summaryrefslogtreecommitdiff
path: root/docs/hardware
AgeCommit message (Collapse)Author
13 days第二代原语优化器: SignLine/ConsensusLion/DitherLion ↵Yuren Hao
实装+本地筛选电池发射 R4后设计原则=只追不翻腾的目标。SignLine核心洞察: 正行/列缩放在sign内是无操作, 作为脉冲幅度 在sign外才有效=因子化自适应的模拟原生形态; ConsensusLion=双漏率电容+符合门的噪声否决; DitherLion=免费噪声做期望软sign。KFAC-lite(追平稳的活动协方差)列为下一轮。 9臂与stage A同协议本地筛选中; 发射纪律教训追加: 发射与等待必须分Bash调用(超时杀树会扫掉setsid)。 Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_014FAPDWQ49M5Ye3NpTndTpn
13 daysEqOLion R4(真实轨迹判决): ↵Yuren Hao
冻结极因子几十步退相关至~0.53=慢漂移前提对符号模式不成立(极均衡 放大病态尾部,小σ奇异向量高速旋转); 子满秩探针NaN, 满秩单迭代仅平基线; 存活空间收窄到温启动 全NS(3-5×)或Occam臂(行/列AGC)升主候选; 限定=最早期轨迹为最难工况, 晚期重dump后再判。 附: --dump_mom探针实装(601步×3矩阵/276MB)。 Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_014FAPDWQ49M5Ye3NpTndTpn
13 daysEqOLion设计+当天三轮证伪: O(1)秩一探针死(两轮), ↵Yuren Hao
真实动量矩阵判决性救活(sign(M) vs sign(polar) 中位仅0.70, 偏斜55-156x, 正交化改写~30%符号=实在载荷); 成本算术修正(O(d)探针/漂移时标≈C次 通过/步, 非1-2%童话); 主分支改为搭车探针(nudged相位数千免费活动向量/步, 代价=数据度量白化 无理论); 下一轮=真实动量逐步轨迹上的追踪测试 Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_014FAPDWQ49M5Ye3NpTndTpn
13 days优化器审计修订: ↵Yuren Hao
OLion(sign输出→正交化器精度要求塌到1bit)推翻Muon模拟死刑的前提; 三条未探路线(平衡态正交化/跟踪摊销/局部规则白化); Hyperball=每矩阵一个AGC的最便宜自适应+跨宽深LR迁移; 价目表实验扩臂 Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_014FAPDWQ49M5Ye3NpTndTpn
13 days表征方案修订(用户纠正): 优化器建模义务不因不造板而消失Yuren Hao
撤回"Muon问题消解"的说法: 消失的只是台架工程风险, 被模拟的系统必须带基底做得出的优化器。 新增: 测量项#6写入路径(单步增量/上下不对称/电荷注入/最小可写增量=硬件版舍入地板, 与fp32 读出地板同构=预注册预测点); 优化器行为级设计节(SGDM/Lion/factored-Adam三候选全部从元件表 取参数, 旧"Lion发散"判决系修复前旧架构须复测); 优化器价目表实验从可选升级为必做(理想数学 版与叠加实测缺陷版两级分开报); 增设混合信号对照臂使全模拟vs混合信号成为可测轴。 Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_014FAPDWQ49M5Ye3NpTndTpn
13 days硬件线转向(08-05用户决定, Samuel建议): ↵Yuren Hao
元件表征方案取代MVP学习板 买样实测imperfection表→器件锚定的Python模拟。清单10类元件~$300, 统一测量协议(DC传递/ 单元间散布/噪声谱/漂移/电容泄漏), 接入现有故障注入机械; 持久缺陷与刷新噪声分开喂, 对应共适应交叉实验的两类理论预期; ADC有效位数直接给对比读出的物理epsilon(接地板律做 预注册预测)。删除学习环路调试的全部风险项。 Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_014FAPDWQ49M5Ye3NpTndTpn
2026-07-31Dillavou通话材料v3: 机制段按RESULT ↵Yuren Hao
77-79改判(读出地板+零成本修复+小尺寸EP胜BP), Q2改为板上对比读出/physical epsilon, cheat sheet加破案更新节 Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_014FAPDWQ49M5Ye3NpTndTpn
2026-07-16RESULT 28: Stage-0 gate #1 verdict (naked resident cells need >=10-12b; T64 ↵Yuren Hao
word-streaming exempt; qcomp = the decisive arm) Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_014FAPDWQ49M5Ye3NpTndTpn
2026-07-15T64 procurement sheet: Taobao AD7528JN @ ¥3/pc (user-verified) — MDAC ↵Yuren Hao
line $11.4k -> ~$1.4k incl 50% screening overage; 3-phase screen protocol tied to Stage-0/Stage-1 gates Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_014FAPDWQ49M5Ye3NpTndTpn
2026-07-13TinyEP hostile audit archived: DEAD per own claim class, SURVIVES-DEGRADED ↵Yuren Hao
rescoped; portable gold = tied-KV DenseAM energy + same-cell bidirectional access Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_014FAPDWQ49M5Ye3NpTndTpn
2026-07-13M-scale supply-chain down-select (20-agent): null demolished (UIUC die modes ↵Yuren Hao
= NO on published record), T64 MDAC winner, grade ceiling B-, claim ruling + GPT TinyEP proposal archived Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_014FAPDWQ49M5Ye3NpTndTpn
2026-07-1318-agent adversarial substrate down-select: incumbent CONFIRMED-MODIFIED ↵Yuren Hao
(-260), step-0 fault gates, refined partner asks Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_014FAPDWQ49M5Ye3NpTndTpn
2026-07-12Hardware outreach v2: gate lifted, clockless-MVP story, Dillavou wave-1, ↵Yuren Hao
brief rewritten to cascade-era Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_014FAPDWQ49M5Ye3NpTndTpn
2026-07-12E-tier wave-1 ledger (RESULT 14 + HW map rows) + cost model v2 recalibrated ↵Yuren Hao
on measured H200 datapoint Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_014FAPDWQ49M5Ye3NpTndTpn
2026-07-11Component->analog-hardware map: full stack (forward path A.1-10, training ↵Yuren Hao
machinery B.11-17, cross-cutting physics C, demo scoping D) with reuse sources, feedback-path costs, tolerance status per item
2026-07-10HW dossier: OLMo2-standard block analog audit (net win; no new ↵Yuren Hao
non-reciprocity; 3 new E-tier tolerance items)
2026-07-09HW: cascade re-pricing under the reuse doctrine — depth becomes a memory ↵Yuren Hao
line-item (time-mux one trainable block), nudged duty drops ~50x, leash instrumentation deleted, SRAM-CIM alone suffices for Demo-1 (device fab optional) Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_014FAPDWQ49M5Ye3NpTndTpn
2026-07-09HW economics retraction: doctrine currency restored (COTS-stitch + FPGA + ↵Yuren Hao
UIUC collab, $5-20k Demo-1 BOM); optimizer state digital-side by design => Adam ~free, Muon unblocked at demo scale, choice is algorithmic Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_014FAPDWQ49M5Ye3NpTndTpn
2026-07-09HW BoM correction: Lion/SGD ruled out for EP training (empirical, abl_lion2 ↵Yuren Hao
blew @600 + prior tests) — factored-Adam is the MANDATORY analog baseline, not premium tier Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_014FAPDWQ49M5Ye3NpTndTpn
2026-07-09HW: optimizer BoM audit per cascade block — Adam analog-compatible via ↵Yuren Hao
factored-AGC (+2x memory, zero ADC), Muon priced out of analog (matrix-matrix NS + full-gradient ADC tax), Lion native baseline Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_014FAPDWQ49M5Ye3NpTndTpn
2026-07-09HW dossier: analog softmax/attention (12 verified claims) — Elfadel-Wyatt ↵Yuren Hao
1993 reciprocal 'entropic resistor' softmax (EP-composable, PAR escape realized), charge-domain KV gain-cells (Leroux 2025), the analog-training gap confirmed unclaimed; +-1% mismatch->+-1% softmax linear tolerance coeff Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_014FAPDWQ49M5Ye3NpTndTpn
2026-07-07Wall-3 theorem addendum: arXiv:2603.26969 (PAR) hardens the J^T obstruction; ↵Yuren Hao
price list + Demo-M reframe Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_014FAPDWQ49M5Ye3NpTndTpn
2026-07-03Initial commit: ept — backprop-free equilibrium transformer (EP)Yuren Hao
Code (ep_run/), organized docs (docs/{method,campaign,hardware,outreach,paper}), analysis scripts (scripts/), ONBOARDING.md entry point. Large data/checkpoints git-ignored (share separately). Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_014FAPDWQ49M5Ye3NpTndTpn
generated by cgit v1.2.3 (git 2.25.1) at 2026-08-18 16:06:30 +0000