This is a great setup. The notes I have are few, and are really perhaps limited to hardware enthusiasts:
The 13900K is notoriously hot-running, and you may find even a 360 AIO will struggle to prevent bottlenecking under all-core CPU rendering load (if that’s what you need to do).
It’s still a little bit pricey for the RTX 4070 Ti, but for Germany that is okay. Check benchmarks to see if you are really likely to see the benefit in spending an extra 30-60 Euros over an RTX 4070 Super.
I think your “weak” point here is the 16 GB of RAM. I would definitely bump that to 32 GB, as Windows will happily take 7-8 GB by itself. But it will work okay with 16 GB. If you keep to 16, remember that in reality, your scene renders will be limited to probably 7 GB or so if you are using the CPU.
For re-use, may may find a slightly higher-power PSU (850 W) may give you better upgradability in the future.
Really minor. I think you are buying an overclocking class CPU for a motherboard that does not support overclocking.
Your choice of Intel is absolutely fine, but as many people correctly state, gives you no (potential) upgrade route compared to AMD AM5 platform. But a 13900K will be very capable for many, many years.
Thx for the replay!
For the RAM I am buying 2x8GB x2, means 32GB,
I looked at a benchmark and i think u r right with the Super.
I am looking forwards to search a “better” motherboard
There’s no need to go for a silly expensive motherboard. Z790 would be nice, but nobody needs 20 power stages. One thing to note with your current choice is the DMI interface, which is slower than the other Raptor lake chipset DMIs.
As for RAM, I think the typical approach is to use 2 x 16 GB over 2 x 2 x 8 GB, as the sticks tend to be paired for timing at the factory. Two sticks I seem to recall require less management by the memory controller then 4. It may not really matter for productivity.
For the record, I have a 4070 Super, and chose it (sort of) on the same basis as my reasoning above. Actually, in my case it was more extreme, as I was looking at the 4070 Ti Super.
Yeah. Ultimately, the 4070 Ti will marginally outperform the 4070 Super. But they get the same VRAM.
One great resource to chose cards is TechTesters on YouTube. They are really good at providing measured comparisons between different OEM cards of the same species. I’d guess a TUF should be in the top half for cooling and noise. But performance wise, they will nearly all be the same by definition.
I have had Rhino crash from lack of memory when rendering a model that was in a 4.5 Gb file.
I had to up my memory from 32 Gb.
I now have 64 Gb on my laptop (I edit on it, and develop plugins, but rarely render on it).
I have 128 Gb on this desktop.
I don’t know that the CPU matters unless you are doing CPU based rendering. Any time I see the Rhino editor struggle, its almost never due to CPU constraints.
I have your case, and the rest of your specs look similar to my desktop.
Though my C drive is a 3.63 Gb M2, and my D drive is similar. But I have a lot of renders and several Yocto builds (Embedded Linux, kernel, software, and firmware build, which use caches for speed) so my C drive was running dangerously low. I also save regular copies of my Rhino files to dropbox, and keep them locally on my desktop. Dropbox is my offline backup, poor man’s version control, and transfer mechanism between machines. I have fibre broadband.
I’ve only put your components in this tool (the SSD is different…)
it gives no warnings… but double check if the parts are correct…
Whatever, you always want good performance in single thread tasks.
That CPU is currently in the top 15, looking at single thread performance.
Without looking at the performance/price ratio (aka when budget doesn’t matter) that will surely be a good choice.
Personally I don’t have the need of “my pc need to be the fastest possible today” … so I would stay a bit away from the top list and increasing a bit the performance/price ratio… maybe a total o 1500 more or less…
Consider buying an additional internal HDD/SSD… for internal backup or else. It’s always useful to have another drive that is not the same of the OS.
One thing that springs to mind with TUF cards is the height. If you are putting it directly to the motherboard, check that the 12VHPWR power cable has enough clearance from the top of the shroud. I seem to recall TUF cards are quite tall, and the cable may hit the glass.
Really doesn’t matter. It’s just the exchange speed of indirect interfaces to the CPU. It affects media and likely the lower PCI-e slots. Don’t worry about it unless you are doing huge amounts of data exchange I think.
My only advice is to just check the GPU + cable clearance, the build seems fine, and enjoy the build. Seems that you have a good selection of parts.
Anything else will just get needlessly convoluted and unlikely to make any difference to your life.
I am finding the following of use. Maybe likewise for you [If I am not to late with this post]. As I am working, I open up Win Task Manager and then check as to what and the amount of X resource[s] are being used for a certain task [ or when running concurrent tasks]. e.g. In Ray Trace mode what resources are being used to render the model of Y size etc. Are some being maximally used ? / the process is taking a long time to finish ? After awhile [ weeks for me] of doing this I get a better idea of what the specifications of a new computer build might be for my work.
