yangzhuxinyzxandOpenAI Codex a17dcdf94e [Doc] Preserve the rejected IQ3 register-pair screen ...

Retain the original four-rank measurements and rejection decision from #1068 without adding the unvalidated prototype. The stored generated-CUDA fingerprint differs from the current prototype generator; these records are historical evidence only.

Co-authored-by: OpenAI Codex <noreply@openai.com>
Signed-off-by: yangzhuxinyzx <153831768+yangzhuxinyzx@users.noreply.github.com>

2026-10-09 09:47:08 +08:00

..

gguf_collective_norm_20261007.json

[Doc] Record matched GGUF collective norm model results

2026-10-07 02:02:32 +08:00

gguf_dflash2_projection_ledger_20261005.json

…

gguf_dmv_context_recovery_20261006.json

[Doc] Separate projection increments from graph-context costs

2026-10-06 23:05:38 +08:00

gguf_dmv_nvlink_model_20261006.json

[Doc] Attribute the matched NVLink projection graph

2026-10-06 21:23:53 +08:00

gguf_dmv_operator_audit_20261006.json

[Doc] Audit original DMV operators against current projection routes

2026-10-06 13:51:28 +08:00

gguf_fp16_short_context_splits_20261007.json

[Kernel] Improve FP16 grouped verification at short contexts (#1022)

2026-10-07 12:07:00 +08:00

gguf_iq2_signed_nibbles_20261007.json

[Doc] Keep IQ2 model evidence compact

2026-10-07 07:20:48 +08:00

gguf_lut4_experts_20261006.json

[Doc] Record matched IQ4 model quality and partition screening

2026-10-06 05:47:14 +08:00

gguf_lut4_experts_lanes8_20261006.json

[Doc] Record matched IQ4 model quality and partition screening

2026-10-06 05:47:14 +08:00

gguf_lut4_experts_lanes16_20261006.json

[Doc] Record matched IQ4 model quality and partition screening

2026-10-06 05:47:14 +08:00

gguf_projection_clock_calibration_20261007.json

[Kernel] Add device warp timing for GGUF projection diagnostics (#1018)

2026-10-07 10:49:56 +08:00

gguf_qkv_profile_calibration_20261007.json

[Kernel] Fuse measured three-format GGUF QKV projections (#1017)

2026-10-07 09:24:37 +08:00

gguf_qkvz_tp4_m8_20261006.json

…

gguf_qpn_attention_inputs_20261006.json

[Doc] Record installed GGUF attention projection gates

2026-10-06 04:24:15 +08:00

gguf_qpn_canonical_streams_20261006.json

[Doc] Retain canonical QPN screening and reject slower routes

2026-10-06 04:34:03 +08:00

gguf_qpn_output_batch_20261006.json

[Doc] Add same-clock head reference and reporting checks

2026-10-06 05:24:52 +08:00

gguf_qpn_projection_batch_20261006.json

[Doc] Attribute the post-QPN graph-linked projection round

2026-10-06 03:57:20 +08:00

gguf_qpn_register_pair_20261008.json

[Doc] Preserve the rejected IQ3 register-pair screen

2026-10-09 09:47:08 +08:00

gguf_qpn_small_outputs_20261006.json

[Doc] Record qualified QPN output projection results

2026-10-06 04:03:45 +08:00

gguf_tp4_verify_cost_20261008.json

[Doc] Measure TP4 GGUF DFlash2 verification cost

2026-10-08 03:07:13 +08:00

sm70_dflash_page832_20261007.json

[Doc] Record qualified DFlash window split speed and quality

2026-10-07 04:43:45 +08:00

sm70_dflash_page1648_20261007.json

[Core] Keep DFlash split windows with E4M3 hybrid pages (#1024)

2026-10-07 12:53:06 +08:00