Guide-yakabatana mibvunzo · Pakati Nhanho

Groq Quiz

Wongorora kuti Groq's LPU inosanganisa sei compiler kuronga, pa-chip ndangariro, uye inference-yekutanga dhizaini yekukurumidza kushumira modhi.

Nzira dzekutungamira dzinoenderanaGroq
Mubvunzo 1 ye 8

Groq's LPU inomirirei?

Tarisa mubvunzo wese8 mibvunzo

Mhinduro uye tsananguro zvinoonekwa paunenge uchitora bvunzo.

  1. Groq's LPU inomirirei?
    • Mutauro Wokugadzirisa Chikamu
    • Logic Processing Unit
    • Linear Performance Unit
    • Latency Prediction Unit
  2. Groq's chips anonyanya kugadziridzwa kuti ndeipi chikamu cheAI workflow?
    • Kudzidzisa mienzaniso kubva pakutanga
    • Inference (kumhanyisa mamodheru akadzidziswa)
    • Data labeling
    • Image compression
  3. Chii chinotsanangudza hunhu hweLPU's architecture?
    • Dynamic hardware kuronga nehombe cache
    • Kuvimba nekutenderera dhisiki kuchengetedza
    • Deterministic, compiler-yakarongwa kuuraya
    • Random rairo yekuraira
  4. Ndiani akatanga Groq, uye nderipi rakanga riri basa rake raizivikanwa?
    • Muvambi weNvidia's GPU mutsara
    • Muvambi we OpenAI
    • Muvambi weCPU
    • Jonathan Ross, abatsira kugadzira Google's TPU
  5. Chii chinopa GroqCloud?
    • API yekuwana yekumhanyisa AI modhi paGroq hardware
    • A social media network
    • Vhidhiyo yekugadzirisa suite
    • Basa rekuchengetedza faira
  6. Ndechipi chirevo chinosiyanisa Groq nenzira kwayo?
    • Groq ndiye Elon Musk's chatbot
    • Groq ikambani yechip, yakasiyana neiyo chatbot Grok
    • Groq ishamwari app
    • Groq inogadzira madhiraivha akaomarara
  7. What role does on-chip SRAM play in Groq's LPU design?
    • It acts as an external archive for completed user conversations
    • It replaces the compiler's instruction schedule
    • It stores and retrieves data close to the compute units with low latency
    • It stores only model outputs after inference ends
  8. In Groq's programmable assembly-line architecture, what does the compiler schedule?
    • Which user requests should be answered first
    • The training labels for each model
    • The order of tokens in the model's vocabulary
    • The movement of data and instructions through compute units