CPU selection and compilation speed

I need to buy a new desktop pc and I am researching CPU selection.
The most important parameters are speed of compilation and efficient processing of big datasets with polars. As examples to show the approximate data size, multiple parquet files 100+ GB in total or one bigger parquet file around 30 GB. Most of the time the datasets are significantly smaller though.

The new system will run on Arch Linux with 64 or 128 GB of memory.

My options for CPU are as I see it currently the AMD Ryzen 9 7950X or the AMD Ryzen 9 7950X3D. The essential difference between the two is 3D V-Cash. The X3D has around 60 MB of additional cache and is more energy efficient.

The question is, if such a difference between the CPUs will make a difference for my use case. Does anyone have a similar configuration?

Thanks a lot in advance for an tip or help.

I would wager that one could drop any of those processors into similarly spec'ed systems and hardly notice any difference in performance. When handling huge data sets other system components will dominate, memory speeds, SSD speeds etc.

I understand. And what about the compilation speed?

1 Like

Afaik the extra cache is mainly useful in gaming.

GamersNexus said that their clang benchmark used to be just a measure of CPU cache. So it might well be that compiling rust -- which is basically a giant hash table benchmark -- would also be much faster with a boatload of cache.

I don't have numbers for it, though. I'm still on a 3950X.

How is the performance on 3950X ?

True. But that is part of the reason I am asking. I was hoping to have some input from more experienced members and ideally members who use one of those cpu. Thanks anyways

according to AMD Ryzen 9 7950X vs. AMD Ryzen 9 7950X3D vs. AMD Ryzen 9 9950X vs. AMD Ryzen 9 3950X Benchmarks - OpenBenchmarking.org
for compiling Wasmer (a Rust compilation benchmark):

AMD Ryzen 9 9950X 16-Core AMD Ryzen 9 7950X 16-Core AMD Ryzen 9 7950X3D 16-Core AMD Ryzen 9 3950X 16-Core
23.00 sec 31.00 sec 33.00 sec 55.00 sec

so the 7950X is slightly better than the 7950X3D, but the 9950X is waay better than both.

Also, note that it's 3D V-Cache, not 3D V-Cash, since it's referring to the CPU's L3 cache, not to money.

3 Likes

if you can wait a few months and don't mind spending some additional money, the 9950X3D is likely faster than all of those, based off how much faster the 9800X3D is compared to the 9700X:

Timed LLVM Compilation

AMD Ryzen 7 9800X3D 8-Core AMD Ryzen 7 9700X 8-Core
485 sec 561 sec
4 Likes

Thanks a lot. Very useful info. After posting the question here, I continued to search for info online and it appears that the only strong point in 7950X3D is it's very good energy efficiency.