I'm currently running a C2D E4300 overclocked to 2.4GHz (Stock is 1.8GHz)...

I did some tests with Norton Utilities a few weeks ago to get a decent idea of what the emulated CPU speed was on an Athlon X2 4000+, and did the same when I got my Core 2 machine shipped back to me.

The DOSBox VM uses the Dynamic core (Max Cycles), SB16 @ 44,100Hz, VGA graphics (Original FS res, no scalers, OpenGL blitting) and no frameskip. GUS, Sound Source and COM ports are disabled.

Since Max cycles make Norton's CPU benchmark results jump as the cycles change, I took a screenshot during the average speed listing. Tests were run on both the Linux and Win32 builds (Performance is basically the same).

Machine 1:

Athlon 64 X2 4000+
1GB DDR667
GeForce 6150LE
Integrated Sound

This pegs the emulated CPU at about the speed of a Pentium 133MHz. I ran a bunch of games and old DOS console emulators and the result confirmed this level of performance.

Blood is very playable at 320x240, and plays passably at 640x480. System Shock is also pretty good at 640x480. Even MP3 playback under a Windows 3.11 install works great, with only mild stuttering if you do too much at once.

Machine 2:

Core 2 Duo E4300 @ 2.4GHz
2GB DDR533
GeForce 7600GT
Integrated Sound

Benches and real world tests peg this at the level of a Pentium Pro 233MHz. Blood runs extremely fast at 640x480 and is very playable at 800x600. Quake runs very well at 640x480, etc.

I didn't take screenshots, but with the Normal core the Athlon scored close to a 486 33MHz under Norton, and the Core 2 just slightly lower than a Pentium 66.

In conclusion, even a low end Core 2 (With a little overclocking) can run basically any DOS app at *very* good speeds.

my a64 3200 runs like a p2 233 though

Funny: I used to run .70 on an Athlon 64 3000+ and Blood had trouble pushing full speed at 320x240 (Blood is one of my favorite benchmarks since I actually owned the game and played it across several generations of PCs I've owned).

This is something a PII at that speed would have absolutely no problem doing, and the difference in speed between a 3000+ and a 3200+ is negligible.

It's difficult to compare DosBox performance across different computers due to many factors:

Crapped up Windows installs over the years. (Soyware/AV/misc hogging processor cycles/memory)
Power savings modes on processors.
Old DosBox versions.
Improper DosBox configurations.
Crappy video/Sound drivers.
Malfunctioning computer hardware
Using DosBox with other games/programs loaded.
etc.

There's also the fact of people turning on FPS counters on old DOS games and claiming that DosBox is slow because they are not getting 60+ fps's in their old DOS games. Even though you were lucky to get more than 15-20+ unless you high-end computers back in the day......or comparing DosBox to NTVDM where the games have hundreds of FPS's per second while DosBox only reaches 30+. (Of course these people just care about speed and benchmarking instead of actual gameplay and compatibility)

The only true comparisons that would actually be worth anything are clean installs of the OS fully updated with the latest drivers......a Linux LiveCD with DosBox would probably be the best bet.

Oh, I'm something of a ricer with operating systems. I build my Linux installs from the CLI up, and my Windows installs are stripped, hyper optimized builds of either Windows 2000 or XP Pro (Got to love nLite and 3+ years as a Windows sysadmin/network tech).

If anything, my results should be faster than norm, not the other way around. 😒

My x2 4200 has an average in the 500s. I still remember all the sysinfo results from my machines 😀. 300 is closer to p100, my p133 managed around 420 😁
But DOSBox does not emulate all the instructions with the same speed so the cpu benchmark is not accurate enough. Games might perform better or worse...