about summary refs log tree commit diff stats
path: root/mu.arc
Commit message (Collapse)AuthorAgeFilesLines
* 1119 - preparing to start porting chessboardKartik K. Agaram2015-04-211-0/+34
| | | | | | | | | Arc version is 15% faster (8.3s vs 9.9s for print-board test) if I use an intermediate array rather than list. I'm starting to question the whole tagged-value design, and the current tagged-value implementation was treating squares as integers in one place anyway, so its benefits for typing are not great. Might as well create a good baseline for the Arc vs C++ performance test.
* 1063 - variable names for surrounding spaces now workKartik K. Agaram2015-04-141-6/+6
| | | | This was a pain to debug.
* 906 - arc averages 800-900 cycles per secondKartik K. Agaram2015-03-141-0/+6
| | | | No context-switching between routines, either.
* 905 - even factorial trace takes too long to parseKartik K. Agaram2015-03-141-1/+11
| | | | Parsing proceeds at 1-3 lines per *second*.
* 899 - undo 898Kartik K. Agaram2015-03-131-11/+4
|
* 898 - experiment: make memory a vectorKartik K. Agaram2015-03-131-3/+11
| | | | | | | But racket takes too long to allocate a million locations, and anything smaller is currently infeasible. Ok, forget how slow writing to memory is, and focus on the C++ version.
* 897 - trying to save a short chessboard logKartik K. Agaram2015-03-131-6/+21
| | | | But it looks like it will take 1500 minutes at 100% CPU!
* 895Kartik K. Agaram2015-03-121-0/+18
|
* 894Kartik K. Agaram2015-03-121-0/+2
|
* 893 - trying to take traces of chessboard appKartik K. Agaram2015-03-121-0/+1
|
* 892 - warning for bugs like 887Kartik K. Agaram2015-03-121-0/+25
|
* 890Kartik K. Agaram2015-03-121-1/+1
|
* 886Kartik K. Agaram2015-03-111-0/+3
| | | | Another test, but now the tests run 50% slower (8 vs 12 minutes)
* 882Kartik K. Agaram2015-03-101-0/+6
|
* 881Kartik K. Agaram2015-03-101-0/+1
|
* 879 - debug prints shouldn't mess up interactive runsKartik K. Agaram2015-03-091-13/+18
|
* 869Kartik K. Agaram2015-03-071-1/+4
| | | | | | | The repeated need for bugfixes after every new test isn't very confidence-inducing. On the other hand, I'm starting to get into a cadence where fixes have been easy to make. We'll see. Monotonic progress continues.
* 858Kartik K. Agaram2015-03-011-2/+1
| | | | | Don't warn when reusing function names in tests, instead just dump section titles in the trace to aid navigation.
* 857Kartik K. Agaram2015-03-011-23/+23
|
* 856 - stop perturbing traces on new functionsKartik K. Agaram2015-03-011-0/+24
| | | | | Still hacky, but now we can standardize the couple of different workflows we need.
* 852 - page-down stitches expanded lines properlyKartik K. Agaram2015-02-281-0/+1
|
* 850Kartik K. Agaram2015-02-271-1/+1
|
* 842Kartik K. Agaram2015-02-261-94/+91
|
* 841 - no obvious optimization targetsKartik K. Agaram2015-02-261-91/+94
| | | | | | | | | | | | | | | | | Top functions (ms and #calls): run-for-time-slice 209953 184 m 63880 99478 canonize 63568 307282 setm 44378 67281 sizeof 37786 130452 absolutize 36587 370943 addr 34567 146896 typeof 17395 63661 parse-instr 11670 99243 space 10988 561036 typeinfo 10691 563272 absolutize seems like a target.
* 831 - wire up trace browser to interactive replKartik K. Agaram2015-02-241-3/+34
|
* 830Kartik K. Agaram2015-02-241-9/+9
|
* 829 - warn if I forgot to freeze a functionKartik K. Agaram2015-02-241-4/+5
|
* 828 - interactive replKartik K. Agaram2015-02-231-8/+20
| | | | | Still klunky since mu has no notion of a return value. I find myself using $print all the time.
* 826 - augment name bindings when extending a functionKartik K. Agaram2015-02-231-3/+4
| | | | This becomes important as we get closer to a repl for mu.
* 825Kartik K. Agaram2015-02-221-0/+3
|
* 761 - now 30% faster than 758Kartik K. Agaram2015-02-161-72/+72
| | | | 4:26 compared to 6:20 on my laptop.
* 760Kartik K. Agaram2015-02-161-72/+76
| | | | This speeds up the final test but not all together.
* 759 - trace.arc.t 13% fasterKartik K. Agaram2015-02-161-7/+3
| | | | Something wrong with my profiling, though. Numbers aren't adding up.
* 758Kartik K. Agaram2015-02-161-5/+5
|
* 757 - collapse now kinda workingKartik K. Agaram2015-02-161-0/+1
| | | | | Tests are getting slow so quickly that I'm tempted to push forward the rewrite to C.
* 753Kartik K. Agaram2015-02-151-0/+2
|
* 751 - newest trace test now passingKartik K. Agaram2015-02-151-4/+3
|
* 750Kartik K. Agaram2015-02-151-11/+9
|
* 748Kartik K. Agaram2015-02-151-2/+47
|
* 744 - test cursor movement in trace browserKartik K. Agaram2015-02-111-4/+61
| | | | | | Don't prevent run-code from clobbering existing functions, but warn because it makes traces easier to read if the different sections of a test can be distinguished.
* 742 - turns out chessboard test logs were incompleteKartik K. Agaram2015-02-111-10/+19
| | | | | Now that we aren't loading system software for every test we can afford to log the loading of test functions.
* 740Kartik K. Agaram2015-02-101-3/+9
|
* 731Kartik K. Agaram2015-02-101-0/+33
|
* 730 - first test for trace browserKartik K. Agaram2015-02-101-2/+2
| | | | | This was painless compared to the chessboard app. Still need to handle input, though. Fingers crossed..
* 726Kartik K. Agaram2015-02-091-2/+2
|
* 725 - now support cursor movementKartik K. Agaram2015-02-091-0/+8
| | | | This had been on hold for some time; dependencies are crutches.
* 724 - stop using charterm for printing textKartik K. Agaram2015-02-091-19/+12
| | | | | | | We'll keep it around for keyboard input now, for the nice scancode names. Ansi escape codes from http://tldp.org/HOWTO/Bash-Prompt-HOWTO/x361.html
* 722 - split at substring, and assorted helpersKartik K. Agaram2015-02-081-1/+50
|
* 721Kartik K. Agaram2015-02-081-6/+6
|
* 720 - substring matching and searchingKartik K. Agaram2015-02-081-5/+62
|