Counter ASIC 3.0 PC 1 AMD: G2 takes the class token from the pack (v4 for the generator-4 candidate), the G2 kit rebuilt from the merged tree, digests unchanged on the Mac, the wrong-token refusal checked
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
This commit is contained in:
igneum-labs
parent
d022e5ef76
commit
be35310f9e
2 changed files with 10 additions and 7 deletions
|
|
@ -38,7 +38,7 @@ The fetch lands at `<jobs>\fetch-ca3-pc1-amd-20261006\{bin,packs,src,SHA256SUMS}
|
|||
|---|---|---|---|---|---|
|
||||
| 1 | `pc1-amd-g1-shadow.ps1` | `packaging/ota/publish-jobs.sh add --kind run --target ae432dc7 --id run-ca3-pc1-amd-g1-shadow-20261006 --script tools/ca3-pc1-amd/pc1-amd-g1-shadow.ps1 --timeout-minutes 40 --title "CA3 PC 1: G1 + AMD ladder on the 9070 XT" --deploy` | about 15 min (7 benches of 60 to 90 dispatches of 2^24 at 0.9 to 2 s each, the card switch up to 150 s, the installed-worker cross-check 40 dispatches) | 9070 XT ALONE (its worker stopped; 5090 and 4070 keep mining) | G1 for `mx8+sh256x27` on the AMD vendor (section 5: `RESULT G1 pack=sh256x27 fingerprint=3d2e8245cc084d07 match=yes` turns "AMD vendor NOT RUN" green; the other five packs give the same for the ladder); the 9070 XT column of the item 8 tables (section 3: `RESULT LADDER pack=<p> ops=<N> mhs= watts= uj= gclk_mhz=`), where the card leaves the latency bound by the 5 percent rule; item 1's AMD per-joule row from the control (`RESULT LADDER pack=mx8-devnet-epoch0 ... uj=`: section 3 "Consequences per tier (item 1)", 16 GB AMD row); section 7 rows "item 8: the 9070 XT rows", "item 1: the 9070 XT rows" |
|
||||
| 1b (the watts re-run) | `pc1-amd-watts.ps1` | `packaging/ota/publish-jobs.sh add --kind run --target ae432dc7 --id run-ca3-pc1-amd-watts-20261006 --script tools/ca3-pc1-amd/pc1-amd-watts.ps1 --timeout-minutes 20 --title "CA3 PC 1: watts on the 9070 XT (control + sh256x27)" --deploy` | about 5 min (2 x 100 dispatches at the 18.9 to 19.3 MH/s job 1 measured = 87 to 89 s each, the switch, 12 s of proof) | 9070 XT ALONE | the two `watts= uj=` cells job 1 left owed: item 1's AMD per-joule row (the control) and the candidate's AMD row in the item 8 table (sh256x27); the sampler is proven on the 12 idle seconds first (`RESULT sampler_raw` x3, `RESULT sampler_idle lines= lines_9070_with_watts=`) and the benches do not run without a 9070 XT watts line (`RESULT watts error=<why>`, status failed, card restored) |
|
||||
| G2 (job 2's slot, after "Ember closed") | `pc1-amd-g2.ps1` | its own small kit first: `tools/ca3-pc1-amd/make-g2-kit.sh "$TMPDIR/igneum-ca3-pc1-amd-g2-kit.zip"` (built 6 October 16:2x UTC: sha256 `f4c029c737740e8b59c68ebc94861bbdede2a968836bb22331677844eea2e7aa`, 198,384 bytes, 35 files; worker exe `1c39e123…`), then `packaging/ota/publish-jobs.sh add --kind fetch --target ae432dc7 --id fetch-ca3-pc1-amd-g2-20261006 --file "$TMPDIR/igneum-ca3-pc1-amd-g2-kit.zip" --dir jobs --extract --title "CA3 PC 1 G2 kit" --deploy`, then `packaging/ota/publish-jobs.sh add --kind run --target ae432dc7 --id run-ca3-pc1-amd-g2-20261006 --script tools/ca3-pc1-amd/pc1-amd-g2.ps1 --timeout-minutes 10 --title "CA3 PC 1: gate G2 on the 9070 XT" --deploy` | about 2 min (two serve-mode jobs of 1,024 nonces; the Apple OpenCL dry run took 2.6 s a pack) | beside the miners (a correctness gate) | the "RX 9070 XT" cell of G2 in `docs/plans/counter-asic-3-gate/hash-gates.md` and the AMD line of G2 in status section 5: `RESULT g2 <pack> found <n> of 1024 distinct=<d> sha256 <digest> file <path> exit=<code>` (the hash lane's form) plus `RESULT G2 pack= found= distinct= digest= digest_match=yes|no`; the found lines go to `g2-<pack>.found` in the job folder, never to stdout; the Mac re-check is `tools/ca3-v4/g2-recheck.sh proto-cuda/packs-ca3-v4/<pack> --digest <hex>` (a digest match is 1,024 of 1,024), or `--file` after `publish-jobs.sh add --kind collect --target ae432dc7 --glob "app/jobs/run-ca3-pc1-amd-g2-20261006/g2-*.found"`. Packs: `packs-ca3-v4/mx8-devnet-epoch0` (control) and `packs-ca3-v4/v4-devnet-epoch0` (the candidate with the era at the devnet epoch-0 chain seeds), NOT the first kit's `sh256x27`: that one is the genesis string-seed pack (`IGNEUM_SEED_BYTES_HEX` of 14 bytes) and the worker's job protocol takes only a 64-hex epoch seed, so it cannot be served (the script refuses such a pack with a RESULT line instead of a malformed job). Expected digests (the verifier through g2-recheck.sh on the Mac, equal to the kit worker's serve mode on Apple OpenCL, 16:21Z, 1,024 of 1,024 both): mx8 `2a1824a2e0834287cf53b33f84cb7322014c85efa9788bece574e4e0730940bd`, v4 `435b976a4de57c5a26c8bada8b3b9e003504687c790d4b856bc56619d3aabd18`; the script carries them and prints `digest_match` itself |
|
||||
| G2 (job 2's slot, after "Ember closed") | `pc1-amd-g2.ps1` | its own small kit first: `tools/ca3-pc1-amd/make-g2-kit.sh "$TMPDIR/igneum-ca3-pc1-amd-g2-kit.zip"` (rebuilt 6 October 16:4x UTC from the merged tree with the generator-4 packs and the merged packfile.h: sha256 `a16165483cbef9c9001897e2e965b95051c4e683940e0dbc9bded002ed0a0e7e`, 198,597 bytes, 35 files; worker exe `7dc3b1ee…`; the 16:2x build `f4c029c7…` carried the generator-3 copies and is superseded), then `packaging/ota/publish-jobs.sh add --kind fetch --target ae432dc7 --id fetch-ca3-pc1-amd-g2-20261006 --file "$TMPDIR/igneum-ca3-pc1-amd-g2-kit.zip" --dir jobs --extract --title "CA3 PC 1 G2 kit" --deploy`, then `packaging/ota/publish-jobs.sh add --kind run --target ae432dc7 --id run-ca3-pc1-amd-g2-20261006 --script tools/ca3-pc1-amd/pc1-amd-g2.ps1 --timeout-minutes 10 --title "CA3 PC 1: gate G2 on the 9070 XT" --deploy` | about 2 min (two serve-mode jobs of 1,024 nonces; the Apple OpenCL dry run took 2.6 s a pack) | beside the miners (a correctness gate) | the "RX 9070 XT" cell of G2 in `docs/plans/counter-asic-3-gate/hash-gates.md` and the AMD line of G2 in status section 5: `RESULT g2 <pack> found <n> of 1024 distinct=<d> sha256 <digest> file <path> exit=<code>` (the hash lane's form) plus `RESULT G2 pack= found= distinct= digest= digest_match=yes|no`; the found lines go to `g2-<pack>.found` in the job folder, never to stdout; the Mac re-check is `tools/ca3-v4/g2-recheck.sh proto-cuda/packs-ca3-v4/<pack> --digest <hex>` (a digest match is 1,024 of 1,024), or `--file` after `publish-jobs.sh add --kind collect --target ae432dc7 --glob "app/jobs/run-ca3-pc1-amd-g2-20261006/g2-*.found"`. Packs: `packs-ca3-v4/mx8-devnet-epoch0` (control, class v3, id 73bcbfe8ccf988f1) and `packs-ca3-v4/v4-devnet-epoch0` (the candidate with the era at the devnet epoch-0 chain seeds; since the program-id fix 7c22d0d generator 4, class "v4", id c120d7963abdcd96, kernels and vectors byte-identical), NOT the first kit's `sh256x27`: that one is the genesis string-seed pack (`IGNEUM_SEED_BYTES_HEX` of 14 bytes) and the worker's job protocol takes only a 64-hex epoch seed, so it cannot be served (the script refuses such a pack with a RESULT line instead of a malformed job). The job line's class token is the pack's own `IGNEUM_PROGRAM_CLASS` (`class=v3` for the control, `class=v4` for the candidate), as `pc2-v4-gates.ps1` does since ca61dec: the worker refuses a token that disagrees with the pack (Mac check 16:41Z: `class=v3` on the v4 pack found 0, the known-bad case). Expected digests (the verifier through g2-recheck.sh on the Mac, equal to the kit worker's serve mode on Apple OpenCL at 16:21Z on the generator-3 copies and again at 16:41Z on the generator-4 packs through the merged host.c and packfile.h with `class=v4`, 1,024 of 1,024 every time; UNCHANGED by the re-export, as the byte-identical kernels require): mx8 `2a1824a2e0834287cf53b33f84cb7322014c85efa9788bece574e4e0730940bd`, v4 `435b976a4de57c5a26c8bada8b3b9e003504687c790d4b856bc56619d3aabd18`; the script carries them and prints `digest_match` itself |
|
||||
| 2 | `pc1-amd-family.ps1` | `packaging/ota/publish-jobs.sh add --kind run --target ae432dc7 --id run-ca3-pc1-amd-family-20261006 --script tools/ca3-pc1-amd/pc1-amd-family.ps1 --timeout-minutes 15 --title "CA3 PC 1: family step costs on the 9070 XT" --deploy` | about 3 min (3 runs x up to 3 AMD devices, 26 variants each, builds dominate) | beside the miners (ratios; every row says `card_state=`) | the "RX 9070 XT" column of the item 6 table (section 3: `RESULT FAMILYBEST name=<f> ratio=<r> exact= path=` per family, `RESULT FAMILY ... variant=` for the form each took: `shfla_bperm` = `ds_bpermute_b32`, the one number that could move R3; `shflx_swz` / `shflx_bperm`; `perm_amd` = `v_perm_b32` through `amd_perm` or `perm_c` emulated; `bfe_amd`; `dot4_amd` = the 5 October sudot4 row re-measured; `mm8_gfx12` / `mm8_gfx11` = the WMMA builtins, exact=unverified); section 7 row "item 6: the 9070 XT step costs" |
|
||||
| 3 | `pc1-amd-derive.ps1` | `packaging/ota/publish-jobs.sh add --kind run --target ae432dc7 --id run-ca3-pc1-amd-derive-20261006 --script tools/ca3-pc1-amd/pc1-amd-derive.ps1 --timeout-minutes 15 --title "CA3 PC 1: dr736 on the 9070 XT" --deploy` | about 4 min (3 packs x a 60-dispatch pass and a 3-dispatch pass) | beside the miners (build, compile and bit-exact rows; the rate as a ratio to mx8 in the same job) | the item 2 table's 9070 XT cells (section 3: `RESULT DERIVE pack=dr736-genesis build_ms= build_ms_cached= build_ms_over_mx8= dataset_ms= fingerprint= match= mhs= ratio_to_mx8=`): the OpenCL compile cost of the derivation program against mx8's (the NVRTC +1.1 s finding on AMD), the 1 GiB daily build against mx8's 72 to 77 ms (5 October), bit-exactness on the AMD vendor; section 7 row "item 2: the 9070 XT rows" |
|
||||
| 4 (optional, confirmed by main) | `pc1-4070-shadow.ps1` | `packaging/ota/publish-jobs.sh add --kind run --target ae432dc7 --id run-ca3-pc1-4070-shadow-20261006 --script tools/ca3-pc1-amd/pc1-4070-shadow.ps1 --timeout-minutes 30 --title "CA3 PC 1: item 8 ladder on the RTX 4070" --deploy` | about 11 min (7 benches of 100 to 200 dispatches at an expected 0.4 to 1 s each, approximate: no 4070 row exists yet; NVRTC per pack; the card switch) | 4070 ALONE (its CUDA worker stopped; 5090 and 9070 XT keep mining) | the "small NVIDIA card binds near 100,000" row of item 8 (section 3, consequences per tier at N = 100,000; section 7 "the 4060-class rows"): `RESULT LADDER pack=<p> ops=<N> mhs= watts= uj= sm_mhz=` with nvidia-smi at 1 Hz on that index; G1 fingerprints on a second NVIDIA card |
|
||||
|
|
@ -65,7 +65,7 @@ G1 on the AMD vendor 7 of 7 (every fingerprint equal to the Mac's and the 5090's
|
|||
| `proto-opencl/family-probe.c` on Apple OpenCL (`cc -framework OpenCL`, M5 Max) | every C-form and emulated row bit-exact on both 32-lane groups in 3 repetitions; ratios to the alu chain beside the Metal probe's (rotr 1.14 against Metal 1.13, shl 0.89 / 0.85, shr 0.89 / 0.86, bfe 0.82 / 0.77, andn 0.84 / 0.75, popc 1.00 / 0.87, clz 1.11 / 1.01, sel 0.82 / 0.76, perm emulated 1.52 / 1.13, shflx and shfla through `__local` 1.96 and 2.04 (Metal native 0.86 and 1.91), dot4 emulated signed 4.66 / 4.73); the Apple `dot(char4,char4)` row exact=no as the 5 October finding said; every AMD builtin skipped (`build=skipped`), the khr, intel and amd_media_ops2 forms `build=failed` with the log line, the run went on; `FAMILYBEST` picked the exact rows |
|
||||
| `proto-opencl/host.c` with the build line, `--bench-pack` on all eight kit packs on Apple OpenCL | self-test PASS (96 of 96 lanes) and the 2^24 fingerprint at base 0 equal to the recorded one on every pack: mx8-devnet-epoch0 `90f794dd556f7a3b`, sh256x13 `59ac286fe2a5a9ef`, sh256x27 `3d2e8245cc084d07`, sh256x53 `4f824b15cf2b124a`, sh256x88 `0572522e39a94d8a`, sh64x52 `9dd010f79d8ca9f4`, dr736-genesis `50e3eaa779da4f1e`, dr736-devnet-epoch0 `9553f6d5c667205a`; `build` line: mx8 70 ms, shadow packs 70 to 88 ms, dr736 337 to 356 ms (+270 ms on Apple OpenCL: item 2's compile cost shows on this harness too) |
|
||||
| mingw cross-builds of both exes | clean (`-Wall -Wextra`), 116 KB and 35 KB |
|
||||
| The kit worker's serve mode on Apple OpenCL (G2 dry run, 16:21 UTC, `with-lock.sh run`, load 5.2 to 5.7) | `packs-ca3-v4/mx8-devnet-epoch0` and `v4-devnet-epoch0`: 1,024 of 1,024 found each (2.6 s), nonces 0 to 1,023, the sorted found file's sha256 equal to the verifier's through `tools/ca3-v4/g2-recheck.sh --file` (the coordinator's `igneum-pow` binary): the two digests above |
|
||||
| The kit worker's serve mode on Apple OpenCL (G2 dry runs, 16:21 UTC at load 5.2 to 5.7 on the generator-3 copies; 16:41 UTC at load 27 to 57, a loaded box, on the generator-4 packs with the merged packfile.h and the `class=v4` token: a correctness run, no number taken) | `packs-ca3-v4/mx8-devnet-epoch0` and `v4-devnet-epoch0`: 1,024 of 1,024 found each (2.6 s), nonces 0 to 1,023, the sorted found file's sha256 equal to the verifier's through `tools/ca3-v4/g2-recheck.sh --file` (the coordinator's `igneum-pow` binary): the two digests above |
|
||||
| `tools/ci/bash-body-check.sh` on the six scripts | 0 bash bodies, all parse |
|
||||
| `tools/ci/kit-path-check.sh` on the six scripts | 6 of 6 kit paths checked before use |
|
||||
| `tools/ci/prover-socket-check.sh`, `copied-sources-check.sh`, `no-conflict-markers.sh` | pass |
|
||||
|
|
|
|||
|
|
@ -1,7 +1,7 @@
|
|||
# Counter ASIC 3.0, PC 1 AMD job G2 (6 October 2026): gate G2 of docs/plans/counter-asic-2-rollout.md section 7 on
|
||||
# the AMD vendor (docs/plans/counter-asic-3-gate/hash-gates.md, the "RX 9070 XT" cell): the CPU verifier exact on
|
||||
# 1,024 hashes from the RX 9070 XT (machine ae432dc7, gfx1201) for the class v4 candidate and the class v3 control.
|
||||
# The hash lane's form (tools/ca3-v4/pc2-v4-gates.ps1, ca3-coord b7cd99e): one serve-mode job per pack of 1,024 nonces
|
||||
# The hash lane's form (tools/ca3-v4/pc2-v4-gates.ps1, ca3-coord ca61dec: the class token from the pack's IGNEUM_PROGRAM_CLASS): one serve-mode job per pack of 1,024 nonces
|
||||
# at target ffffffffffffffff and prehash 00..01 (every nonce a "found g2 <nonce> <hash>" line), through the kit's
|
||||
# igneum-worker-opencl.exe (--serve --pack <dir> --device <gfx1201>); the found lines never go to stdout: each pack's
|
||||
# lines go to g2-<pack>.found in this job's folder (nonce order, LF, no BOM) and ONE line per pack reports the count,
|
||||
|
|
@ -10,8 +10,9 @@
|
|||
# 1,024 of 1,024; on a mismatch the file is fetched with `publish-jobs.sh add --kind collect --glob
|
||||
# "app/jobs/<job id>/g2-*.found"` and diffed line by line.
|
||||
# Packs (the G2 kit, fetch job fetch-ca3-pc1-amd-g2-20261006): proto-cuda/packs-ca3-v4/v4-devnet-epoch0 (the candidate
|
||||
# mx8+sh256x27 composed with the era at the devnet epoch-0 chain seeds) and packs-ca3-v4/mx8-devnet-epoch0 (the class v3
|
||||
# control). The first kit's sh256x27 is the genesis STRING-seed pack (IGNEUM_SEED_BYTES_HEX of 14 bytes), which the
|
||||
# mx8+sh256x27 composed with the era at the devnet epoch-0 chain seeds; generator 4, class "v4", program id
|
||||
# c120d7963abdcd96 since the program-id fix 7c22d0d, kernels and vectors byte-identical) and packs-ca3-v4/mx8-devnet-epoch0
|
||||
# (the class v3 control, id 73bcbfe8ccf988f1). The first kit's sh256x27 is the genesis STRING-seed pack (IGNEUM_SEED_BYTES_HEX of 14 bytes), which the
|
||||
# worker's job protocol cannot take (epoch_seed must be 64 hex), so G2 runs on the chain-seed packs the 5090 and the
|
||||
# Mac ran (8 x 1,024 of 1,024 there). Beside the miners (correctness, not rate): no card is switched; the process
|
||||
# list says what the card carried. Never quits, pauses, resumes or updates the installed app; never writes
|
||||
|
|
@ -71,9 +72,11 @@ function G2-JobLines([string] $d) {
|
|||
$epoch = ([regex]::Match($ph, '#define IGNEUM_SEED_BYTES_HEX "([0-9a-f]+)"')).Groups[1].Value
|
||||
$day = ([regex]::Match($ph, '#define IGNEUM_DAY_BYTES_HEX "([0-9a-f]+)"')).Groups[1].Value
|
||||
$tokens = ''
|
||||
$cls = [regex]::Match($ph, '#define IGNEUM_PROGRAM_CLASS "(v[23])"')
|
||||
$cls = [regex]::Match($ph, '#define IGNEUM_PROGRAM_CLASS "(v[234])"')
|
||||
$era = [regex]::Match($ph, '#define IGNEUM_ERA_SEED_HEX "([0-9a-f]+)"')
|
||||
if ($cls.Success -and $cls.Groups[1].Value -eq 'v3') { $tokens = ' class=v3'; if ($era.Success) { $tokens += ' era=' + $era.Groups[1].Value } }
|
||||
# the class token is the pack's own (v3 or v4 since ca3-v4-node 7c22d0d: a class v4 pack is generator 4 and says "v4";
|
||||
# the worker refuses a token that disagrees with the pack)
|
||||
if ($cls.Success -and $cls.Groups[1].Value -ne 'v2') { $tokens = ' class=' + $cls.Groups[1].Value; if ($era.Success) { $tokens += ' era=' + $era.Groups[1].Value } }
|
||||
return @("job g2 $g2Pre ffffffffffffffff 0 1024 $epoch $day$tokens", 'quit')
|
||||
}
|
||||
$ok = 0; $digestMatch = 0
|
||||
|
|
|
|||
Loading…
Reference in a new issue