WARNING:
Benchmarks are worth nothing in real life, they are lab tests that give results which only indicate performance for that specific test.
But anyway, they give an aproximate overview concept of the performance of a specific system, never an acurate one though.
This comparison is totally uncientific and utterly wrong.

So, after all this nonsense said to avoid trolls, what follows is some interesting results for the Amiga geeks out there.
When 68k Amigas are mentioned, amiga users usually try to refer to a particular x86 equivalent, so as to compare its performance.
So that is what I actually did using the Linux BogoMIPS test, on both sytstems and got the following conclusions:

Performance equivalences aproximates (taken from Linux BogoMIPS benchmarks)

68000 7                = 8086 5
68020 14 A1200 no FAST = 286 8
68020 14 w/fastmem =     386DX 20
68030 16               = 386DX 25
68030 25               = 386DX 40
68030 40               = 486SX 20
68030 50               = 486DX 25
68040 25               = 486DX 33
68040 40               = pentium 66      
68040 50 overclocked =  pentium 90
68060 50               = pentium 233 (no MMX type cpu)




Atari vs Amiga benchmark test (linux BogoMIPS again)

Amiga 3000 16 mhz    3.97 BogoMIPS
Atari Falcon 16 mhz  3.95 BogoMIPS

Amiga is faster!!!!!
Just by a very small margin =0


PS: I am sorry, i was bored

its sure close in some circumstances, but my table from "real life" benchmarks are like that:
68000 7 = 8086 5
68020 14 A1200 no FAST = 286 12
68020 14 w/fastmem = 386DX 20
68030 16 = 386DX 16
68030 25 = 386DX 25
68030 40 = 386DX 40
68030 50 = 486SX 25
68040 25 = 486DX 33
68040 40 = 486DX2 66
68060 50 = pentium 90-100(no MMX type cpu)

@StormLord
an Amiga 3000 16mhz slower than an Amiga 1200 with fastmem? I believe you missed on that one. The other ones look reasonable

as i have seen, things are like this,

020 and 030 are the same speed as it comes to mhz per mhz, ofcourse there are 030's with more Mhz than 020's

040 is 4x faster than 030 in the same Mhz
a 040@25 is twice faster than 030@50

060 is about 2x than 040 in the same mhz because of the superscalar.


so 040 WAS a BIG progress!

Again, yes you are right, the 68040 was a great breakthru.
If using 68000 and 68020 programs in a 68060, the performance compared to the 68040 at  the same clock rate is the same.

Except if the cache size (8 vs. 4 kB) makes a difference. Or different FPU.