Rough comparison to Teraflops

[David (PassMark)]

FLOPs is a measure of floating point operations per second.
But it is rather vague. As it doesn't really define
- the exact type of operation (additional, multiplication, square roots, etc..)
- the data size (64bit, 32bit, 16bit, etc...)
- the size of the data buffer and if that buffer can fit into the CPU's cache
- the rounding mode used
- single thread or multi-thread
- if AVX & other SIMD instructions can be used
- if the mix of instructions and variables allows hyper threading and pipelining to be effectively used.
- etc....

The closest measure in PerformanceTest is the "Floating point maths" test which returns a result in Millions of operations per second (MOps)
So, for example, my i7-4770 CPU 3.4Ghz 4 core 8 threads, has a result of 13,088 MOps.

So ~13 GigaFLOPs for a i7-4770.

Which is almost 4 operations per clock cycle.

[Johnksellers]

I am curious about cpubrnchmark.net benchmarks.

Is the numeric result roughly proportional to Flops?

And, what would the benchmark result that would roughly comparable to one Teraflop?