GitHub - ggml-org/ggml: Tensor library for machine learningYou signed in with another tab or window. Reload to refresh your session. You signed out in another tab or window. Reload to refresh your session. You switched accounts on another tab or window. Reload to refresh your session. Dismiss alert
Intel Publishes Blazing Fast AVX-512 Sorting Library, Numpy Switching To It For 10~17x Faster Sorts - Phoronix
Intel Publishes Blazing Fast AVX-512 Sorting Library, Numpy Switching To It For 10~17x Faster Sorts Written by Michael Larabel in Intel on 15 February 2023 at 04:00 PM EST. 51 Comments Intel recently published an open-source C++ header file library for high performance SIMD-based sorting, which initially is focused on providing a lightning fast AVX-512 quicksort implementation. As of today that co
GitHub - ggml-org/whisper.cpp: Port of OpenAI's Whisper model in C/C++
Stable: v1.7.6 / Roadmap High-performance inference of OpenAI's Whisper automatic speech recognition (ASR) model: Plain C/C++ implementation without dependencies Apple Silicon first-class citizen - optimized via ARM NEON, Accelerate framework, Metal and Core ML AVX intrinsics support for x86 architectures VSX intrinsics support for POWER architectures Mixed F16 / F32 precision Integer quantization
プãƒã‚°ãƒ©ãƒ ã®ãƒ•ã‚¡ã‚¤ãƒ«èªã¿æ›¸ãã§ã€Œmmapã€ã‚’使ã†ã¨é€Ÿåº¦ãŒå‘上ã™ã‚‹ç†ç”±ã¨ã¯ï¼Ÿ
ソフトウェア開発ã«ãŠã„ã¦ã€Œãƒ•ã‚¡ã‚¤ãƒ«ã®èªã¿æ›¸ãã€ã¯ä½¿ç”¨é »åº¦ã®é«˜ã„æ“作ã§ã‚ã‚Šã€é€Ÿåº¦ã®å‘上ã¯ã‚½ãƒ•ãƒˆã‚¦ã‚§ã‚¢å…¨ä½“ã®ãƒ‘フォーマンスã«å¤§ãã影響ã—ã¾ã™ã€‚ãã‚“ãªãƒ•ã‚¡ã‚¤ãƒ«ã®èªã¿æ›¸ãæ“作を行ã†éš›ã«ã€Œmmapを使ã£ãŸæ–¹ãŒé€šå¸¸ã®ã‚·ã‚¹ãƒ†ãƒ コールよりも高速ã«ãƒ•ã‚¡ã‚¤ãƒ«ã‚’æ“作ã§ãã‚‹ç†ç”±ã€ã«ã¤ã„ã¦ã€ãƒ–リティッシュコãƒãƒ³ãƒ“ア大å¦ã®å‡†æ•™æŽˆã§ã‚ã‚‹Alexandra Fedorovaæ°ãŒèª¬æ˜Žã—ã¦ã„ã¾ã™ã€‚ Why mmap is faster than system calls | by Alexandra (Sasha) Fedorova | Medium https://sasha-f.medium.com/why-mmap-is-faster-than-system-calls-24718e75ab37 OS上ã§ãƒ¦ãƒ¼ã‚¶ãƒ¼ãŒãƒ—ãƒã‚°ãƒ©ãƒ を実行ã™ã‚‹éš›ã€ãƒ—ãƒã‚°ãƒ©ãƒ ã¯ã€Œãƒ¦ãƒ¼ã‚¶ãƒ¼ç©ºé–“ã€ã¨ã€Œã‚«ãƒ¼ãƒãƒ«ç©ºé–“ã€ã¨å‘¼ã°ã‚Œã‚‹2種類ã®é ˜åŸŸã‚’利用ã—ã¾ã™ã€‚
x86ã§doubleãŒfloatより速ã„ã‹ã©ã†ã‹ã‚’検証ã—ã¦ã¿ãŸ - Qiita
昔話 ãã‚Œã¯æ˜”々ã®ã“ã¨ã€‚ x86ã«ã¯æµ®å‹•å°æ•°ç‚¹æ¼”ç®—ã‚’è¡Œã†æ‰‹æ®µãŒãªãã€å¤–付ã‘ã®æµ®å‹•å°æ•°ç‚¹æ¼”算ユニットを接続ã™ã‚‹ã¨ã„ã†æ‰‹æ³•ã§ã€æµ®å‹•å°æ•°ç‚¹æ¼”算を実ç¾ã—ã¦ã„ãŸã®ã§ã‚ã£ãŸã€‚ x87ã¨å‘¼ã°ã‚ŒãŸãã‚Œã¯ã¨ã¦ã‚‚エクセレントãªã‚·ã‚¹ãƒ†ãƒ ã§â€¦ã¨ã„ã†è©±ã¯wikipediaã«è²ã‚ã†ã€‚ https://ja.wikipedia.org/wiki/Intel_8087 é‡è¦ãªã®ã¯ã€x87ãŒå†…部表ç¾ã¨ã—ã¦80bitã®æ‹¡å¼µå€ç²¾åº¦ã‚’使ã£ã¦ã„ã‚‹ã€ã¨ã„ã†ã“ã¨ã€‚ ã“ã‚Œã®ãŠã‹ã’ã§ã€x87ã«ãŠã„ã¦ã¯ã€ç¢ºã‹ã«~~doubleã®ã»ã†ãŒé€Ÿã‹ã£ãŸ (floatã ã¨doubleã¸ã®ã‚ャストコストãŒç™ºç”Ÿã™ã‚‹ãŸã‚)~~ 嘘ã ã‚ãれ。ASM見ãŸã‚‰åˆ¥ã«ã‚ャストã¨ã‹ã—ã¦ãªã‹ã£ãŸã‚。 ã©ã£ã¡ã‹ã¨ã„ã†ã¨ä¸¸ã‚ã®å½±éŸ¿ã§ç²¾åº¦ãŒç•°ãªã‚‹ã“ã¨ã®ã»ã†ãŒé‡è¦ã ã‚。 改ã‚ã¦èª¿ã¹ã¦ã¿ã‚‹ã¨ã€doubleãŒé€Ÿã„ã¨ã•ã‚Œã¦ã„る資料ã«ã¤ã„ã¦ã¯ã‚ã‚“ã¾ã‚Šãªã„ã“ã¨ã«æ°—付ã。 (åŒç‰ã¨ã—ã¦ã„る資料
1
ランã‚ング
ランã‚ング
メンテナンス
リリースã€éšœå®³æƒ…å ±ãªã©ã®ã‚µãƒ¼ãƒ“スã®ãŠçŸ¥ã‚‰ã›
最新ã®äººæ°—エントリーã®é…ä¿¡
j次ã®ãƒ–ックマーク
kå‰ã®ãƒ–ックマーク
lã‚ã¨ã§èªã‚€
eコメント一覧を開ã
oページを開ã
GitHub - ggml-org/llama.cpp: LLM inference in C/C++
GitHub - tanakamura/instruction-bench: instruction-bench
{{#tags}}- {{label}}
{{/tags}}