SA260324R-0060_r1

SA260324R-0060

① SA Source

• Source: 開啟完整 SA 文章
• Section: SRAM and Memory Hierarchy
• Line hint: 69

Context Before

SRAM machines such as Groq’s LPU therefore enable very fast time to first token and tokens per second per user but

Evidence

at the expense of total throughput, as their limited SRAM capacity quickly gets saturated by weights, with little left over for KVcache that grows as more users are batched.

Context After

GPUs win for throughput and cost as we have shown.

② Atomic Claim

但代價是 total throughput 較低，因為有限的 SRAM 容量很快會被模型 weights 填滿，剩餘可供 KVcache 使用的空間不多，而 KVcache 會隨 batch 中使用者增加而成長。

• Epistemic Mode: ASSERTED
• Mapping Status: PARTIAL

③ Semantic Frame

{
  "frame_type": "CLAIM_ONLY",
  "qualifiers": {
    "condition_text": "as more users are batched",
    "numeric_mentions": [],
    "temporal_mentions": []
  },
  "reason": "句中 their 承接前一條 SRAM machines／Groq LPU；為避免擅自補主詞，以 claim-only 保留完整因果鏈。",
  "statement": "total throughput 較低，因有限 SRAM 容量很快被 weights 填滿，留給 KVcache 的空間有限，而 KVcache 隨 batch 使用者增加而成長。"
}

④ Canonical Entity Mapping

Role	Surface Label	Canonical Target
—	—	—

⑤ Human Review

請在 Properties 逐項確認：

• 原文 → Atomic Claim 是否忠實
• Atomic Claim → Semantic Frame 是否忠實
• Canonical Entity mapping 是否正確
• Epistemic mode 是否保留原文語氣
• 最後選擇 review_action

Review state

Markdown 內文不是正式 approval。只有 Apply bridge 寫入的 Decision Ledger event 才是正式決策。