search

hahnyuan / LLM-Viewer

Analyze the inference of Large Language Models (LLMs). Analyze aspects like computation, storage, transmission, and hardware roofline model in a user-friendly interface.
MIT License
275 stars 32 forks source link

请问hardware_param里面的onchip_buffer指的是哪部分的size？ #10

Closed pangsg closed 1 week ago

pangsg commented 2 months ago

我查了一下资料，v100的L1和SRAM大小一共是128KB，但是repo里面写的V100 onchip_buffer是20480e3，这块想请教下指的是哪部分buffer

hahnyuan commented 2 months ago

Please see the chart in https://images.nvidia.com/content/volta-architecture/pdf/volta-architecture-whitepaper.pdf Page 10. The V100 has 20480 KB register file.

pangsg commented 2 months ago

thanks a lot.

---原始邮件--- 发件人: "Zhihang @.> 发送时间: 2024年7月16日(周二) 下午4:43 收件人: @.>; 抄送: @.**@.>; 主题: Re: [hahnyuan/LLM-Viewer] 请问hardware_param里面的onchip_buffer指的是哪部分的size？ (Issue #10)

Please see the chart in https://images.nvidia.com/content/volta-architecture/pdf/volta-architecture-whitepaper.pdf Page 10. The V100 has 20480 KB register file.

— Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you authored the thread.Message ID: @.***>