8 core Any Value Over 4 in TS19?

Trainlover16 - Emphatically - and Demonstrably, it is the Graphics Processing Unit (GPU) that makes the most significant performance improvement in T:ANE or TRS19.
There's a point of diminishing returns reached with increased CPU power and core frequency, since the engine doesn't scale beyond using a few cores, whereas (other factors being equal) the faster your GPU, the better Trainz performance will become. (Admittedly, these components are interdependent, with high performance GPUs needing matching performance from CPUs for optimal frame rates/ responsiveness to avoid 'bottlenecking'.)
 
Last edited:
Graphic Processing Units (GPU) also function as a CPU for some large-number computations. Adding a more expensive CPU, with the same GPU, does not always allow the anticipated gains. The exact relationship between the two computer elements depends upon the actual program, which is a living element - under constant change. Thus the marriage of a GPU to a CPU may be great today, a new release of software may unbalance that marriage.

"Love And Marriage" (popular song) is appropriate in a graphics computer (Trainz) setup. "You can't have one (CPU) without the other(GPU)". Buy a better CPU and that GPU may begin to operate at 100%, sometimes (HOT). Just like a marriage, it may become burned up and quit.

While the N3V programs may be a reasonable cost, the hardware needed to achieve their full potential may not. AND, the next release may make demands that can force some to spend a lot of money just to get the full potential of the wonderful new features and keep the marriage balanced.
 
The additional cores in the processor are quite noticeable if you are going to do things that require it. I noticed a big difference when running a DBR, or processing content in Content Manager. The additional cores also relieve the CPU from needing to handle the system tasks while performing other CPU-intensive actions since system tasks are moved to another core while additional cores are used for other things, and these additional cores allow the CPU to work efficiently with multiple tasks. Downloading content, while editing an asset, for example, will work much better with more than one core and a decent amount of RAM.

This is also quite noticeable in other tasks outside of Trainz. I use a MIDI-based music program called Pianoteq to drive my high-end Roland LX-17 digital piano. This software makes use of multiple CPU cores when converting the MIDI recorded performances into .wav files for further editing. On my 4-core laptop, the conversion from MIDI to .wav takes upwards of 15 minutes while my desktop with the 5960 can process this in a matter of a couple of minutes. These recordings are not small either with some topping 20 minutes if not longer. The final .wav file is about 500 GB or larger before editing.

The other thing to consider as PC_Ace mentioned is a faster GPU (video card) should have a fairly high-end processor to support it, otherwise, the CPU can't process the data handed to it by the video card quick enough. Remember the CPU is the heart more than the brain in a computer. It handles the processing of numbers and system requests as it receives interrupt calls and memory requests from the various peripherals. In the olden days, the CPU did the bulk of the work since they were generally faster than the graphics cards and other peripherals, but today the CPU is an equal partner since many of these devices, especially the graphics cards, have their own GPU (graphics processing unit), and substantial RAM. With GPU's being built to process numbers quickly, they are used more often than not now to handle that task, and the fast CPU is used to move the data out to the supporting hardware such as RAM, and disk.

All things being equal, having an under-powered CPU can cause what is called bottle-necking. What this means is the CPU is too slow to process the data given to it from the GPU and other peripherals so data sits there in memory waiting to process. The aim for computer hardware is to move data as quickly as possible and out for the users to use, and if things sit there in the wings waiting for processing, this causes bad performance.
 
Back
Top