| Commit message (Collapse) | Author | Age | Files | Lines |
|---|
| |
|
|
| |
showed against its own words: identity (233 samples and the ends of 100,000 and 200,000 messages: 0 bytes differ), taking over (70 kill -9s, all open with every message done), torn tails (867 op boundaries opened, 568 within cuts); "every write a save makes is a self-delimiting, self-verifying extent" false as written (zeros ahead, installs by rename under process-numbered names, directory syncs, 8 MB segments taken): file operations shipped, still no replication layer (the store's own ops cut at its own syncs, heard the ack); the price on tmpfs (204 to 400 us, 903 B a message, 42% zeros the follower could make); what was not tried (a follower started again and heard from its files, pure-serve's checkpoint thread, a real medium and two machines, a segment let go, asynchronous, choosing the owner)
|
| |
|
|
| |
and the numbers: an owner (pure-run --store) with shim.so preloaded ships every change its store makes to its files (ops.h: create, write, truncate, allocate, rename, unlink, mkdir), cut at each fsync or fdatasync, over pure link (shipper.c, follower.c), and its sync returns once the follower has the cut (synchronous: one cut in flight, no log); identity: 100,000 messages (200,081 cuts; 34 compactions, an image written again) and 200,000 (400,149 cuts; 67 compactions, a second log segment), the stores compared byte for byte every 1,999 and 3,001 cuts and at the end: 0 differences, the follower's store opened as owner holding every message; takeover: 70 kill -9s (owner alone, the whole node, the node with cuts of 512 and 64 B records): all open, hold every message the owner said done, none not sent, the entries in order (and equal to the owner's own store each time); prefixes: 867 op boundaries of a 100,000-message stream (568 within cuts, 266 around installs) opened as stores: all pass; the price: 204 to 400 us a message on tmpfs (155 us waiting for acks, two cuts a message), 903 bytes shipped a message, 42% of them zeros written ahead (a compaction's new file and the tree's tail)
|
| |
|
|
| |
(app.c's BORDER: rows whole and finite, the budget and memory begun again, the one state the handler changes written apart and made the state's at a faultless end into a buffer no reader holds, events and changes counted; every end answers all 18,672 lines, the broken-rule build refused at both hand ends): machine instructions a code point, the runtime 40,365, its empty handler 1,595, the hand border with shared 592 and with plain 379, Vega's C given the same payload 240: 68x, 2.5x C (cycles 3.2x), the border 1,595 to about 22; what went: the rows rebuilt into a tree each message (pure_list_read's make_tree), 54 states walked five times an event (40 of them t.X's registers), the last event's nested lists, a bytes state set whole through a text's piece tree; no new fact (N2, today's proof, the language, forms chosen)
|
| |
|
|
| |
LIST_VALUE) and one analysis widened (N2 the handler's changes), not three new as 7ffb114 says: the 9.43x of the event's lists as typed buffers needs no new fact (keyed and indexed are declared, an event's list is one by its place; a typed buffer does every operation an unkeyed, unindexed one gets), only the form chosen and the memory charge's contract; seven fns (range and six), not six
|
| |
|
|
| |
machine instructions a point (event aside) to 564 emitted by hand, 2.5x Vega's C (225; cycles 3.6x), factors 1.46 (dispatch), 3.02 (memory slots and calls), 1.10 (whole numbers), 1.22 (ranges flat), 9.43 (the event's lists typed buffers), 1.23 (Text sharing), the contract left (bounds 1.12, budget and memory 1.39) and the algorithm (1.61); the facts: today's proof licenses exact (each fn's registers, lists and changes: none of the six has lists or changes), then three new (N1 whole numbers, N3 a let list's operations, N4 unchanged after LIST_VALUE) and one analysis widened (N2 the handler's changes); the event itself now the largest layer (1,566 a point: the points decoded into a list, lists cleared and freed, malloc, breaks through the store tree), outside this chain; a premise corrected (the event's lists are FLAT arrays since 021, not trees: the 9.43x is the generic list path); the let lists' factor, broken by a share before, measured as the chain's largest
|
| |
|
|
| |
line, decide and its fns written once against a rung's macros; every rung answers all 18,672 lines as the test and exactly as exact; with LB18 made wrong every rung is refused): machine instructions a code point, the handler: exact 8,785 (registers and plain calls; doubles, fmod, whole and bounds checks twice, the event's lists FLAT PureLists with list_held and grows at each append, Text's lists copied, ranges the stored tree, the budget), integers 7,979, ranges' columns flat 6,561, the event's lists typed buffers 696, Text sharing them 564, no bounds checks 502, no charges 362 (the same algorithm in C), Vega's C 225; the executor 40,189 and stitched 28,089 hold the event's own 1,566 besides (an empty handler: decoding the points into a list, clearing and freeing the event's lists, malloc, the breaks state through the store tree)
|
| |
|
|
| |
executor 14,488 machine instructions a point, stitched 8,144, emitted exact 2,474, with integers 1,587, a flat column 1,083, both 226, C 129 (dispatch 1.78x, memory slots and the call's machinery 3.29x, the numbers' and lists' form 10.9x, bounds and budget 1.75x C); the bar set before (3x C's round) missed by exact (217) and passed by both (19.8); the premises checked: an unchanged list proven (function_changes), a whole-number index not (number slots, INDEX_AT checks it: a fact for the proof to take on), a local's lifetime not tested here; "no layer holds most, each 1.4-1.5x" and the let lists' "broken by arithmetic" were shares inside the executor, not factors; "one machine instruction an instruction is 2.5x C" was wrong (both does 269.5 in 226); what is bundled (memory slots with the call), the flat column's cost moved to changes (unmeasured), and 36% of the app only
|
| |
|
|
| |
could license it (range.c; every way's answers the executor's and the stitched build's, for all 61,285 points): a code point's machine instructions in the executor 14,488, stitched 8,144, exact (registers, a plain call; doubles, fmod, the finite and whole checks, the tree read, the budget) 2,474, integers 1,587, a flat column 1,083, both 226, C 129; a round of halving exact 217, both 19.8, C 11.3: the bar (3x C's round) missed by exact, passed by both
|
| |
|
|
| |
C's with the test before; now Pure's boundaries are read from its view, and a wrong variant is refused with 0 lines right); the executor and the stitched build, rebuilt at adccf58, both answer all 18,672 lines, and the stitched build's numbers stand (40,189 to 28,089 machine instructions a point, 1.43x; cycles 1.32x); 029 says so
|
| |
|
|
| |
built 17.3%, numbers 13.5%, lists read 13.0%, moves 9.0%, calls 7.1%, malloc 4.0%: no layer holds most of the cliff, three generic ones (interpretation, the lists' form, the numbers') 1.4-1.5x each, the cliff their product; range()'s binary search 36% of the machine instructions and 269.5 of 549.9 executed (an algorithm Vega's C does not use); a candidate broken by arithmetic before code (let lists held again as versioned states: 13,000 of 40,236 at most, 1.5x); the bar's arithmetic (one machine instruction an instruction is still 2.5x C: the instructions must fall too); "the general JIT 008 forbids" corrected (008's task A asks for one execution principle from the proof; it forbids the executors' split and scaled-down techniques)
|
| |
|
|
| |
address every 100,003 user instructions from perf_event_open's ring, the answers checked first; where.py names each by addr2line's inlined chain: the step() case its line is in, or the function outside): the app's 40,236 a code point are dispatch 30.6% (the switch and run_calls' loop), lists built 17.3%, numbers 13.5%, lists read 13.0%, moves 9.0%, calls 7.1%, malloc 4.0%; no mechanism near most; range()'s binary search alone 14,489 a point (36%, variant v0r): 11.4 rounds of about 1,270, of which floor division with fmod 186 and a stored read 240
|
| |
|
|
| |
2.19x against 4x, its premise wrong (the source already keeps t.cls[i] in locals: the reads only the proof merges are 16.4 of 549.9), (b) 1.31x against 3x, and the path capped (interpretation is 22 of an instruction's 74 machine instructions: stitching bodies gives at most 1.42x); what was tested (step() on constant instructions, frame and checks as they are) and what was not (an emitter given the proof, the general JIT 008 forbids); where it goes (runtime machinery: a call 1,375, a stored read 331, the lists 22,651 a point) and the 7x the batch lanes give that a fn in a block loses; the 1,774 and 3.2-3.9 ns corrected (the budget's count: 549.9 run, 11.7 ns each) here and in 140; the op counts kept
|
| |
|
|
| |
the numbers: (a) the executor runs 549.9 instructions a code point (the budget's 1,774 counts blocks as written), moves 205.5, constants 76.7 and reads of the same list, version, field and place in an event 16.4, so merging leaves 251.3: 2.19x, the bar 4x (ops-count.patch); (b) the executor's own steps stitched on constant instructions, a function a basic block (stitch.c, native-hook.patch), answers checked: 40,703 to 28,601 machine instructions a point, 15,566 to 11,850 cycles: 1.31x, the bar 3x (cost.c); Vega's C 220 and 82; where it goes: dispatch 22 machine instructions an instruction of 74, an empty handler 1,583 a point, the lists built 22,651 more, the Text record 2,951, decide 12,988 (events.c, variants); a stored list's read 331, a block fn's call 1,375, an arithmetic step 248 in the executor and 35 in batches (unit.c, units)
|
| |
|
|
| |
rejected by outside feedback 008): two principles, each a layer deleted because another already holds what it holds, each with one experiment that can break it before code; 1 the proof is the compiler (the verifier proves stack places and types, loop bounds, each fn's change set, no aliasing between lists, and the three executors throw it away; a C compiler must guess what Pure has proven, so verified aup should be easier to compile to C's speed than C; delete the executor as a path, the batch lanes and the AArch64 backend, emit by stitching the executor's own op bodies; break it on UAX #14: the merging the proof licenses 4x fewer instructions, no dispatch 3x less a instruction), 2 the commit is the replica (a save's extents self-delimiting and self-verifying, the link's heard their durable ack: no replication layer, the follower's files the leader's byte for byte; break it with two hosts on one Pi, cmp at sampled saves, kill -9 and open as owner; checkpoints and compaction to ship or make deterministic; the owner-choosing rule deferred); schema evolution an open question (one shape in the tree, projected on read: unmeasured, and leaves are columns); effects and screens already one boundary, backends
|
| |
|
|
| |
in a loop's body becomes a branch out of it, unmeasured): ③'s first work settles it
|
| |
|
|
| |
Pure, Vega; the messenger, TPC-C, 140's specs, the host, links and store measured), what is verified and what is not apart, each unverified one under 029's order; the document app's table and notes kept below as its time's record; the gaps section points to 029
|
| |
|
|
| |
gaps as architecture and order, each hung on an invariant already there (deterministic events, the chained commit log, ordered once delivery, views as pure functions) with a first proof on the Pi, its pass number and a stop rule: 1 the computing cliff and views' fns (the slow specs all runner 2: fns send whole blocks to the executor; fns spread at load below the language, bytes as columns, fewer instructions, CALL in a view whose callees change nothing), 2 the living store's evolution (a migration a pure fn of the new module, rewritten in batches while the app answers; 030's items-wide at 1M under 5 ms unanswered, the store the same bytes), 3 distribution (committed bursts replayed by a follower, an epoch fenced owner chosen by two of three, topology as a directory app's views; two hosts on one Pi first), 4 Vega past the grid (a display list as a view type, layout in aup with text measured by the host, IME and clipboard as input events, native windows last), alongside 5 effects whose results come back as events, one entry, a C ABI v0; the app contract gathered as draft v0, the fixed limits deferred
|
| |
|
|
| |
19,284), how, what checks it (586 of 600 loads at a freed address, answers right; ASan clean), what a mutant comparing addresses showed (nothing: requests hold their module and freeing waits a pass), the push group woken again, the bindings left (24 B an input a module); the bank's queries within 0.4% (s50, s51); CHECKPOINT's remaining list without it
|
| |
|
|
| |
module a Loaded with holds (each snapshot of it, each request read with it); one replaced is retired and freed on the apps thread once no app runs it, every connection thread has begun a pass since (woken for it), and nothing holds it; who keeps a module to ask later whether it is the same asks its serial (core: PureModule.serial, a reader's module_serial, pure_snapshot_module, pure_snapshot_release says it freed; serve: routes, groups, links, bindings, queries, the published snapshot), as an address may come back; a link binds under its receiver's lock, a link naming an input under binding_lock; 030's memory.py: 4,000 changes 0 bytes a change (was 19,284), 4,100 freed; stress.py: readers, view workers, push streams and a body across changes, 586 of 600 loads at a freed module's address, every answer right, ASan 1,000 changes clean; a push group that looked while its app resumed missed the new snapshot (published before the app took requests) and waited for its next save: woken again once it takes them (server_view_workers had failed once under load)
|
| |
|
|
| |
183x), not two runs' numbers
|
| |
|
|
| |
rebuild left: the core's three gaps, the one-shot timer and lending list values as decision points, the work unit, the fixed limits, no frozen spec; byte and text work 61-186x C; aup 1.47-2.04x Python's lines; three machines last seen together 10-05; one entry point, packages, hosts beyond Linux, a C ABI; the evidence table still the document app's); the direction row's frictions and list values
|
| |
|
|
| |
486 runs (41 modules) the same with gcc and clang on the Pi (Pure 0896b38, aup 1efe1c7; aup's modules the same bytes under clang); the Mac and WSL2 not reached (ssh timed out, no route), the bundle and r140.sh ready for them
|
| |
|
|
| |
-Wbitwise-instead-of-logical stopped a clang build since 027's 4f-2, so a Mac's: the same frees either way, the second malloc not tried when the first fails); Pure builds and its unit tests pass with clang
|
| |
|
|
| |
recorded run by run for machines to compare (aup's 018 checks as 033's record.py does, its 027 stages', 028's bytes state, 140's frictions; Pure's module_specs, the nine specs on the three paths, Pure's specs copied under the temporary directory so no machine's paths are in an input): 486 runs of 41 modules on the Pi
|
| |
|
|
| |
about 10% of its time, JSON tree 6%, HTTP 2%, inflate 0.06%, the rest none; the cost a tree built item by item, 64 ns an item and 250 ns a copy): lend and detach (CoW) noted as the decision point (a branch on every event list's change for 10/6/2/0/0/0), a compile-time move ruled out (it changes LIST_VALUE's contract), the arrays copied whole built (UAX #14 -6.8%, its share to about 3%); 041's five works within 0.03% (b49, b50), the bank's queries within 0.3% (s49, s50), noise between builds larger
|
| |
|
|
| |
opcode, no change to what a module means): nest() made every list value item by item into a new tree (64 ns an item, 250 ns a copy fixed); an event's list in FLAT arrays, not keyed, its cells numbers, bools and integers, is now copied column by column (pure_list_copied: 600 items 38,268 to 985 ns, 60 items 3,926 to 324); UAX #14's nine lists a line 6,917 to 6,445 ns a code point (-6.8%, the same instructions); JSON tree's, HTTP's and inflate's (trees: texts or lists in their records, or few) as they were; copied_lists in test-module (random lists against item by item, then each changed apart; trees, keys and texts not copied so)
|
| |
|
|
| |
change, no new opcode): lists declared together, targets together (straight when no value can read what a target changes), a list given a list value whole, abs, tick.period, hex where numbers go; lines to Python 1.47-2.04x (were 1.47-2.15x): UAX #14 176 to 169, layout 116 to 110, HTTP 148 to 146 and its last workaround gone; instructions of the four whose modules changed within 0.2%, inflate's and UAX #14's modules the same; the six specs' checks pass on the three paths, module_specs nine; RTO's checks no longer rewrite G (tick.period); the suite table's rows 5, 8 and 9
|
| |
|
|
| |
Python count holds the four lines of its left-to-right total (116 against 50 without them)
|
| |
|
|
| |
examples/layout.aup: a subset of CSS Flexbox 1 written down first, one line, row or column, padding, gap, six justify-content, four align-items, grow and shrink with 9.7's loop, cross sizes stretched, nothing sized by content): 2,000 random trees (29,388 nodes; 7,659 items frozen at their padding by the loop) are Taffy 0.15's within 0.01 (layout_taffy, built by cargo; the worst 8.3e-5) and layout.py's bit for bit on the three paths, the paths the same; the single-pass subset first tried refused 66% of trees, so the loop came in; Taffy given min sizes equal to the padding (given 0 it floors at 0 while resolving lengths, where 9.7 floors the content box), the reference adding from the left (sum compensates since Python 3.12); module_specs in make check: 100 trees (tests/specs/layout); bars: 116 lines against Python's 54, 4,754 ns a node, Taffy's 644 (7.4x), pure Python 5,396 (Pure 1.14x faster), 1,494 instructions a node (9.7's loop 44%); four frictions; aup's rows read by position no longer copy (99,067 ns a node before); the suite table's row 8; CHECKPOINT's direction row
|
| |
|
|
| |
(aup's examples/rto.aup, the timer a deadline in the state seen at each periodic tick): 20 scenarios (10 seeds, ticks every 10 and 100 ms; a closed loop: first sendings lost one in five, the receiver acknowledging cumulatively) and their 241,612 events right against rto.py, the model of the same rules (no known implementation takes given events), on the three paths, the paths the same; module_specs in make check: one scenario at both periods (tests/specs/rto, 2,297 events); bars: one workaround (the periodic tick: 2,795.8 ticks a timeout at 10 ms, 281.6 at 100, seen 4.27 and 49.23 ms late on average; ticks 93% and 57% of the counted work), 56 lines against Python's 30, three frictions (an input's period not readable, no abs, send a keyword); the one-shot timer's shape as a decision point, one spec needing it: noted, not made; the suite table's row 10
|
| |
|
|
| |
record copies its nine lets' lists a line: with building them, 260 of 1,774 instructions a code point, after inflate's code()), and a keyed list's floor seek (the last item keyed x or below) first wanted (UAX #14's ranges, a binary search by position, 360 instructions a code point), one spec behind it
|
| |
|
|
| |
opcode (aup's examples/linebreak.aup, the properties as data: 2,787 ranges from the UCD 16.0 Vega's text is frozen at, built by Vega's text/unicode.py, loaded once): every line of LineBreakTest.txt (16,672) and 2,000 random lines (their boundaries as linebreak.py, the Python reference, which passes the test whole, works them out) right on the three paths, the paths' outputs the same; module_specs in make check: the ranges and the lines kept in tests/specs/linebreak (derived from the UCD, the Unicode License v3's notice), run by pure-run --views; bars: written, no workaround (three frictions: hex literals always i64, number no let's name, a let's list one a line), 176 lines against Python's 99, 6,920 ns a code point against Vega's C 34.7 (199x), pure Python 2,523 (Pure 2.7x slower), 1,774 instructions a code point (setup 260, the ranges' binary search 360, decide about 1,150 a boundary); the decisions before 8 (box layout: Taffy can be fetched by cargo, a subset to choose first; as a handler) and 10 (a one-shot timer is a core primitive one spec needs: noted, not made); the suite table's row 5; CHECKPOINT's direction row
|
| |
|
|
| |
holds a table (research 140's UAX #14 holds 2,787 ranges; printed at each of 16,672 commands, its first run's captured output used the Pi's memory up and it was reset)
|
| |
|
|
| |
the three paths as before (zlib's 640 streams and 240 damaged, dnspython's 400 messages and 300 damaged, h11's 200 streams on every cut and 200 damaged); lines inflate 132 to 121, DNS 115 to 106, HTTP 159 to 148 (Python's 1.5-1.9x, was 1.6-2.0x); time inflate 1,152 to 1,139 ns a byte (102x C), DNS 244 to 221 (61x, instructions 63.6 to 49.9 a byte: == in its compression), HTTP 278 to 270 (83x, with the free empty join); left: DNS's view calls no block fn, HTTP's let list takes no list value; module_specs, the checks and the timing tools read the bytes states (inflated, pending); the suite table's rows for 1, 3 and 9
|
| |
|
|
| |
change), so a bytes state joined while empty (HTTP's pending + packet.data) no longer copies the packet: HTTP a packet at a time 288 to 270 ns a byte at the same instructions, as the old app (270.0); counts: b48 against b47 knapsack without batches -0.42%, join -0.16%, BFS with batches -0.12%, the rest within 0.03%; the bank's queries (s47 and s48 in turns, six each) in_credit -0.34%, field -0.34%, found -0.22%, per_branch -0.20%, large -0.18%, ranks -0.13%, none the same, the seven answers the same bytes
|
| |
|
|
| |
patch kept): a block fn whose only return is its last line, with no loop or called where the stack is empty, written out where it is called, its parameters slots of its own (one set a fn and kinds of arguments in a handler); the three specs' checks passed, but inflate 191.2 instructions a byte as before and HTTP 61.4 to 62.1 (+1.1%, 277.7 to 274.9 ns): the hot fns (inflate's decode, HTTP's crlf) return early inside a loop, a fn with a loop cannot be written out where values are on the stack (Pure opens no loop above values), and storing a short body's parameters costs what CALL and PICK did; inflate's 23% was the expression fn bit's (a loop, a let and conversions gone with the call); what would make it pay is Pure's (loops above values, or a cheaper frame), one spec behind it; the decision point's list of what was made after it
|
| |
|
|
| |
flagged PURE_TEXT_BYTES in state_flags (so serve's same_states compares it for queries and replaces), its TEXT_LOAD giving bytes and its TEXT_SET taking them, a cursor's ops on it refused (anchor in bytes) and an anchor in it refused at load; one store file text-NAME, the values record marking each bytes state (kind 7, its name, 8 zero bytes), a text state's file that is no UTF-8 its initial value when marked and store_error when not (as before); a replace carrying bytes to a text only when its draft is UTF-8, else opening the store's draft anew; pure-run shows a bytes state's hex; unit test bytes_states (seven starts from a store: 0xff and 0x00 kept, bytes to text to bytes to text with no UTF-8 and with it; three replaces; four refused); counts: b47 against b46 BFS with batches +0.12%, the rest within 0.001%; the bank's queries (s45 and s47 in turns, six each, bytes compared among them) in_credit -0.31%, per_branch -0.18%, ranks -0.11%, the rest within 0.01%, the seven answers the same bytes; Pure, Vega and aup checks green; 008, 028
|
| |
|
|
| |
it): state type 6 a module's bytes, held as a text is (one draft, app->texts) with state_bytes marking it, no new opcode (TEXT_LOAD gives and TEXT_SET takes bytes for it), no anchor in it; one store file text-NAME, the values record (PVS1) marking each bytes state (kind 7, its name) so a text file that is no UTF-8 is a move when the record says bytes (the initial value) and corruption when it does not (store_error, as before); text and bytes carried over each other when exact (text to bytes always, bytes to text when UTF-8), on a replace and on a start, A to B to A raising no stale value; aup's state x: bytes; the checks planned, two names for the file and a new runtime kind turned down, list sharing and emptying left for when their cost shows
|
| |
|
|
| |
TEXT_COMPARE's argument takes PURE_COMPARE_BYTES (bit 8) beside its relation, its two then bytes (UTF-8 asked of neither), as CONCAT's argument says bytes; verified as A_ORDER (a relation, and the bit), the executor relating by the argument's relation alone; the JIT, the tier and the batch executor leave TEXT_COMPARE to the executor as before; unit test bytes_compared: eight relations of a constant's bytes, none, 0xff and 'A' (0xff after 'A', a longer after its first byte), four refused (the bit on texts, no bit on bytes, a text with bytes, a relation past NE); counts: b46 against b45 the same on all five works, with and without batches; 008
|
| |
|
|
| |
with no new opcode (aup's examples/http.aup: each packet an event, its bytes added to what the last left, every request they complete read, h11's grammar): 200 streams (132,623 bytes, 578 requests: methods, targets, HTTP/1.1 and a last 1.0, mixed-case names, spaces around values, Content-Length and chunked bodies with extensions and trailers, some ending in part of a request) fed whole, at random boundaries and a byte at a time (the 50 under 400 bytes), the requests read h11's on every cut and the bytes left those after them, the three paths' outputs the same bytes; 200 damaged streams fed whole as http_requests.py (the Python reference) reads them, 35 refused by both and 165 the same; module_specs in make check: 80 streams and their reading (tests/specs/http, no h11 at check time) whole on the three paths and cut 88 ways; bars: written, three workarounds (the bytes left in a list of one: bytes are no state; names and codings compared as text: bytes have no ==; field lines returned in a record: a fn returns no bare list), 159 lines against Python's 79, 278 ns a byte against the same algorithm in C 3.27 (85x), pure Python 67 (its bytes methods are C: Pure 4.1x slower), h11 284; 140's decision point after 027: bytes alone as a state (inflate, HTTP) and bytes' == (DNS, HTTP) now each met by two written specs, so the two-spec rule lets them into the core (their design not begun); a view calling a block fn met by one (DNS); aup's three its compiler's; CHECKPOINT's direction row
|
| |
|
|
| |
loops it compiled not looked at
|
| |
|
|
| |
examples/dns.aup: a message read by a handler into its header and records, names uncompressed in their wire form, those in NS, CNAME, PTR, MX, SOA and SRV data too, each pointer pointing before the last; written back by the view wire, the owners compressed): 400 messages dnspython made (148,582 bytes, 3,466 records: A, AAAA, NS, CNAME, PTR, MX, SOA, TXT, SRV, EDNS, shared suffixes, 63-byte labels, bytes above 0x7f, 245-byte names) read on the three paths as dnspython reads them (each section's records as a set) and written back to bytes dnspython reads as the same message, the paths' outputs the same bytes; 300 damaged messages as dns_message.py (the Python reference) reads them, 185 refused by both and 115 the same records and bytes; module_specs in make check: 200 of those messages (tests/specs/dns, no dnspython at check time) read on the three paths, written back and read again to the same state and bytes, 23 cut ones failing; dnspython shuffles a set's records unless want_shuffle is off, and merges a set's differing TTLs to the least; bars: written, two workarounds both the core's (bytes have no ==: TEXT_COMPARE takes text alone; a view calls no block fn: CALL is in handlers and functions alone), 115 lines against Python's 72, 244 ns a message byte read and written against the same algorithm in C 3.65 (67x), glibc's resolver reading alone 4.4, pure Python 131 (Pure 1.9x slower), 218 compiled; 63.6 instructions a byte, reading 42%, writing 58%; CHECKPOINT's direction row
|
| |
|
|
| |
let's list only in its handler's own block are aup's (Pure's function signatures already take lists, type 4, moved by number and copied before a change; a block's list emptied by LIST_REMOVE or kept apart), a bytes value as a state alone is Pure's state kinds (a record alone aup could spread into states as it spreads record fields; bytes have no kind: text asks UTF-8), small block fns' inlining aup's; the two-spec rule gates the core alone
|
| |
|
|
| |
examples/inflate.aup, codes read a bit at a time from canonical counts as zlib's puff.c): 640 streams zlib made (6,488,480 bytes: random bytes, words, runs and repeats 32,768 back; levels 0, 1, 6, 9 and zlib's five strategies, so stored, fixed and dynamic blocks) inflated to the original's bytes on the three paths, the paths' outputs the same bytes; 240 damaged streams (cut, a bit flipped) read as inflate.py (the Python reference) reads them, 165 refused by both and 75 the same bytes; module_specs in make check: 144 streams on the three paths, 21 cut ones failing and leaving the last bytes; bars: written (no new opcode), three workarounds (lists passed in records: a fn takes no list; the output a list of one: a bytes value is no state; a block's tables built in a fn: a let's list lives in the handler's own block), 132 lines against Python's 81, 1,152 ns an output byte against the same algorithm in C 10.7 (108x) and zlib 3.5 (329x), pure Python 1,214; 191 instructions a byte (a block fn called for each bit was 246 and 1,553 ns: small block fns' inlining is 2c's leftover); the four timing tools' message calls fixed for 4b's API (no keys); CHECKPOINT's direction row
|
| |
|
|
| |
directory: 10-13's -o /dev/null made a compiled probe say Permission denied): three compile as written, the other seven's diagnostics now mostly naming the form, so the same needs written in 027's forms (probes-027/): a byte an i64, bits on i64s, a record holding itself through a list, a recursive fn in a handler with its result type, a character by for and char, a variant in a list's item all compile; three gaps left, each a change of the core: a view calling a recursive fn or a fn with a block, a record or a variant as a state by itself, a character at a cursor; the suite table's column after 027 (2 and 4 written, 6 and 7 work, the first needs of 1, 3, 5 and 9 compiling, 8's layout as a view blocked by the first gap)
|
| |
|
|
| |
and of counts; i64 the exact integers, bits and money), / a number's division and // the floored quotient (FLOORDIV, 95 opcodes now 96), an i64 sum exact, money i64 minor units shown by .Nd, a replace and stored values moved when exact; the bar: TPC-C's money and rates integers in Pure and SQLite alike, every row of every table exactly the same after 20,000 transactions; core 22,766 lines to 22,791, serve.c 5,058 as it was, aup's compile.c 5,720 to 5,726; counts: b45 against stage 4's end (b42) within 0.04% on all five works; the bank's queries (s42 and s45 in turns, six each) per_branch -0.58%, in_credit +0.56%, ranks +0.20%, the rest within 0.01%, the seven answers the same bytes; TPC-C in RAM (W=1, 64 terminals, 20 s, in turns, three each): s42 on its own app (money numbers) 46,092 transactions a second, s42 on stage 5's (money i64) 47,303, s45 on it 47,287: the runtime the same (-0.03%), the app 2.6% faster with i64 money (the server's CPU a transaction 42.3 to 39.6-39.8 us, not taken apart), new-order p99 3,584 to 3,456 us; 027's status: the five stages made; CHECKPOINT's direction row through stage 5
|
| |
|
|
| |
(pure_cell_as) for a replace while running, the stored values at a start and a changed record's list moved (in memory and from its image): a whole number in [-2^63, 2^63) to an i64 (-0 as 0), an i64 within 2^53 to a number, 0 and 1 to and from a bool, all else the new default; aup's test number_kinds_carried runs it on states and a stored list's field, on a replace and on starts from a store; 008's replace and start say so for i64, and the values and a list image's i64 cells (two's complement, little-endian 8 bytes); no code changed, so no counts; 027
|
| |
|
|
| |
takes digits (bits 20-24 as FIXED's), an integer's last N digits written after a point (1234 at two is 12.34, -5 -0.05, 0 0.00; width and zeros after the sign), verification allowing digits on FIXED and DECIMAL alone; unit test: ten rows of formats_and_held (-2^63 at two, 25 digits, a number's whole 1234, its 12.5 a limit); TPC-C's money i64 cents and its rates ten-thousandths (tables, messages, load, the transactions), SQLite's INTEGER columns and int64 bindings alike, the comparison now exact (no 1e-9 tolerance): after 20,000 transactions every row of every table the same (grouped SQLite too) and 3.3.2 holding on both; counts: b45 against b44 BFS without batches +0.09%, the rest within 0.001%; the bank's queries (s44 and s45 in turns, six each) in_credit -0.59%, per_branch -0.34%, ranks -0.20%, the rest within 0.01%, the seven answers the same bytes; 008 (NUM_TEXT's digits, a message row's i64 cells), 027 and 092
|
| |
|
|
| |
the result checked against the range (overflow past it), so a sum past the range midway and back in it at the end is the answer and an i64 sum's order is none of its business; the batch executor's folds are numbers' lanes alone, summed in the items' order, as before; 008's FOLD says so; counts: b44 against b43 within 0.04% (BFS without batches -0.03%); the bank's queries (s43 and s44 in turns, six each) per_branch -1.32% (all six below s43's lowest: a larger Fold lays the number sum's loop otherwise), in_credit -0.72%, ranks -0.26%, the rest within 0.01%, the seven queries' answers the same bytes on both; 008 and 027
|
| |
|
|
| |
default of literals and of counts, positions, lengths and u32 fields, whole to 2^53 exactly; i64 the exact integers past it, bits and money, never mixed, moved by number() and i64() when exact), / a number's division, // the floored quotient, i64 folds exact (only the result in range, so in any order), number folds in order unless the batch executor's 2^53 test says any order is exact, money as i64 minor units shown by .Nd, a replace carrying number, bool and i64 when exact; 5a: opcode FLOORDIV (95, 96 now): a number's floored quotient as Python's float //, exactly (a less its floored remainder, divided, to the nearest whole; a zero signed as a / b; by zero divide), a new opcode so the JIT, the tier and the batch executor leave it to the executor; unit test floor_division (7 // 2, -7 // 2, 7 // -2, 1 // 0.1 = 9, 0 // -3 = -0, 1e300 // 7, by zero); counts: b43 against b42 within 0.001% but BFS without batches -0.09%; the bank's queries (s42 and s43 in turns, six each) -0.20% to 0.00%; 008 and 027
|
