Why don't you try to work with a single-core processor if the single-thread performance is so important?
Despite being so active in these deeply technical subjects, you clearly don't know what "single threaded performance" means.
You should think of single threaded performance as
performance per core, because that's what it really is, and forms the theoretical upper limit of multithreaded scaling; cores * performance per core.
All workloads split over multiple cores will encounter diminishing returns with increased core count, as synchronizing more cores is inevitably going to take more time. This overhead might not be significant if you're doing a large batch job that takes minutes or even hours, but if it's an interactive application or a game, then you have a very critical time limit before the application becomes laggy and non-responsive. Since there is an overhead cost with each thread you synchronize, balancing how threads share data and the size of work chunks is essential for good multithreaded performance. In such cases faster cores will lead to better utilization and less stalls and lag, essentially you can scale to more cores before performance gains become negligible.
As I said in my previous post, single threaded performance helps "everything". Whether an application uses 1 or 128 threads, an increase in single threaded performance is nearly always going to benefit a computational workload, and sometimes even help multithreaded performance even more due to less overhead.
The applications are thread-count starved - we all need more cores because we are always limited by the speed of execution of a single thread, the only way to overcome this limit is to use more cores.
If an application is in fact starved for more threads, then more threads are good.
But applications have to be carefully designed to scale well. Large batch jobs are "easy", while applications like Phoshop etc. are harder. That's why you often see with such applications that more cores helps a little up to a point, but faster cores always help.
Unless an application benefits from thread isolation (which some web server tasks do), faster cores are always going to perform better than more cores. If you have the option between a CPU that has 50% more cores or one that's 50% faster per core, the latter will nearly always win.
Oh, and to be honest - this PassMark thread is heavily AMD optimised.
Typical rookie mistake.
There are really no such thing as "AMD optimized" or "Intel optimized".
Just because a piece of software is performing better on one specific piece of hardware, doesn't mean it's "optimized" for it. In 99.9% of cases it simply comes down to the resource balance. The exception are the few cases where unique ISA features are utilized, or the application intentionally runs a slower code path for certain hardware.
