Micron Technology’s Post

Alphawave Semi announces a robust 9.2 Gbps HBM3E sub-system (PHY + Controller IP) silicon platform. This #chiplet-enabled platform, based on Micron Technology HBM3E, offers best-in-class performance and exceptional power efficiency. Based on the company’s proven HBM3E IP, the platform takes chiplet-enabled memory bandwidth to new heights of 1.2 Terabytes per second (#TBps), addressing the demand for ultra-high-speed connectivity in high-performance compute (#HPC) and accelerated compute for generative artificial intelligence (#AI) applications. To find out more, visit: https://lnkd.in/eiwPisce #AlphawaveSemi #Chiplets #AI #Connectivity #CustomSilicon #HPC

𝗧𝗦𝗠𝗖 𝗦𝘁𝗮𝗰𝗸𝘀 𝘁𝗵𝗲 𝗖𝗵𝗶𝗽𝘀 𝘄𝗶𝘁𝗵 𝟯𝗗 𝗦𝘆𝘀𝘁𝗲𝗺-𝗼𝗻-𝗪𝗮𝗳𝗲𝗿 𝗧𝗲𝗰𝗵𝗻𝗼𝗹𝗼𝗴𝘆 TSMC’s 3D System-on-Wafer (SoW) technology is set to revolutionize high-performance computing by stacking logic and memory, pushing the limits of AI and HPC. 𝗪𝗵𝘆 𝗶𝘁 𝗺𝗮𝘁𝘁𝗲𝗿𝘀 ‣ TSMC’s CoW-SoW platform combines wafer-scale logic with stacked high-bandwidth memory, promising massive gains in bandwidth density. ‣ Early adopters like Cerebras and Tesla have already realized performance-per-watt improvements using TSMC’s initial InFO-SoW technology. ‣ Production of CoW-SoW is expected by 2027, which could unlock exascale-class AI processing. 𝗥𝗲𝗮𝗱 𝘁𝗵𝗲 𝗳𝘂𝗹𝗹 𝗮𝗿𝘁𝗶𝗰𝗹𝗲: https://lnkd.in/ejtBZmAj #TSMC #CoW #SoW #3DChip #InFO #SoW

ANNOUNCING THE FASTEST #AI CHIP ON EARTH📣 Cerebras proudly announces CS-3: the fastest AI accelerator in the world. 🌍🌏🌎 The CS-3 can train up to 24 trillion parameter models on a single device. 🤯 The world has never seen #artificalintelligence at this scale. CS-3 specs: ⚙ 46,225 mm2 silicon | 4 trillion transistors | 5nm ⚙ 900,000 cores optimized for sparse linear algebra ⚙ 125 petaflops of AI compute ⚙ 44 gigabytes of on-chip memory ⚙ 1,200 gigabytes of external memory ⚙ 21 PByte/s memory bandwidth ⚙ 214 Pbit/s fabric bandwidth 📰 Press Release: https://lnkd.in/gC8KGQRN 👨🎓 Learn More: https://lnkd.in/gXZ-QBJn

Unlocking Performance: From Data Center GPU Max Series to Gaudi 2 AI Accelerators and Xeon Scalable Processors, Intel is reshaping the landscape of HPC and AI. Learn more: https://bit.ly/3RUKt5Z #AITechnology #Solutions

At the foundation of any #AI discussion is #Data. Where data exists at every point that it is reasoned over is a critical consideration when creating #infrastructure for AI. From #NAND through #HBM and the #DRAM that lives in between, this will be a fascinating discussion at Dell Technologies #DellTechWorld Monday morning. Arrive well caffeinated. What can be done about power efficiency as AI scales? What are the states of the art in volatile memory and persistent storage? What does the future hold? Samsung Semiconductor and I will explore these and other questions. The Futurum Group Wharton Executive Education | CTO Program Wharton Executive Education | Digital Transformation for Senior Executives Program

Check out this special report from Ann Mutschler of Semiconductor Engineering and explore the linkage between #AI accelerators and optimized interconnects. Arteris network-on-chip IP and SoC integration software are critical catalyts to tomorrow's innovations. https://bit.ly/3RbWi5Q. Learn more about Arteris products here: https://bit.ly/3GrK3x4 #semiconductor #innovation #SystemOnChip #NetworkOnChip #AI #ArtificialIntelligence #Arteris

#4_Trillion_Transistors : #Cerebras #WSE_3 : #5nm_TSMC/ : #256_ExaFLOPs Source : https://lnkd.in/gNMx-vd3 : 13/03/2024 : "#Cerebras Systems Unveils World’s Fastest AI Chip with Whopping 4 Trillion Transistors https://lnkd.in/gPMsSrQZ : Third Generation #5nm #Wafer_Scale_Engine_3 (#WSE_3) Powers Industry’s Most #Scalable #AI_Supercomputers, Up To #256_exaFLOPs via #2048_Nodes" "Cerebras Systems, the pioneer in accelerating generative AI, has #Doubled down on its existing #World_Record of #Fastest #AI_chip with the introduction of the #Wafer_Scale_Engine_3. The #WSE_3 delivers #Twice the #Performance of the previous record-holder,the #Cerebras #WSE_2, at the same power draw and for the same price. Purpose built for training the industry’s largest AI models, the #5nm-based, #4_Trillion_Transistor WSE-3 powers the Cerebras CS-3 AI supercomputer, delivering 125 petaflops of peak AI performance through #900000_AI_Optimized #Compute_Cores. Key Specs: 4 trillion transistors 900,000 AI cores 125 petaflops of peak AI performance #44GB on-chip #SRAM #5nm #TSMC process External memory: 1.5TB, 12TB, or 1.2 #PB #Trains AI models up to #24_Tillion_Parameters #Cluster_Size of up to #2048_CS_3 systems." [ #Groq LPU : https://lnkd.in/gwXAhx6P) : 20/02/2024 #500_token/ Sec. : Fastest #Inferene_Engine : https://groq.com/ #Groq LPU : A #Software_Defined #Tensor_Streaming_Multiprocessor for #Large_Scale #Machine_Learning ] [ 13pgs : https://lnkd.in/gAZns7PU) : 11/11/2023 #China : #5_ExaFlop/s #40_Million_Core #Sunway_Supercomputer by : Rongfen Lin & others Paper : "5 ExaFlop/s HPL-MxP Benchmark with Linear Scalability on the 40-Million-Core Sunway Supercomputer" Source : https://lnkd.in/gP6s4PfH ] [ https://lnkd.in/gpT_y6Dm) : https://lnkd.in/gwe2cNrv #5Exaflops : #Meta #SUPERCOMPUTER Expected to Operate : #RSC ( AI #Research #Supercluster) will serve #Training #Data at #16terabytes / #Second ]  [ https://lnkd.in/gpT_y6Dm) : https://lnkd.in/gR5PxDPJ) : 18/01/2024 #META : #350000_Nvidia_H100s for #AI_Supercomputing  "We're building an absolutely massive amount of infrastructure to support this. By the end of this year, we're going to have around 350,000 Nvidia H100s, or 600,000 H100 equivalents of compute if you include other GPUs." Source : https://lnkd.in/g5nRxigG ] [ https://lnkd.in/dm4cs3fs) : 21/07/2023 #G42 and #Cerebras Unveil World’s #Largest #4_Exaflop #Supercomputer for #AI : #12_Exaflop by #Mid_2024 & #36_Exaflop #by #End_2024 : Source Video : https://lnkd.in/dxCkieJ3 " Launching today with its first of #9_interconnected #AI_Supercomputers, the #Condor_Galaxy system will reach a combined AI training capacity of #36_ExaFLOPs"] [ https://lnkd.in/dqpTXHGP) : : Jun 30, 2023 With #22000_NVIDIA_H100 : #LARGEST_SUPERCOMPUTER IN THE WORLD IS IN THE MAKING BY #NVIDIA & #InflectionAI : completed by late 2023 : "Microsoft-backed AI startup Inflection raises $1.3 billion from Nvidia and others"]

For compute-bound AI problems, such as CNNs, the Agilex™ 5 family provides the only FPGA fabric infused with AI. This family enhances our traditional DSP blocks with AI tensor capability which are embedded into the core fabric to vastly increase the amount of AI compute. Agilex™ 5 delivers up to 56 INT8 TOPS* of AI performance and is well-suited to the needs of new intelligent embedded edge solutions. Read our FPGAi blog to learn more: https://intel.ly/4dbBXYj #FPGAi #AcceleratingInnovators

I am really looking forward to moderating this panel of distinguished experts from Samsung Semiconductor at Dell Technologies big, big show. The Futurum Group Wharton Executive Education | CTO Program Wharton Executive Education | Digital Transformation for Senior Executives Program Let me know if you will be there.

Performance, Durability, and Reliability For Any Deployment. Supermicro, using Intel Corporation processors, builds intelligent edge solutions that deliver high-performance compute, unparalleled edge AI capabilities, and reliability needed to provide superior application performance in a broad range of environments. 🌐 https://serverfactory.com/ ✉️ [email protected] 📞 +44 (0)20 3432 5270 #ServerFactory #Supermicro #Intel #Edge #Innovation #AI