/[gxemul]/trunk/doc/technical.html
This is repository of my old source code which isn't updated any more. Go to git.rot13.org for current projects!
ViewVC logotype

Diff of /trunk/doc/technical.html

Parent Directory Parent Directory | Revision Log Revision Log | View Patch Patch

revision 23 by dpavlin, Mon Oct 8 16:19:37 2007 UTC revision 24 by dpavlin, Mon Oct 8 16:19:56 2007 UTC
# Line 10  Line 10 
10    
11  <!--  <!--
12    
13  $Id: technical.html,v 1.72 2006/02/18 15:18:15 debug Exp $  $Id: technical.html,v 1.74 2006/06/17 10:16:22 debug Exp $
14    
15  Copyright (C) 2004-2006  Anders Gavare.  All rights reserved.  Copyright (C) 2004-2006  Anders Gavare.  All rights reserved.
16    
# Line 70  because the host architecture and emulat Line 70  because the host architecture and emulat
70  compared just like that.  compared just like that.
71    
72  <p>Performance depends on several factors, including (but not limited to)    <p>Performance depends on several factors, including (but not limited to)  
73  host architecture, host clock speed, which compiler and compiler flags  host architecture, target architecture, host clock speed, which compiler
74  were used to build the emulator, what the workload is, and so on. For  and compiler flags were used to build the emulator, what the workload is,
75  example, if an emulated operating system tries to read a block from disk,  what additional runtime flags are given to the emulator, and so on.
76  from its point of view the read was instantaneous (no waiting). So 1 MIPS  
77  in an emulated OS might have taken more than one million instructions on a  <p>Devices are generally not timing-accurate: for example, if an emulated
78  real machine.  operating system tries to read a block from disk, from its point of view
79    the read was instantaneous (no waiting). So 1 MIPS in an emulated OS might
80    have taken more than one million instructions on a real machine.
81    
82  <p>Also, if the emulator says it has executed 1 million instructions, and  <p>Also, if the emulator says it has executed 1 million instructions, and
83  the CPU family in question was capable of scalar execution (i.e. one cycle  the CPU family in question was capable of scalar execution (i.e. one cycle
# Line 85  penalties that are not simulated by GXem Line 87  penalties that are not simulated by GXem
87    
88  <p>Because of these issues, it is in my opinion best to measure  <p>Because of these issues, it is in my opinion best to measure
89  performance as the actual (real-world) time it takes to perform a task  performance as the actual (real-world) time it takes to perform a task
90  with the emulator. Typical examples would be "How long does it take to  with the emulator, e.g.:
 install NetBSD?", or "How long does it take to compile XYZ inside NetBSD  
 in the emulator?".  
   
 <p>So, how fast is it? :-)&nbsp;&nbsp;&nbsp;Answer: it varies.  
   
 <p>The emulation technique used varies depending on which processor type  
 is being emulated. (One of my main goals with GXemul is to experiment with  
 different kinds of emulation, so these might change in the future.)  
91    
92  <ul>  <ul>
93    <li><b>MIPS:</b><br>    <li>"How long does it take to install NetBSD onto a disk image?"
94          There are two emulation modes. The most important one is an    <li>"How long does it take to compile XYZ inside NetBSD
95          implementation of a <i>dynamic binary translator</i>.          in the emulator?".
         (Compared to real binary translators, though, GXemul's bintrans  
         subsystem is very simple and does not perform very well.)  
         This mode can be used on Alpha and i386 host. The other emulation  
         mode is simple interpretation, where an instruction is read from  
         emulated memory, and interpreted one-at-a-time. (Slow, but it  
         works. It can be forcefully used by using the <tt>-B</tt> command  
         line option.)  
   <p>  
   <li><b>All other modes:</b><br>  
         These use a kind of dynamic translation system. This system does  
         not recompile anything into native code, it only uses tables of  
         pointers to functions written in (sometimes machine-generated) C  
         code. Speed is lower than what can be achieved using real binary  
         translation into native code, but higher than when traditional  
         interpretation is used. With some tricks, it will hopefully still  
         give reasonable speed. The ARM and PowerPC  
         emulation modes use this kind of translation.  
96  </ul>  </ul>
97    
98    <p>So, how fast is it? :-)&nbsp;&nbsp;&nbsp;Answer: it varies.
99    
100    
101    
102    
103    
Legend:
Removed from v.23  
changed lines
  Added in v.24
  ViewVC Help
Powered by ViewVC 1.1.26