Search:

Never mind, found the cause, I used the wrong value to compare against, used _DST.mdig instead of _NUM->mdig

Attempted to fix a rounding error but the fix seems to be ignored, I got other stuff to do now so hoping someone will spot the cause by the time I come back, here's the code & output:
GitHub -...

Figured out the fix:


exp -= ((__DST.upto - final.bit) - __SRC.bits);

Recheck the docs on scanf, your using the output wrong if memory serves me correctly

Managed to fix the mantissa in division but still having problems with the exponent in that field, hoping someone here can see where I'm going wrong.
GitHub - awsdert/alu at...

Managed to fix it, this was all I needed to add after all: GitHub - awsdert/alu at 5f4c54b0303fa98c63a776212cd65a9241371707


dmov = _DEXP.from - d1st.bit;
smov = _SEXP.from - s1st.bit;
mov =...

Appears the video was not as informative as it needed to be, ended up removing something that I thought was needed because of the video and got more correct results, still fails at a point and need...

Forgot to use to separate bias's when I switched to assuming the temporary floats where double the size of the initial variables, fixed that and then analysed more of the flow and found that it's...

Uncovered & fixed a bug with alup_mov_flt2flt, turned out the mantissa transfer was overriding 1 bit too many and corrupted the exponent as a result, still can't get the final exponent correct...

Found a more informative video for FPN multiplication, made some changes to the function I'm using and I think I now how the mantissa correct (at least it looks correct from the raw binary), the...

Precisely why I keep saying it needs to be fixed

I thought it was C# given how the file extension is *.cs, I figured that was where that abomination of "double." came from

Okay 1st you didn't fix the condition brackets so you still need to fix that, also:


else (rdoMale.Checked == true && rdo30Over.Checked == true)

How does that even compile? Did you mean this?...

First I don't see how that compiles because from what I see you have an unclosed if condition at the start, also that condition is hard to read as is, start by using this format on it:


if
(
...

Managed to fix it somehow, not quite sure how though so if someone feels like analysing the code here's the commit holding the code that passes the tests, the fix however leads me to believe I may...

Well didn't expect much, not that it matters now but sorry I went silent for a while, combination of work and being to into gaming, reading and youtube lead me to leave this to the wayside for a...

Just uploaded the changes now, got called to start work early (2hrs overtime before my normal hours) so I'll just leave you to it if you're still interested in trying, here's the vid that made me...

I think I was still misunderstanding something in the previous iteration, gotta shoot off to work so I'll just dump the code and output.
Output:


make check.run
...
Running suite(s): ALU...

OK I think I've nearly got it right on addition & subtraction, somehow missed the bias in the exponent though, anyone able to see where I lost the bias?


typedef int_t (*alup__addsub_int2int)(...

Well I figured out why the other tests were passing, I had fed the same value into both sides of the operation, fixing that resulted in the expected failures, still haven't figure out where I'm going...

Got back to it 2day instead, anyways I tried a couple of things then noticed I needed some variable from the exponent matching stage (forgot what it was and for what purpose by the time I combined...

I'm guessing noone else had any ideas either... or just wasn't interested, I'm betting on the lata, eh I'll just come back to it 2mw

I went back to the original bug where floating subtraction did not yield negative numbers in the appropriate cases, I tried applying my new knowledge of how negative numbers are represented to...

As it happens my previous code where the negative was temporarily converted to positive prior to analysis and movement was actually correct, seems the bug I encountered that led me to believe...

Currently trying to fix an integer to float function, does positive numbers just fine but chokes on negatives, seems to be just one off so was looking for the potential fix in the final part of the...