X-Git-Url: http://git.mutantstargoat.com/user/nuclear/?p=dosdemo;a=blobdiff_plain;f=README.md;h=496d3957e442c5a064c9cd3ed10c416de63e9e41;hp=4804bd7ca79e980a022a261d7658bb3f78ff2564;hb=b6c89d9ffbb08c5286448de4773290f0924a420b;hpb=ecc362e702b77d5c1334656e6f4dc9b17abcb767

diff --git a/README.md b/README.md
index 4804bd7..496d395 100644
--- a/README.md
+++ b/README.md
@@ -25,11 +25,62 @@ which will in turn start dosbox, which will execute the DOS binary! If the gods
 are slumbering in valhalla, just typing `dosbox demo.exe` should do the trick.
 
 
-SDL backend
------------
-Run make to build (assuming make on your system is GNU make).
+Building and running with the SDL backend
+-----------------------------------------
+Run make to build (assuming make on your system is GNU make), or use the visual
+studio 2013 project on Windows.
 
 The SDL backend will scale the framebuffer up, by the factor specified in the
 `FBSCALE` environment variable. So run the demo as: `FBSCALE=3 ./demo` for
 a 3x scale factor, or just export the `FBSCALE` env var in the shell you are
 going to use for running the demo. The default scale factor is 2x.
+
+Datafiles
+---------
+The demo datafiles are in their own subversion repo. To checkout the data files
+run the following in the demo root directory:
+
+  svn co svn://mutantstargoat.com/datadirs/dosdemo data
+
+Random optimization details about the Pentium1 (p54c)
+-----------------------------------------------------
+Use cround64 (util.h) for float -> integer conversions, instead of casts.
+
+Performance measurement with RDTSC:
+    perf_start();
+    /* code under test */
+    perf_end(); /* result in perf_interval_count */
+
+Cache organization (L1): 8kb data / 8kb instruction
+128 sets of 2 cache lines, 32 bytes per cache line.
+
+Addresses which are multiples of 4096 fall in the same set and can only have
+two of them in cache at any time.
+
+U/V pipe pairing rules:
+ - both instructions must be simple
+ - no read-after-write or write-after-write reg dependencies
+ - no displacement AND immediate in either instruction
+ - instr. with prefixes (except 0x0f) can only run on U pipe.
+ - prefixes are treated as separate 1-byte instructions (except 0x0f).
+ - branches can be paired if they are the second instr. of the pair only.
+
+Simple instructions are:
+ - mov reg, reg/mem/imm
+ - mov mem, reg/imm
+ - alu reg, reg/mem/imm (alu: add/sub/cmp/and/or/xor)
+ - alu mem, reg/imm
+ - inc reg/mem
+ - dec reg/mem
+ - push reg/mem
+ - pop reg
+ - lea reg,mem
+ - jmp/call/jcc near
+ - nop
+ - test reg,reg/mem
+ - test acc,imm
+
+U-only pairable instructions:
+ - adc, sbb
+ - shr, sar, shl, sal with immediate
+ - ror, rol, rcr, rcl with immediate=1