Hey guys, it's been a while since I've contributed anything here, but over the last week I had some free time.
Last week I took 8 different finals, and in the time after the tests I spent crafting a tool to help me with a new passion of mine.

Speed cubing, or the process of competitively solving a Rubik's cube.

At first I included a simple timer, that was originally measured in milliseconds, but after further work I formatted the time into a 00:00.00 format and added a list of previous times, automagic saving, and statistic keeping (best, total average, average of 10, soon to be median 3 of 5 and 10 of 12)

So, the controls:
Enter: Start/stop the timer
"Y"/"N": After stopping the timer you have the option whether or not to add it to the list of times, the options should be obvious
"D": Clears the current list of times
"S"/"R": Saves/Recalls one saved list of times

Since I wrote this on calc, and not with the assistance of an emulator, I have no screenshots for you, but here's the code (hopefully it works after transcribing it from my calc)

I'll probably add new features here as I make them.
Note: You can just compile this with luna or something and run it on calc.

So if any of you guys are cubers, maybe you'll use this to practice when you can't access a computer 

tn, msg = 0, "";
t, n, aa, a10, b = 0, 0, 0, 0, 0
save = false
times = {};

function a()
local min, avg10, avga = 2^32, 0, 0
for i = 1, tn do
local v = times[i]
if v < min then
min = v
end
if i <= 10 then
avg10 = avg10 + v
end
avga = avga + v
end
b = min
aa = math.floor(avga/tn)
a10 = math.floor(avg10 / math.min(tn, 10))
end

function f(t)
return ("%02d:%02d.%02d"):format(t/60000, t/1000%60, t/10%100)
end

function on.paint(g)
if c then
t = (timer.getMilliSecCounter() - n)
end
g:drawString("time: " .. f(t), 0, 0, "top");
g:drawString(msg, 0, 20, "top");
g:drawString("best:              " .. f(b), 150, 0, "top");
g:drawString("average (10): " .. f(a10), 150, 20, "top");
g:drawString("average (all): " .. f(aa), 150, 40, "top");
for i = 1, 8 do
local v = times[i];
if not v then break end
g:drawString(i .. ")      " .. f(v), 0, 20+i*20, "top")
end
timer.start(0.01);
end

function on.timer()
timer.stop()
platform.window:invalidate()
end

function on.charIn(c)
if save then
save = false
msg = ""
if c == "y" then
table.insert(times, 1, t)
tn = tn + 1
a()
on.charIn("s")
end
end
if c == "d" then
t = 0
times = {};
tn = 0;
msg = "";
a();
elseif c == "s" then
local s = "" .. tn
msg = "saved"
for i = 1, tn do
s = s .. " " .. times[i]
end
var.store("cube_times", s .. " ");
elseif c == "r" then
local s, b , n = var.recall("cube_times") or "0 ", {}, 0;
msg = "loaded";
for w in s:gmatch("%d+%s") do
n = n + 1;
b[n] = tonumber(w);
end
tn = table.remove(b, 1);
times = b;
a();
end
end

function on.enterKey()
 if save then
 save = false
msg = ""
return
end
if c then
save = true;
msg = "save? (y/n)";
c = nil;
else
c = true;
msg = "";
n = timer.getMilliSecCounter();
end
end

on.charIn("r")



I am using assembly, but this question about linking in general, so this section seems the most low-level and appropriate.

I've dabbled with assembly a few times before, but one problem that always bugged me is the linker I used.
I already had MinGW installed, so I just used GCC to link my object files, however this always created huge, bloated files.
I suspect this is from including crap I don't need, but I have no idea how to disable this.

So, what suggestions would you guys make for a Win7 x86/x64 Linker, that is either non bloaty like GCC, or it is easy to disable all this bloat.

Tips for disabling GCC's bloat are accepted as well.

Note: I'm using NASM as my assembler.

What are your opinions of this commercial ?
I'm personally rather surprised by hearing Dubstep in a commercial.

So, I somehow was linked to one of Taylor Mali's poetry videos on youtube, and I was absolutely hooked.
I think he's amazing, but I want to hear what you guys think

here's a list of some of his videos I've watched on youtube:

Welp, I am in need of more posts before I can join the IRC apparently, so I will introduce you all to a project I'm working on

Since Lua is currently the fastest programming language for up-to-date NSpire calculators, I thought I would share this relevant project.
This really is for those who want to get the absolute most out of Lua. It's very impractical to write an entire program from Lua assembly, since the compiler does a decent job for most non-time-critical functions.

Lua is actually a semi-compiled language (at least that's what I'm going to call it).
It is compiled to bytecode, and than interpreted dynamically on a virtual machine.

You can execute Lua bytecode directly in Lua with the loadstring() function.

loadstring("\27\76\117\97\81\0\1\4\4\4\8\0\56\0\0\0\64\67\58\47\85\115\101\114\115\47\74\97\115\111\110\47\68\111\99\117\109\101\110\116\115\47\80\114\111\103\114\97\109\109\105\110\103\47\76\117\97\47\77\79\68\83\47\116\101\115\116\46\108\117\97\0\13\0\0\0\13\0\0\0\0\0\0\2\1\0\0\0\30\0\128\0\0\0\0\0\0\0\0\0\1\0\0\0\13\0\0\0\0\0\0\0\0\0\0\0")();

Now I'd go over the bytecode format, but you can learn all of that here
This document also includes all of the information on all of the instructions to the Lua VM.

Now, onto the project.
(In retrospect, the name I chose was rather silly, but meh, it applies)

Basically, it includes the following:

• An assembler for a custom syntax of my own design (feature ridden)
• Something I would call a Linker, that takes a table of all the relevant information, and compiles it into Lua bytecode
• Soon to be a disassembler, which will turn a normal Lua function into Lua assembly.

You can find the repository here: https://github.com/NecroBumpist/MODS

So, time for "Hello World!" in Lua assembly
local mods = require("MODS.Current.main");
local bytecode = mods.Assemble(
[[
; Hello_World.lasm
.options 0, 0, 0, 2        ; a directive, which sets some properties, the '2' meaning MaxStackSize

getglobal    0,    <'print'>        ; <> denotes a constant, i've yet to bother with a directive for doing constants
loadk        1,    <'Hello World!'> ; The LuaVM is register based, we put the function in one register, then arguments above it
call         0,    2,    1          ; now we call register 0, with (2-1) parameters, accepting (1-1) values back
return       0,    0                ; and now we return
]]);

loadstring(bytecode)() -- turns the bytecode into a function, then we call that function

>Hello World!


So,  I'd like to hear what some of you Lua users who are looking for speed improvements think
I plan on updating the repository soon to include:

• Write some proper documentation
• Updates to the assembler to include NASM like Macros (debugging right now)

In ten or twenty minutes the repository should be much cleaner

Comments, questions, or concerns ? Or Lua Assembly general thread.