2019-03-04 |
Luke Kenneth Casso... | remove chain dependence, calculate ediffs in parallel... |
tree | commitdiff |
2019-03-04 |
Luke Kenneth Casso... | comment out unneeded code for now |
tree | commitdiff |
2019-03-04 |
Luke Kenneth Casso... | convert to only use one multi-shifter |
tree | commitdiff |
2019-03-04 |
Luke Kenneth Casso... | rename stickybit variable |
tree | commitdiff |
2019-03-03 |
Luke Kenneth Casso... | unit test for multi-bit shift right with merge (sticky... |
tree | commitdiff |
2019-03-03 |
Luke Kenneth Casso... | cleanup |
tree | commitdiff |
2019-03-03 |
Luke Kenneth Casso... | small optimisation, move subtraction of -126 from expon... |
tree | commitdiff |
2019-03-03 |
Luke Kenneth Casso... | add 3 extra unit tests |
tree | commitdiff |
2019-03-03 |
Luke Kenneth Casso... | limit count leading zeros to stop exponent shift-amount... |
tree | commitdiff |
2019-03-03 |
Luke Kenneth Casso... | fix shift class syntax errors (untested) |
tree | commitdiff |
2019-03-03 |
Luke Kenneth Casso... | use priority encoder for normalisation in single cycle... |
tree | commitdiff |
2019-03-03 |
Luke Kenneth Casso... | add in FPNumShiftMultiRight class |
tree | commitdiff |
2019-03-02 |
Luke Kenneth Casso... | use bool() function instead of reduce(or_) |
tree | commitdiff |
2019-03-02 |
Luke Kenneth Casso... | got single-cycle align working again (accidental combin... |
tree | commitdiff |
2019-03-02 |
Luke Kenneth Casso... | turn FPOp into module |
tree | commitdiff |
2019-03-02 |
Luke Kenneth Casso... | move put_z to PutZ class |
tree | commitdiff |
2019-03-02 |
Luke Kenneth Casso... | reorg pack setup |
tree | commitdiff |
2019-03-02 |
Luke Kenneth Casso... | reorg corrections setup |
tree | commitdiff |
2019-03-02 |
Luke Kenneth Casso... | remove unneeded function call |
tree | commitdiff |
2019-03-02 |
Luke Kenneth Casso... | remove global z as output from specialcases, use sc... |
tree | commitdiff |
2019-03-02 |
Luke Kenneth Casso... | remove unneeded variable, use module overflow to get... |
tree | commitdiff |
2019-03-02 |
Luke Kenneth Casso... | managed to make round signal an output from normalisati... |
tree | commitdiff |
2019-03-02 |
Luke Kenneth Casso... | tidyup, remove unneeded intermediate |
tree | commitdiff |
2019-03-02 |
Luke Kenneth Casso... | tidyup, remove unneeded intermediate |
tree | commitdiff |
2019-03-02 |
Luke Kenneth Casso... | reorg FPRound move setup function |
tree | commitdiff |
2019-03-02 |
Luke Kenneth Casso... | add comment about add0+add1 stages |
tree | commitdiff |
2019-03-02 |
Luke Kenneth Casso... | add1 module setup reorg |
tree | commitdiff |
2019-03-02 |
Luke Kenneth Casso... | remove temporary external z, use add0 output, connect... |
tree | commitdiff |
2019-03-02 |
Luke Kenneth Casso... | reorg: move add0 setup function |
tree | commitdiff |
2019-03-02 |
Luke Kenneth Casso... | use correct local output from pack chain |
tree | commitdiff |
2019-03-02 |
Luke Kenneth Casso... | reorganise normalisation init: move setup function... |
tree | commitdiff |
2019-03-02 |
Luke Kenneth Casso... | add module links (gives useful names) |
tree | commitdiff |
2019-03-02 |
Luke Kenneth Casso... | complicated way to create a loop inside the normalisati... |
tree | commitdiff |
2019-03-01 |
Luke Kenneth Casso... | remove variable overflow |
tree | commitdiff |
2019-03-01 |
Luke Kenneth Casso... | connect corrections to pack without global z |
tree | commitdiff |
2019-03-01 |
Luke Kenneth Casso... | split roundz from norm z |
tree | commitdiff |
2019-03-01 |
Luke Kenneth Casso... | pass add0 z through to add1 independently |
tree | commitdiff |
2019-03-01 |
Luke Kenneth Casso... | add new temporary z for result chain |
tree | commitdiff |
2019-03-01 |
Luke Kenneth Casso... | connect add1 to norm1 overflow without global store |
tree | commitdiff |
2019-03-01 |
Luke Kenneth Casso... | merge normalise_1 and normalise_2 stages |
tree | commitdiff |
2019-03-01 |
Luke Kenneth Casso... | commennt use of intermediates |
tree | commitdiff |
2019-03-01 |
Luke Kenneth Casso... | store zero-extended a and b in temp signals |
tree | commitdiff |
2019-03-01 |
Luke Kenneth Casso... | store tests in temp signals |
tree | commitdiff |
2019-03-01 |
Luke Kenneth Casso... | experimenting with chaining Overflow module |
tree | commitdiff |
2019-02-28 |
Luke Kenneth Casso... | use output from align as input to add0 |
tree | commitdiff |
2019-02-28 |
Luke Kenneth Casso... | remove commented-out code |
tree | commitdiff |
2019-02-28 |
Luke Kenneth Casso... | use GetOpMod for b |
tree | commitdiff |
2019-02-28 |
Luke Kenneth Casso... | create and use GetOp module |
tree | commitdiff |
2019-02-28 |
Luke Kenneth Casso... | move fpnum_b to class FPGetB |
tree | commitdiff |
2019-02-28 |
Luke Kenneth Casso... | narrowing down rounding error to use of Norm1 module |
tree | commitdiff |
2019-02-28 |
Luke Kenneth Casso... | separate denormalisation module and use it |
tree | commitdiff |
2019-02-28 |
Luke Kenneth Casso... | use denorm exponent signal |
tree | commitdiff |
2019-02-28 |
Luke Kenneth Casso... | sorting out unit tests, comply with IEEE754 on RISCV |
tree | commitdiff |
2019-02-28 |
Luke Kenneth Casso... | recompiled sfpy, testing FP16 again |
tree | commitdiff |
2019-02-27 |
Luke Kenneth Casso... | whoops, overflow not right, reverting |
tree | commitdiff |
2019-02-27 |
Luke Kenneth Casso... | add failed test |
tree | commitdiff |
2019-02-27 |
Luke Kenneth Casso... | assign tests to signals |
tree | commitdiff |
2019-02-27 |
Luke Kenneth Casso... | create single and multi shift cycle, single doesnt... |
tree | commitdiff |
2019-02-27 |
Luke Kenneth Casso... | more chains between inputs and outputs |
tree | commitdiff |
2019-02-27 |
Luke Kenneth Casso... | move of = Overflow() out of FPADD, use chain |
tree | commitdiff |
2019-02-27 |
Luke Kenneth Casso... | remove tot from FPADD, use chain |
tree | commitdiff |
2019-02-27 |
Luke Kenneth Casso... | connect add0 to add1 |
tree | commitdiff |
2019-02-27 |
Luke Kenneth Casso... | create add1 stage module and use it |
tree | commitdiff |
2019-02-27 |
Luke Kenneth Casso... | try some more chaining of inputs to outputs |
tree | commitdiff |
2019-02-27 |
Luke Kenneth Casso... | pass output from normalise_2 to input of roundz |
tree | commitdiff |
2019-02-27 |
Luke Kenneth Casso... | create add0 stage module and use it |
tree | commitdiff |
2019-02-27 |
Luke Kenneth Casso... | name modules correctly |
tree | commitdiff |
2019-02-27 |
Luke Kenneth Casso... | whoops norm2 using norm1 mod |
tree | commitdiff |
2019-02-27 |
Luke Kenneth Casso... | create normalise_2 module and use it |
tree | commitdiff |
2019-02-27 |
Luke Kenneth Casso... | put exponent > 126 logic in FPNumBase, use it in norm... |
tree | commitdiff |
2019-02-27 |
Luke Kenneth Casso... | split out first stage normalisation to module and use it |
tree | commitdiff |
2019-02-27 |
Luke Kenneth Casso... | reduce random case test numbers as well |
tree | commitdiff |
2019-02-27 |
Luke Kenneth Casso... | reduce number of unit test runs to get quicker more... |
tree | commitdiff |
2019-02-27 |
Luke Kenneth Casso... | split special cases into separate module and use it |
tree | commitdiff |
2019-02-27 |
Luke Kenneth Casso... | create and use FPPack module |
tree | commitdiff |
2019-02-27 |
Luke Kenneth Casso... | create and use corrections submodule |
tree | commitdiff |
2019-02-27 |
Luke Kenneth Casso... | rounding done in module |
tree | commitdiff |
2019-02-27 |
Luke Kenneth Casso... | get roundz state to put answer in explicit output,... |
tree | commitdiff |
2019-02-27 |
Luke Kenneth Casso... | clean up unit_test_single get_case based on how dual_ad... |
tree | commitdiff |
2019-02-27 |
Luke Kenneth Casso... | clear STB immediately after setting, stops add1 repeati... |
tree | commitdiff |
2019-02-26 |
Luke Kenneth Casso... | moving internal strobe test forward is ok |
tree | commitdiff |
2019-02-25 |
Luke Kenneth Casso... | invert stb/ack between add1 and add2 |
tree | commitdiff |
2019-02-24 |
Luke Kenneth Casso... | experimenting with dual add |
tree | commitdiff |
2019-02-23 |
Luke Kenneth Casso... | trying different testing for 2nd round |
tree | commitdiff |
2019-02-23 |
Luke Kenneth Casso... | use function to get chain of v/ack/stb |
tree | commitdiff |
2019-02-23 |
Luke Kenneth Casso... | yippee got dual add chained together |
tree | commitdiff |
2019-02-23 |
Luke Kenneth Casso... | whoops revert decode inside module FPNumIn, causing... |
tree | commitdiff |
2019-02-23 |
Luke Kenneth Casso... | add dual unit test |
tree | commitdiff |
2019-02-23 |
Luke Kenneth Casso... | move unit test order |
tree | commitdiff |
2019-02-23 |
Luke Kenneth Casso... | remove unneeded class declaration |
tree | commitdiff |
2019-02-23 |
Luke Kenneth Casso... | add a dual-chained add experiment |
tree | commitdiff |
2019-02-23 |
Luke Kenneth Casso... | store logic-test conditions in intermediates |
tree | commitdiff |
2019-02-21 |
Luke Kenneth Casso... | isolate inputs and outputs in FPGetA class |
tree | commitdiff |
2019-02-21 |
Luke Kenneth Casso... | FPADD need no longer be derived from FPBase |
tree | commitdiff |
2019-02-21 |
Luke Kenneth Casso... | remove explicit code-adding of states, use for-loop... |
tree | commitdiff |
2019-02-21 |
Luke Kenneth Casso... | move putz to separate class |
tree | commitdiff |
2019-02-21 |
Luke Kenneth Casso... | move pack to separate class |
tree | commitdiff |
2019-02-21 |
Luke Kenneth Casso... | move corrections to separate class |
tree | commitdiff |
2019-02-21 |
Luke Kenneth Casso... | move rounding to separate class |
tree | commitdiff |
2019-02-21 |
Luke Kenneth Casso... | move normalisation stages to separate classes |
tree | commitdiff |
next |