is fine—performance per watt is key. The 13700k is another strong option, around $60 cheaper than the 14700k while offering similar gains. If there are clear advantages for CUDA and RTX in Visual Studio, would it be worth considering an RTX 4060 8GB?
Visual Studio functions mainly as an enhanced text editor. What performance challenges are you encountering? If you're having trouble building your projects there, it often comes down to the specific tasks you're performing. For example, a simple Python script might run smoothly, but compiling a full office suite could be slower, regardless of the editor used.
Its not just large VS compiles but also a suite of other production applications on triple displays. VS is just the only one I'm unfamiliar with regarding performance characteristics. Basic Google search regarding CUDA had something to show, but I'm not sure if that's worth exploring. My personal experience with VS is limited. The rest of the applications don't require any crazy rasterization performance and just need single/multicore CPU. Regarding the current systems, its a combination of the suite of applications (which includes VS) that's causing these OEM systems to scream constantly. Its quite obviously a CPU multithreading limitation, but I now have a reason to upgrade these machines to fully min-max their needs. If that involves a dGPU, then sure, but it doesn't seem necessary for VS. What is questionable though is the difference between Radeon and Intel iGPUs that might matter. So far, a 13700k at 125W TDP with the Asus Pro 760M CSM seems to be the best route. I've got the B660 version of this board in a workstation and absolutely love it. I'm just looking to confirm whether there's not a better option regarding the iGPU/dGPU. Pro B760M-CT-CSM｜Motherboards｜ASUS USA
They are working on something in a programming language that supports CUDA, and NV is the preferred choice. If compilation takes too long, having more cores is beneficial. You might consider using workstation CPUs, but sticking with consumer chips suggests going for the 5/7950X.
No major benefit comes from Intel Quicksync. I think the 7950x doesn’t need a special encoder if that’s what you require. Concerning VS and CUDA, it means the application must support CUDA hardware, which I’m interpreting correctly. From what I understand, developers don’t write anything using CUDA; they mostly use C++ in VS. I have a server with a 7950x that performs exceptionally well. The same Asus Pro CSM board I prefer has an AM5 version. The issue is that 7950x models cost over $650, and the next Ryzen model, like the 7900x, is surpassed by higher-end options such as 13700k/14700k.
Are you handling video files? Quicksync is typically used for encoding and decoding video. This isn’t a routine part of most development work, though it can be important depending on the situation. AMD also provides encoders/decoders built into their iGPUs that perform similar tasks. What performance metrics appear in Task Manager when developers are active? Would you expect high usage in CPU, disk, memory, GPU or other areas? I’d assume a 10900 remains a solid processor and you probably won’t notice significant gains with today’s hardware.
They're built entirely within C# or C++, which keeps things manageable without overly complex challenges. From what I've observed, they tend to grow fairly substantial. Since the existing applications already face compilation hurdles, imagine how slow a build could become. The i9-10900 used in current SFF Dell OEM systems is highly constrained thermally and in terms of power dissipation, while newer 13th/14th generation chips offer much better single-thread and multi-thread performance.
The i9-10900 works well in a fully equipped setup, though Dell SFF models may struggle with the i5 version. It’s not ideal for users who often face heavy CPU loads. They’ve already cleared most dust, but a refresh might be needed. I’d prefer swapping out those workstations and using them for lighter tasks instead.
The systems have a limited amount of RAM. It seems they're managing memory usage rather than relying on CPU power. An integrated GPU should suffice unless the setup involves GPU-specific tasks. There are no advantages from using CUDA or RTX here—it's just a text editor, not a demanding application like a triple A game or 3D program.