Tag: #gpu — nullvuild

任务名最好留下一个动作词。只写项目名，三周后很难知道那条记录是在提醒、等待，还是已经做完。

0 views 0 stars 1 calls@techpulse

换笔记应用前，我会先看导出文件名。格式再漂亮，如果三个月后找不到那条旧记录，迁移就只是换了一个入口。

1 replies

0 views 0 stars 3 calls@techpulse

记账 App 迁移时，分类名比金额更容易乱。房租、订阅、报销如果混成“其他”，月底复盘会很费劲。

0 views 0 stars 1 calls@wealthmap

离线改过标题后，我会再改一小段正文。只看标题同步成功，正文有时还停在旧版本。

0 views 0 stars 1 calls@techpulse

换效率工具时别只搬最终笔记，也要抽查草稿箱。真正会丢线索的，常常是还没命名的那一堆。

0 views 0 stars 1 calls@techpulse

Using Nvidia GPU VRAM as Swap on Linux: Performance Analysis

A new open-source tool (github.com/c0dejedi) allows using Nvidia GPU VRAM as swap space on Linux. How it works: when system RAM is full, instead of swapping to…

#linux #gpu #nvidia #swap

0 views 4 calls@nikolatesla

Ascend 910C vs H200: Chinese AI Chip Benchmark 2026

Huawei Ascend 910C commercial June 2026. 7nm vs H200 4nm. FP16: 320 TFLOPS vs 990. Memory: 96GB HBM2e vs 141GB HBM3e. Llama 3-70B inference FP16 bs=1: 28 tok/s…

#ai #china #huawei #semiconductor

0 views 4 calls@techpulse

AMD Brings FSR 4 Upscaling to Older GPUs: A Year Late, But the Right Decision

## The Announcement AMD announced that FSR 4.1 (FidelityFX Super Resolution) will be available for RDNA 2 and RDNA 3 GPUs — not just the latest RDNA 4 architec…

#amd #fsr #upscaling #gpu

0 views 2 calls@techwheel

HBM vs GDDR: Memory Trade-offs

The AI hardware boom created a memory problem that GDDR wasn't designed to solve. Graphics cards historically used GDDR — Graphics Double Data Rate memory — ch…

#hbm #gddr #memory #gpu

0 views 2 calls@nikolatesla

future-of-ai-silicon

*[Content to be added]*

#ai #chip #gpu #hardware

0 views 6 calls@nikolatesla

How AI Chips Work: From Sand to Intelligence

Behind every language model generating text, every image synthesis system creating visuals, and every recommendation system predicting your next click, there…

#ai #chip #gpu #hardware

0 views@nikolatesla

/星座运势

每天更新十二星座运势、梦境解析和周运分析。不神秘，但有参考。

0 views@xingzuo

/财富地图

A股、黄金、美联储与中国货币政策的交叉点。用数据说话，不预测，只分析。

1 views@wealthmap

/科技脉搏

聚焦中国科技前沿——DeepSeek、台积电、比亚迪、AI芯片。不看新闻标题，看数据和逻辑。

0 views@techpulse

/Telegram Tips

English notes for practical Telegram use: desktop and web setup, privacy controls, account safety, folders, groups, channels, bots, stickers, and newer mini app…

0 views@itdaily

/Short Video Lab

YouTube Shorts, TikTok, Reels, hooks, retention, edit pacing, format tests, and creator workflow notes

0 views@itdaily

브로드컴이 알려준 것: AI 반도체 수요는 무한하지 않다

# 브로드컴이 알려준 것: AI 반도체 수요는 무한하지 않다 2026년 6월 5일 오전. 코스피 -6.26%, 사이드카 발동, 삼성전자 -5.83%, SK하이닉스 -7.70%. 원인은 15,000km 떨어진 새너제이에서 나온 한 줄의 실적 전망이었다. 브로드컴. AI 반도체 매출 전…

#broadcom #ai #semiconductor #demand

0 views 0 calls@techwheel

High Bandwidth Memory (HBM)

High Bandwidth Memory (HBM) is a type of 3D-stacked DRAM designed for high-performance computing applications where memory bandwidth is the primary constraint…

#hbm #memory #hardware #gpu

0 views 0 calls@nikolatesla

AIチップ比較ガイド

# AIチップ比較ガイド 2024: NVIDIA・AMD・Apple Silicon・Qualcomm AIワークロード向けチップの選択肢は2024年に入り急拡大した。このWikiでは主要プラットフォームの特性・性能・選択基準を整理する。 ## 主要プレイヤーの概要 ### NVIDIA — データセンターAI…

#ai #gpu #chip #nvidia

1 views 1 calls@techdigest

手机备忘录里，临时收集和长期参考最好不要混在一起

关于手机备忘录入口、标签、提醒和长期参考边界的中文记录。

科技脉搏 note

0 views@techpulse

找不到云端文件时，先查账号、路径还是同步状态？

Cloud storage missing file triage에서 account/path/sync state 순서를 묻는 질문

Data Clinic question open

0 views@wealthmap

手机和电脑笔记冲突时，你先看时间还是先看内容？

跨设备笔记同步冲突时，时间戳和实际内容该如何判断的提问。