Be more careful with markdown notation

Author: markmi

README.md

@@ -1,4 +1,4 @@
-##### Some_acpphint_curves_with_notes.md
+##### Some\_acpphint\_curves\_with\_notes.md
 This covers:
 
 0. Cortex-A57 cores with and without smaller cache level(s) being shared
@@ -36,7 +36,7 @@ Effects of competing for the shared RAM cache level(s) can be seen:
    (left side of goldenrod vs. blue curves)
 1. Computing is somewhat slower.
    (right side of goldenrod vs. blue curves)
-2. The curve structures show a RAM cache changes at around n*256 KiByte, for n
+2. The curve structures show a RAM cache changes at around n\*256 KiByte, for n
    being the number of hardware threads in use.
 3. The curve structures show a RAM cache change at about 8 MiByte.
 4. The 1-hardware-thread curves show a RAM cache change at about 32 KiByte.
@@ -116,7 +116,7 @@ Notes:
 0. The x and y ranges for all the plots for 1950X CCD/CCX 2/4 thread
    combinations are the same, making comparison/contrast easier.
 1. The curves for allowing (hardware) thread migration without SMT involvement
-   show very clear evidence of around n*512 KiByte and 16 MiByte RAM cache
+   show very clear evidence of around n\*512 KiByte and 16 MiByte RAM cache
    changes, n being the number of (hardware) threads.
 2. The shared smaller cache level(s) for SMT usage have competing usage and make
    them show messier transitions.
@@ -154,7 +154,7 @@ Notes:
 0. The x and y ranges for all the plots for 1950X CCD/CCX 2/4 thread
    combinations are the same, making comparison/contrast easier.
 1. The curves for allowing (hardware) thread migration without SMT involvement
-   show very clear evidence of around n*512 KiByte, 16 MiByte, and 32 MiByte RAM 
+   show very clear evidence of around n\*512 KiByte, 16 MiByte, and 32 MiByte RAM 
    cache changes, n being the number of (hardware) threads, except for the 2CCD 4CCX case that has more cache per (hardware) thread.
 2. The shared smaller cache level(s) for SMT usage have competing usage and make
    them show messier transitions.

Some_acpphint_curves_with_notes.md

@@ -1,6 +1,6 @@
-##### acpphint_example_data/LICENSE.md
+##### acpphint\_example\_data/LICENSE.md
 
-###### Licensing for generated-data-files, `.gp` files, and plot images (.png) . . .
+###### Licensing for generated-data-files, `.gp` files, and plot images (\*.png) . . .
 There is no intent for any of the generated data files, the example `.gp` files
 for use with gnuplot, or the plot images to have any licensing restrictions,
 unlike, for example, the C++ source code and the associated binary files.

acpphint_example_data/LICENSE.md

@@ -1,4 +1,4 @@
-#### acpphint_example_data/README.md
+#### acpphint\_example\_data/README.md
 In general these notes are as of when they were written.
 
 This covers:
@@ -32,8 +32,8 @@ WDK23_CA72-FBSD/
 WDK23_CA78C-FBSD/
 ```
 (I do not distribute binaries. But I also do not distribute the
-source sys_cpubinding_lgplv2_1_okay.cpp that I used to prove
-that a Linux was supportable for WANT_CPU_LOCKDOWN builds.
+source sys\_cpubinding\_lgplv2\_1\_okay.cpp that I used to prove
+that a Linux was supportable for WANT\_CPU\_LOCKDOWN builds.
 I just wanted to avoid covering yet more licensing criteria.)
 
 `acpphint_example_data/` also contains some `*.gp`
@@ -76,7 +76,7 @@ The fields are:
 1.  median trial's mean time per lap in sec
     (median from sort by mean time per lap)
 2.  QUality ImProvement/sec (a.k.a QUIPS)
-3.  Quality (a.k.a QUIPS * previously listed time per lap, 
+3.  Quality (a.k.a QUIPS \* previously listed time per lap, 
     like in the original pthread HINT)
 4.  kernel vectors total memory use
 

acpphint_example_data/README.md

@@ -1,4 +1,4 @@
-#### acpphint_src/Code_issues_vs_original_HINTs.md
+#### acpphint\_src/Code\_issues\_vs\_original\_HINTs.md
 This file documents properties of source code
 differences with the original serial and
 pthreads HINT benchmarks.
@@ -27,7 +27,7 @@ For this variation on HINT, RUNTM is computed instead
 of being directly adjustable. PATIENCE, MXPROC, DSREFS,
 ISREFS, and MSIZE are not used.
 
-For this variation on HINT, STILL_UNKNOWN was
+For this variation on HINT, STILL\_UNKNOWN was
 added to EFlag as the first value.
 
 The original serial and pthread HINTs did not use
@@ -58,7 +58,7 @@ for readability):
                                    * sizeof(ISIZE) * 2));
 ```
 but it later miscalculated the memory use for those (missing
-any '*2' or tracking of the largest index actually used
+any '\*2' or tracking of the largest index actually used
 in errs and in ixes):
 ```
     memuse = (int)(qdata[i-1].n
@@ -168,7 +168,7 @@ That is a problem vs the description of MXPROC: "Maximum
 number of processors to use in shared memory configuration".
 With each "processor" having HI and LO entries in gs the
 maximum "processor" count is MXPROC/2 in order to avoid
-trashing memory and accessing unowned memory. "2*" could be
+trashing memory and accessing unowned memory. "2\*" could be
 considered as missing from the gs declaration. Or the hint.h
 description could be considered wrong.
 
@@ -275,7 +275,7 @@ initialization was very different: It was based on setting up
 the scatter decomposition.
 
 At no stage was there an explicit single cell spanning the whole
-scx*scy area so no individual area calculation spanned a big
+scx\*scy area so no individual area calculation spanned a big
 enough area to have overflow/truncation/rounding issues. Nor did
 any sum of areas that was peformed.
 

acpphint_src/Code_issues_vs_original_HINTs.md

@@ -1,4 +1,4 @@
-#### acpphint_src/README.md
+#### acpphint\_src/README.md
 This covers:
 1.  What `acpphint` is a variation on . . .
 2.  How to build . . .
@@ -133,7 +133,7 @@ say to make builds take less time.
 
 The default currently is to build ul (unsigned long) and which ever of ull
 (unsigned long long) or ui (unsigned int) that have the same maximum value
-as unsigned long. There is a macro DSIZE_ALL_ISIZE_ALL that, when defined,
+as unsigned long. There is a macro DSIZE\_ALL\_ISIZE\_ALL that, when defined,
 leads to a wide variety of DSIZE ISIZE combinations in most of the programs.
 
 `acpphint_kernels.cpp` has, for example (abbreviated/line-split),
@@ -248,7 +248,7 @@ See `../*.md` for more overall information.
 Overall, each `acpphint*-NAME` is licensed as GPLv2 (not earlier, not later).
 But building the acpphint related code does build some OpenBSD-vintage-ISC
 licensed source files, those under `other_src_used/` instead of under
-`acpphint_src/`. Some resulting object files are linked into the acpphint*-NAME
+`acpphint_src/`. Some resulting object files are linked into the acpphint\*-NAME
 related programs.
 
 See the `acpphint_src/LICENSE.md` for the GPLv2 license text that applies to the 
@@ -257,7 +257,7 @@ Programs that do not have names that start with `acpphint` do not contain code
 from `acpphint_src/` ( just from `../other_src_used/` ).
 
 See the `../other_src_used/LICENSE.md` for the OpenBSD-vintage-ISC license text
-that applies to just the source files in other_src_used/ and to any binaries not 
+that applies to just the source files in other\_src\_used/ and to any binaries not 
 containing code from `acpphint_src/` source files.
 
 Note: The original HINT's copyright holder, Iowa State University Research

acpphint_src/README.md

@@ -1,4 +1,4 @@
-#### acpphint_src/Usage_differences_with_original_HINTs.md
+#### acpphint\_src/Usage\_differences\_with\_original\_HINTs.md
 This file documents differences with the original serial and
 pthreads HINT benchmarks that matter for use.
 
@@ -23,7 +23,7 @@ format appropriate for at least gnuplot):
 1.  median trail's mean time per lap in sec 
     (median from sort by mean time per lap)
 2.  QUality ImProvement/sec (a.k.a QUIPS)
-3.  Quality (a.k.a QUIPS * previously listed time per lap, 
+3.  Quality (a.k.a QUIPS \* previously listed time per lap, 
     like in the original pthread HINT)
 4.  kernel vectors total memory use
     (not an understated figure like the original
@@ -51,7 +51,7 @@ HINT counts.
 ##### argv[?] use . . .
 The output files have the argv[0] text as the file name
 prefix and text after that identifying the specific test.
-The example makefile.* files generate very long command
+The example makefile.\* files generate very long command
 names that encode lots of context information.
 
 The default filename suffix is ".txt" .
@@ -124,7 +124,7 @@ for readability):
                                     * sizeof(ISIZE) * 2));
 ```
 but later miscalculated the memory use for those (missing any
-'*2'):
+\*2):
 ```
     memuse = (int)(qdata[i-1].n
            * (sizeof(RECT)+sizeof(DSIZE)+sizeof(ISIZE)));

acpphint_src/Usage_differences_with_original_HINTs.md

@@ -1,7 +1,7 @@
-#### other_src_used/README.md
+#### other\_src\_used/README.md
 `acpphint` is a variation on the original HINT benchmark, but written in C++ and with various choices specific to my use.
 
-While, overall, `acpphint` is licensed as GPLv2 (not earlier, not later), building the `acpphint` related code does build some OpenBSD-vintage-ISC licensed source files, these are under `other_src_used/` . Some of these object files are linked into the acpphint related programs.  There is one program (cpp_clock_info_main) which only uses `other_src_used/` source code (including itself). The ISC licensed source code and these object files are not derivatives of the original HINT benchmark code and have other potential uses.
+While, overall, `acpphint` is licensed as GPLv2 (not earlier, not later), building the `acpphint` related code does build some OpenBSD-vintage-ISC licensed source files, these are under `other_src_used/` . Some of these object files are linked into the acpphint related programs.  There is one program (cpp\_clock\_info\_main) which only uses `other_src_used/` source code (including itself). The ISC licensed source code and these object files are not derivatives of the original HINT benchmark code and have other potential uses.
 
 See the `other_src_used/LICENSE.md` for the OpenBSD-vintage-ISC license text that applies to just the source files in `other_src_used/` . Each distributed `other_src_used/` C++ source file also contains the OpenBSD-vintage-ISC license text.