rydotã®ãƒ–ãƒƒã‚¯ãƒžãƒ¼ã‚¯ - ã¯ã¦ãªãƒ

rydot id:rydot

rydotã®ãƒ–ãƒƒã‚¯ãƒžãƒ¼ã‚¯ (12,630)

${{author_name}}$

{{{comment_expanded}}}

{{label}}

{{#is_bookmark}}ãƒªã‚¹ãƒˆ{{/is_bookmark}}{{^is_bookmark}}ãƒªãƒ³ã‚¯{{/is_bookmark}}

${{author_name}}$
{{author_name}}{{created}}
{{ #comment }}{{ comment }}{{ /comment }}
- {{ label }}

${{author_name}}$

{{{comment_expanded}}}

{{label}}

{{#is_bookmark}}ãƒªã‚¹ãƒˆ{{/is_bookmark}}{{^is_bookmark}}ãƒªãƒ³ã‚¯{{/is_bookmark}}

Bolt Graphicsã€RTX 5090ã®10å€é€Ÿã„ã¨ã„ã†ã€ŒZeusã€ GPUã‚’ç™ºè¡¨ | XenoSpectrum
ãƒ†ã‚¯ãƒŽãƒã‚¸ãƒ¼ Bolt Graphicsã€RTX 5090ã®10å€é€Ÿã„ã¨ã„ã†ã€ŒZeusã€ GPUã‚’ç™ºè¡¨ ç±³å›½ã®ã‚¹ã‚¿ãƒ¼ãƒˆã‚¢ãƒƒãƒ—Bolt Graphicsã¯ã€NVIDIA RTX 5090ã¨æ¯”è¼ƒã—ã¦æœ€å¤§10å€ã®æ€§èƒ½ã‚’èª‡ã‚‹ã¨ã„ã†æ¬¡ä¸–ä»£GPUã€ŒZeusã€ã‚’ç™ºè¡¨ã—ãŸã€‚æœ€å¤§2.25TBã®ãƒ¡ãƒ¢ãƒªå®¹é‡ã¨å†…è”µ800GbEãƒãƒƒãƒˆãƒ¯ãƒ¼ã‚¯ã‚’ç‰¹å¾´ã¨ã—ã€ãƒ¬ãƒ³ãƒ€ãƒªãƒ³ã‚°ã€ç§‘å¦è¨ˆç®—ã€ã‚²ãƒ¼ãƒŸãƒ³ã‚°åˆ†é‡Žã«é©æ–°ã‚’ã‚‚ãŸã‚‰ã™ã€‚é–‹ç™ºè€…ã‚ãƒƒãƒˆã¯2025å¹´å¾ŒåŠã€é‡ç”£ã¯2026å¹´å¾ŒåŠã‚’äºˆå®šã—ã¦ã„ã‚‹ã€‚ é©æ–°çš„ã‚¢ãƒ¼ã‚ãƒ†ã‚¯ãƒãƒ£ã¨åœ§å€’çš„ãªæ€§èƒ½ (Credit: Bolt Graphics) ã‚«ãƒªãƒ•ã‚©ãƒ«ãƒ‹ã‚¢å·žã‚µãƒ‹ãƒ¼ãƒ™ãƒ¼ãƒ«ã‚’æ‹ ç‚¹ã¨ã™ã‚‹Bolt Graphicsã¯ã€GPUã‚¢ãƒ¼ã‚ãƒ†ã‚¯ãƒãƒ£ã‚’ä¸€ã‹ã‚‰å†è¨è¨ˆã—ãŸã€ŒZeusã€ã‚’ç™ºè¡¨ã—ãŸã€‚æ³¨ç›®ã™ã¹ãã¯ã€æ˜¨ä»Šã®ãƒˆãƒ¬ãƒ³ãƒ‰ã¨ã¯ç•°ãªã‚ŠAIã«ã¯ä¸€åˆ‡è¨€åŠã›ãšã€é«˜å“è³ªãƒ¬ãƒ³ãƒ€ãƒªãƒ³ã‚°ã€HPCï¼ˆé«˜æ€§èƒ½ã‚³ãƒ³ãƒ”ãƒ¥ãƒ¼ãƒ†ã‚£ãƒ³ã‚°ï¼‰ã€ã‚²ãƒ¼ãƒŸãƒ³ã‚°ã«ç‰¹åŒ–
rydot 2025/03/10
ãƒªãƒ³ã‚¯
æœ€é©åŒ–è¶…å…¥é–€
ã“ã®ã€Œæœ€é©åŒ–è¶…å…¥é–€ã€ã¯ã€ã€Žæœ€é©åŒ–ã—ãŸã„ï¼ã€ã¨æ€ã£ãŸæ™‚ã«ã€æœ€åˆã«å‚è€ƒã«ãªã‚Šãã†ãªã‚‚ã®ã‚’ã¤ã‚‰ã¤ã‚‰ã¨èªžã‚Šã¾ã—ãŸã€‚
rydot 2025/03/08
ãƒªãƒ³ã‚¯
C++ ã‚ªãƒ–ã‚¸ã‚§ã‚¯ãƒˆã®ã‚«ã‚¹ã‚¿ãƒ ãƒ“ãƒ¥ãƒ¼ã‚’ä½œæˆã™ã‚‹ - Visual Studio (Windows)
Visual Studio ã® Natvis ãƒ•ãƒ¬ãƒ¼ãƒ ãƒ¯ãƒ¼ã‚¯ã‚’ä½¿ç”¨ã™ã‚‹ã¨ã€[ãƒãƒ¼ã‚«ãƒ«] ã‚¦ã‚£ãƒ³ãƒ‰ã‚¦ã€[ã‚¦ã‚©ãƒƒãƒ] ã‚¦ã‚£ãƒ³ãƒ‰ã‚¦ã€ãƒ‡ãƒ¼ã‚¿ãƒ’ãƒ³ãƒˆãªã©ã€ãƒ‡ãƒãƒƒã‚¬ãƒ¼ã®å¤‰æ•°ã‚¦ã‚£ãƒ³ãƒ‰ã‚¦ã§ã®ãƒã‚¤ãƒ†ã‚£ãƒ–åž‹ã®è¡¨ç¤ºæ–¹æ³•ã‚’ã‚«ã‚¹ã‚¿ãƒžã‚¤ã‚ºã§ãã¾ã™ã€‚ Natvis ã®è¦–è¦šåŒ–ã¯ã€ä½œæˆã—ãŸåž‹ã‚’ãƒ‡ãƒãƒƒã‚°ä¸ã«ã‚ˆã‚Šè¦‹ã‚„ã™ãã™ã‚‹ã®ã«å½¹ç«‹ã¡ã¾ã™ã€‚ Natvis ã¯ã€ä»¥å‰ã®ãƒãƒ¼ã‚¸ãƒ§ãƒ³ã® Visual Studio ã§ä½¿ç”¨ã•ã‚Œã¦ã„ãŸ autoexp.dat ãƒ•ã‚¡ã‚¤ãƒ«ã‚’ã€XML æ§‹æ–‡ã€ã‚ˆã‚Šé«˜åº¦ãªè¨ºæ–ã€ãƒãƒ¼ã‚¸ãƒ§ãƒ³ç®¡ç†ã€è¤‡æ•°ãƒ•ã‚¡ã‚¤ãƒ«ã®ã‚µãƒãƒ¼ãƒˆã§ç½®ãæ›ãˆã‚‹ã‚‚ã®ã§ã™ã€‚ Natvis ã®è¦–è¦šåŒ– Natvis ãƒ•ãƒ¬ãƒ¼ãƒ ãƒ¯ãƒ¼ã‚¯ã‚’ä½¿ç”¨ã—ã¦ã€è‡ªåˆ†ã§ä½œæˆã—ãŸåž‹ã®è¦–è¦šåŒ–ãƒ«ãƒ¼ãƒ«ã‚’ä½œæˆã™ã‚‹ã¨ã€é–‹ç™ºè€…ãŒãƒ‡ãƒãƒƒã‚°ä¸ã«åž‹ã‚’ç¢ºèªã—ã‚„ã™ããªã‚Šã¾ã™ã€‚ ãŸã¨ãˆã°ã€æ¬¡ã®å›³ã¯ã€ã‚«ã‚¹ã‚¿ãƒ è¦–è¦šã‚¨ãƒ•ã‚§ã‚¯ãƒˆã‚’é©ç”¨ã—ã¦ã„ãªã„ãƒ‡ãƒãƒƒã‚¬ãƒ¼ ã‚¦ã‚£ãƒ³ãƒ‰ã‚¦ã«è¡¨ç¤ºã•ã‚ŒãŸ Windows::UI::XAM
rydot 2025/03/05
ãƒªãƒ³ã‚¯
GitHub - NVIDIA-RTX/RTXNTC: NVIDIA Neural Texture Compression SDK
rydot 2025/03/02
ãƒªãƒ³ã‚¯
Day 1: RCUã¨ã¯ä½•ã‹
ã‚¤ãƒ³ãƒˆãƒãƒ€ã‚¯ã‚·ãƒ§ãƒ³ ã“ã‚Œã¯sanoãŒä¸€äººã§RCU(Read Copy Update)ã«é–¢ã™ã‚‹è©±ã‚’ã¾ã¨ã‚ã¦Rustã§RCUã‚’å®Ÿè£…ã‚’ã—ã¦ã„ãã‚¢ãƒ‰ãƒ™ãƒ³ãƒˆã‚«ãƒ¬ãƒ³ãƒ€ãƒ¼ã§ã‚ã‚‹ã€‚ 25æ—¥é–“ã®è¨˜äº‹ã‚’RCUã®è©±é¡Œã ã‘ã§æ›¸ãåˆ‡ã‚Œã‚‹ã‹ä¸å®‰ã ãŒã€ä¸€ã¤ä¸€ã¤ã®è¨˜äº‹ã‚’èªè€…ãŒã€Œé¢ç™½ã„ã€ã€Œå‹‰å¼·ã«ãªã‚‹ã€ã€Œã‚ã‹ã‚Šã‚„ã™ã„ã€ã¨æ€ãˆã‚‹ã‚ˆã†ãªã‚‚ã®ã«ã—ã¦ã„ã“ã†ã¨æ€ã†ã€‚ ä»¥ä¸‹ã®ã‚ˆã†ãªæµã‚Œã§RCUã«ã¤ã„ã¦æ·±ãæŽ˜ã‚Šä¸‹ã’ã¦ã„ãï¼š ã¾ãšRCUã¨ã¯ä½•ã‹ã¨ã„ã†åŸºæœ¬çš„ãªèª¬æ˜Žã‹ã‚‰å…¥ã‚Šã€æ¬¡ã«å®Ÿè£…ã®åœŸå°ã¨ãªã‚‹ä¸¦è¡Œå‡¦ç†ã®çŸ¥è˜ã‚’è§£èª¬ã™ã‚‹ã€‚ ãã®å¾Œã€ã“ã‚Œã‚‰ã®çŸ¥è˜ã‚’çµ„ã¿åˆã‚ã›ã¦RCUã®å‹•ä½œåŽŸç†ã‚’ç†è§£ã—ã€æœ€å¾Œã«Rustã«ã‚ˆã‚‹å®Ÿè£…ã«å–ã‚Šçµ„ã‚“ã§ã„ãã€‚ 12æœˆ1æ—¥ã®æ™‚ç‚¹ã§sanoã¯RCUã®å®Ÿè£…ã‚’çµ‚ãˆã¦ã„ã‚‹ã®ã§èˆˆå‘³ãŒã‚ã‚Œã°å®Ÿè£…ã‚’å…ˆå–ã‚Šã—ã¦èªã‚“ã§ã»ã—ã„ã€‚ å®Œæˆã—ãŸRCUã®å®Ÿè£…: ã¾ãŸã€ä½•ã‹è³ªå•ã‚„æŒ‡æ‘˜ãŒã‚ã£ãŸã‚‰Xã«ã¦æŒ‡æ‘˜ã‚’ã—ã¦ã„ãŸã ã‘ã‚Œã°ã€åŠ ç†ä¿®æ£ã‚’ã™ã‚‹ã¤ã‚‚ã‚Šãªã®ã§ã©ã†ãžã‚ˆã‚ã—ã
rydot 2025/03/01
ãƒªãƒ³ã‚¯
Massively Parallel Simulations of Binary Black Hole Intermediate-Mass-Ratio Inspirals
rydot 2025/02/28
ãƒªãƒ³ã‚¯
Cooperative Groups: Flexible CUDA Thread Programming | NVIDIA Technical Blog
InÂ efficient parallel algorithms, threads cooperate and share data to perform collective computations. To share data, the threads must synchronize. The granularity of sharing varies from algorithm to algorithm, so thread synchronization should be flexible. Making synchronization an explicit part of the program ensures safety, maintainability, and modularity. CUDA 9 introduces Cooperative Groups, w
rydot 2025/02/28
ãƒªãƒ³ã‚¯
An efficient GPU-based h-adaptation framework via linear trees for the flux reconstruction method
rydot 2025/02/28
ãƒªãƒ³ã‚¯
CUDA: QuadTree & Octree, p.1
rydot 2025/02/28
ãƒªãƒ³ã‚¯
Chapter 37. Octree Textures on the GPU
rydot 2025/02/28
ãƒªãƒ³ã‚¯
Cista++ | Homepage
C++ Serialization & Reflection Cista++ is a simple, open source (MIT license) C++17 compatible way of (de-)serializing C++ data structures. Single header. No mac ros. No source code generation. Raw performance - use your native structs. Supports modification/resizing of deserialized data! Supports complex and cyclic data structures including cyclic references, recursive data structures, etc. Save 5
rydot 2025/02/28
ãƒªãƒ³ã‚¯
GitHub - niXman/yas: Yet Another Serialization
rydot 2025/02/28
ãƒªãƒ³ã‚¯
Surface Simplification using Intrinsic Error Metrics
rydot 2025/02/26
ãƒªãƒ³ã‚¯
Diverse Counterfactual Explanations (DiCE) for ML â€” DiCE 0.11 documentation
rydot 2025/02/26
ãƒªãƒ³ã‚¯
Keenan Crane - Discrete Differential Geometry
rydot 2025/02/26
ãƒªãƒ³ã‚¯
GitHub - pola-rs/polars: Dataframes powered by a multithreaded, vectorized query engine, written in Rust
>>> import polars as pl >>> df = pl.DataFrame( ... { ... "A": [1, 2, 3, 4, 5], ... "fruits": ["banana", "banana", "apple", "apple", "banana"], ... "B": [5, 4, 3, 2, 1], ... "cars": ["beetle", "audi", "beetle", "beetle", "beetle"], ... } ... ) # embarrassingly parallel execution & very expressive query language >>> df.sort("fruits").select( ... "fruits", ... "cars", ... pl.lit("fruits").alias("lite
rydot 2025/02/25
ãƒªãƒ³ã‚¯
SmartCore
rydot 2025/02/25
ãƒªãƒ³ã‚¯
Torch.manual_seed(3407) is all you need: On the influence of random seeds in deep learning architectures for computer vision
In this paper I investigate the effect of random seed selection on the accuracy when using popular deep learning architectures for computer vision. I scan a large amount of seeds (up to $10^4$) on CIFAR 10 and I also scan fewer seeds on Imagenet using pre-trained models to investigate large scale datasets. The conclusions are that even if the variance is not very large, it is surprisingly easy to
rydot 2025/02/24
ãƒªãƒ³ã‚¯
æ•°å¦è€…ãŒæ›¸ã„ãŸæ·±å±¤å¦ç¿’è¬›ç¾©ã€€æ•°å¦ã¨éŒ¯è¦–ã®ç§‘å¦é¤¨é›»åå‡ºç‰ˆéƒ¨
å¤§å¦1å¹´ã§å¦ã¶å¾®åˆ†ã¨è¡Œåˆ—ã‚’ä½¿ã£ã¦ï¼Œç‰¹ã«ç•³ã¿è¾¼ã¿ãƒ‹ãƒ¥ãƒ¼ãƒ©ãƒ«ãƒãƒƒãƒˆãƒ¯ãƒ¼ã‚¯ã®æ·±å±¤å¦ç¿’ã®åŸºç¤Žäº‹é …ã‚’æ•°å¦çš„åŽ³å¯†æ€§ã‚’é‡è¦–ã—ã¦è§£èª¬ã—ã¾ã™ï¼Žæ•°å¦ã¯ã©ã®ã‚ˆã†ã«ãªã£ã¦ã„ã‚‹ã®ã ã‚ã†ã‹ï¼Œã¨ã„ã†ç–‘å•ã‚’æŒãŸã‚ŒãŸå ´åˆã«å‚è€ƒã«ãªã‚‹ã“ã¨ã‚’ç›®æŒ‡ã—ã¦ã„ã¾ã™ï¼Žé©å®œï¼Œæ”¹è¨‚ç‰ˆãƒ»å¢—è£œç‰ˆã‚’ç™ºä¿¡ã—ã¦ã„ãäºˆå®šã§ã™ã€‚æ›´æ–°ã¯ã“ã®ã‚µã‚¤ãƒˆã«æŽ²è¼‰ã—ã¦ã„ãã¾ã™ã€‚ å…¬é–‹æ—¥ 2025/1/6. Ver. 1.2 (2025/1/7) typo. Ver. 1.3 (2025/1/8ï¼‰typo. Ver. 1.4 (2025/1/10) æ–‡ã®å¾®å°ä¿®æ£ æœ€æ–°ç‰ˆã€€Appendix Ver. 1.3ï¼ˆ2025/1/12ï¼‰ã€€æ–‡ã®è¿½åŠ ãƒ–ãƒ©ã‚¦ã‚¶ãƒ¼ã«ã‚ˆã£ã¦ã¯ pdf ãŒæ–‡å—åŒ–ã‘ç‰ã™ã‚‹å ´åˆãŒã‚ã‚Šã¾ã™ã€‚ãã®å ´åˆã¯ pdf ã‚’ãƒ€ã‚¦ãƒ³ãƒãƒ¼ãƒ‰ã—ã¦ã”è¦§ãã ã•ã„ã€‚ å¼•ç”¨ã®éš›ã¯ æ–°äº•ä»ä¹‹ï¼Œ æ•°å¦è€…ãŒæ›¸ã„ãŸæ·±å±¤å¦ç¿’è¬›ç¾©ï¼Œç•³ã¿è¾¼ã¿ãƒ‹ãƒ¥ãƒ¼ãƒ©ãƒ«ãƒãƒƒãƒˆç¯‡ï¼ŒMATLABãƒ—ãƒã‚°ãƒ©ãƒ ä»˜ãï¼Œ2025.
rydot 2025/02/24
ãƒªãƒ³ã‚¯
A Survey on Knowledge Distillation of Large Language Models
rydot 2025/02/21
ãƒªãƒ³ã‚¯
1 2 3 4 5 6 7 8 9 10 æ¬¡ã®ãƒšãƒ¼ã‚¸