| [hatari-devel] Hatari emulated performance vs Falcon |
[ Thread Index |
Date Index
| More lists.tuxfamily.org/hatari-devel Archives
]
- To: hatari-devel@xxxxxxxxxxxxxxxxxxx
- Subject: [hatari-devel] Hatari emulated performance vs Falcon
- From: Douglas Little <doug694@xxxxxxxxxxxxxx>
- Date: Fri, 1 Feb 2013 16:07:03 +0000
- Dkim-signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=googlemail.com; s=20120113; h=mime-version:x-received:date:message-id:subject:from:to :content-type; bh=c0EHo3vTctWeWnwcqGXT1KSnVJUIJJMOJT2I315WTZU=; b=LyKDw0L7rGBWtLXFGuxIs13ReiganxfYOoZj9PZ9nzqYF2m4lGppM7nZylF8Jc7jWF tZ6bS08C4MYdyIiq+L5rc91puVTdSKdMsltKmKUBwG2JGy/JORLzzFw26GULXEZyQ/Cw c3ru7EtdpzcyfwLI/MmUa+AMBPr4gBlUJuqNKT+KpTiyWz0M5845Kxqcrzp6oVuD+0L4 slxKo48XEfUx9i7h5cDf6bnEXnR8//zXGNe5d0qVua1vmz99OvIySTvRU3kOTwSAUrpR orbs+vdUTknU9OCTOk6LJFlsP+3YqHEHdYhMCU3XnKgO+GcxrEPuYFDbnWIt6hH5TSWu Dcng==
Hi,
I have been running some experiments in the background today, comparing Hatari against a real Falcon, for the BadMood project.
1) It's already understood that the d-cache is not taken into account and that does affect timing in some places, but I'm looking at other differences which don't involve the d-cache.
2) We have also recognized a difference in some DSP timings (discussed earlier) so I'll ignore that for now too.
Would I be correct in assuming that other timings should be close most of the time? I'm not yet sure if the i-cache behaviour is supposed to influence timing currently (My tests suggest the i-cache timings are also ignored).
So would I be correct in assuming that comparisons between Falcon and Hatari are best conducted with BOTH caches disabled?
Thanks,
D.