[新聞] IBM-AMD 合作量子超級電腦瞄準藥物發現

看板Stock (股票)作者 (潤潤潤)時間3小時前 (2025/08/26 21:29), 3小時前編輯推噓7(7010)
留言17則, 9人參與, 1小時前最新討論串1/1
原文標題: IBM-AMD Alliance Targets Drug Discovery, Climate Modeling With Quantum Supercomputers 原文連結:https://tinyurl.com/3u7643zt 發布時間:Tue, August 26, 2025 at 7:53 PM GMT+8 記者署名:Anusuya Lahiri 原文內容: International Business Machines Corp. (NYSE:IBM) and Advanced Micro Devices Inc. (NASDAQ:AMD) are partnering to advance quantum-centric supercomputing, combining IBM’s quantum hardware and software expertise with AMD’s strength in high-performance computing and AI acceleration. 國際商業機器公司 (NYSE:IBM) 與超微半導體公司 (NASDAQ:AMD) 正在合作推進以量子為 核心的超級運算,結合 IBM 在量子硬體與軟體的專長,以及 AMD 在高效能運算與人工智 慧加速的優勢。 The collaboration, announced Tuesday, is aimed at building scalable, open- source platforms that integrate quantum and classical systems to solve problems beyond the reach of traditional computing. Target applications range from drug discovery and climate modeling to advanced optimization. 這項於週二宣布的合作,旨在建立可擴展的開源平台,整合量子與傳統系統,以 解決傳統運算無法觸及的問題。應用範圍包括藥物研發、氣候建模與先進優化。 As part of the effort, AMD’s CPUs, GPUs, and FPGAs will be integrated with IBM’s quantum systems, supporting IBM’s push to deliver fault-tolerant quantum computers before the end of the decade. The companies intend to demonstrate hybrid quantum–classical workflows later this year, a step toward bringing commercially viable applications into view. 在這項努力中,AMD 的 CPU、GPU 與 FPGA 將與 IBM 的量子系統整合,支持 IBM 在本十 年結束前交付具備容錯能力的量子電腦的推進。雙方預計於今年稍晚展示量子–經典混合 工作流程,邁出商業應用的第一步。 The move comes as momentum in quantum computing accelerates globally. According to McKinsey, the quantum technology market, spanning computing, communications, and sensing, is projected to grow from $4 billion in 2024 to as much as $97 billion by 2035, with quantum computing alone potentially accounting for $72 billion. 隨著全球量子運算的發展加速,這一舉措應運而生。據麥肯錫預測,量子科技市場(涵蓋 計算、通訊與感測)將從 2024 年的 40 億美元,增長至 2035 年的 970 億美元,其中 光是量子運算就可能佔 720 億美元。 By 2040, McKinsey estimates the sector could nearly double again to $198 billion as breakthroughs in error correction, qubit scalability, and secure communications open new frontiers in finance, defense, and life sciences. Governments worldwide have already pledged more than $10 billion in funding for 2025, underscoring the strategic importance of the field. 到 2040 年,麥肯錫估計該產業可能幾乎再翻倍,達到 1,980 億美元,隨著錯誤修正、 量子位擴展性與安全通訊的突破,將開啟金融、國防與生命科學的新前沿。全球各國政府 已承諾 2025 年投入超過 100 億美元資金,凸顯該領域的戰略重要性。 Competition in the space is intensifying. Alphabet’s (NASDAQ:GOOGL) (NASDAQ: GOOG) Google Quantum AI recently unveiled its Willow chip, a superconducting system that demonstrated exponential error reduction as qubit counts scale. 該領域競爭日益激烈。Alphabet (NASDAQ:GOOGL)(NASDAQ:GOOG) 的 Google Quantum AI 最近發布了其 Willow 晶片,這是一個超導系統,隨著量子位數增加展現出指數級的錯誤 降低。 Willow completed a benchmark in under five minutes that would take an estimated ten septillion years for a conventional supercomputer, marking a significant step in real-time error correction. Willow 在不到五分鐘內完成了一個基準測試,而傳統超級電腦估計需要 10 的 24 次方 年(十兆兆兆年),標誌著即時錯誤修正的一大突破。 Nvidia Corp. (NASDAQ:NVDA) has also deepened its presence in quantum through a partnership with Google Quantum AI to accelerate device development and simulation, a collaboration that Wedbush analysts have flagged as a long-term growth driver alongside Nvidia’s dominance in AI semiconductors. Nvidia (NASDAQ:NVDA) 也透過與 Google Quantum AI 的合作加深在量子領域的布局,協 助加速裝置開發與模擬。Wedbush 分析師指出,這一合作與 Nvidia 在 AI 半導體的主導 地位並列,將成為長期增長的驅動力。 The IBM–AMD partnership adds another layer to the evolving ecosystem, where established players and startups alike are racing to make quantum systems practical at scale. IBM–AMD 的合作為正在演變的生態系統再添一層,老牌企業與新創公司正競相推動量子 系統的大規模實用化。 心得/評論: https://i.imgur.com/6xVaXte.jpeg
技術上 IBM 需要 AMD 的 CPU/GPU/FPGA 來打造混合運算架構 商業上 避免落入 Nvidia–Google 聯盟的生態圈 政策上 AMD 與政府超級電腦計畫有深厚關係,有助於 IBM 拿下公共資金與科研應用 -- ※ 發信站: 批踢踢實業坊(ptt.cc), 來自: 150.117.197.145 (臺灣) ※ 文章網址: https://www.ptt.cc/bbs/Stock/M.1756214973.A.CE3.html ※ 編輯: TyuzuChou (150.117.197.145 臺灣), 08/26/2025 21:30:10

08/26 21:33, 3小時前 , 1F
AMD夜盤本來跌的 原來噴這個?
08/26 21:33, 1F

08/26 21:45, 2小時前 , 2F
樓下教主
08/26 21:45, 2F

08/26 21:48, 2小時前 , 3F
我不是,我只會東南亞糞島
08/26 21:48, 3F

08/26 21:55, 2小時前 , 4F
蒙R波
08/26 21:55, 4F

08/26 22:04, 2小時前 , 5F
丸子,之前跟IBM合作的都…
08/26 22:04, 5F

08/26 22:05, 2小時前 , 6F
完了..
08/26 22:05, 6F

08/26 22:37, 2小時前 , 7F
丸子…
08/26 22:37, 7F

08/26 22:50, 1小時前 , 8F
不就IBM和google吹太兇,所以只能找蘇嬤和老黃合作
08/26 22:50, 8F

08/26 22:50, 1小時前 , 9F
, 目前量子要實用化至少要2030,在此之前只能靠數
08/26 22:50, 9F

08/26 22:50, 1小時前 , 10F
位孿生
08/26 22:50, 10F

08/26 22:52, 1小時前 , 11F
用GPGPU模擬量子運算是現階段最可靠方案
08/26 22:52, 11F

08/26 23:00, 1小時前 , 12F
就算量子電腦實用化,那其實應用也沒比ASIC廣多少
08/26 23:00, 12F

08/26 23:01, 1小時前 , 13F
只有特定密碼學領域可以把指數複雜度變多項式時間
08/26 23:01, 13F

08/26 23:03, 1小時前 , 14F
還有一些像是最佳化問題、線性代數求解
08/26 23:03, 14F

08/26 23:04, 1小時前 , 15F
可以讓時間複雜度降低,或是直接模擬量子系統
08/26 23:04, 15F

08/26 23:04, 1小時前 , 16F
所以像是化學藥物發現就很有用
08/26 23:04, 16F

08/26 23:04, 1小時前 , 17F
但是並非像是很多人想像的直接幹爆現有的超級電腦
08/26 23:04, 17F
文章代碼(AID): #1ehRQzpZ (Stock)
文章代碼(AID): #1ehRQzpZ (Stock)