From aeba8ea5ba346a926ab26fa750f2805718ff5368 Mon Sep 17 00:00:00 2001 From: Vasudev Menon Date: Sat, 12 Oct 2024 23:48:28 +0530 Subject: [PATCH] add more posts --- ...en-9800X3D-Cinebench-scores-gets-leaked.md | 35 +++++++++++++++++++ ...ng-Chiplet-PHY-at-64Gbps-in-3nm-process.md | 27 ++++++++++++++ ...ect-Fanless-Cooling-System-Architecture.md | 27 ++++++++++++++ ...loud-providers-and-enterprise-customers.md | 34 ++++++++++++++++++ 4 files changed, 123 insertions(+) create mode 100644 source/_posts/AMD-Ryzen-9800X3D-Cinebench-scores-gets-leaked.md create mode 100644 source/_posts/Eliyan-Corp-delivers-highest-performing-Chiplet-PHY-at-64Gbps-in-3nm-process.md create mode 100644 source/_posts/HP-Enterprise-introduces-Direct-Fanless-Cooling-System-Architecture.md create mode 100644 source/_posts/Nvidia-Blackwell-has-1-year-of-waitlist-due-to-huge-demand-from-cloud-providers-and-enterprise-customers.md diff --git a/source/_posts/AMD-Ryzen-9800X3D-Cinebench-scores-gets-leaked.md b/source/_posts/AMD-Ryzen-9800X3D-Cinebench-scores-gets-leaked.md new file mode 100644 index 0000000..bfdd99b --- /dev/null +++ b/source/_posts/AMD-Ryzen-9800X3D-Cinebench-scores-gets-leaked.md @@ -0,0 +1,35 @@ +--- +title: AMD Ryzen 9800X3D Cinebench scores gets leaked +date: 2024-10-12 19:51:13 +tags: + - amd + - ryzen + - x3d + - 3d v-cache + - windows + - linux + - cinebench + - benchmark + - leaked + - zen 5 + - vertical cache + - chiplet design + - tsmc +--- + +### Quick Report + +Ryzen 9800 3D V-Cache is rumoured to be launching soon during Q4 2024 and uplift the bad underwhelming Ryzen 9000 desktop reception due to lackluster performance compared to 7000 3D V-Cache and X series. Since then AMD has been working with Microsoft to improve the Zen 5 and even backport branch prediction update that improved performance of Zen 3, Zen 4 and Zen 5. + + +According to leaked screenshot or a short video consisting of HWINFO64 clock info and Cinebench 2024 from Bilibili forums. It was posted on [X aka Twitter by HXL][def] + +### Source(s) + +- [TPU][def2] +- [Bilibili][def3] +- [X][def] + +[def]: https://x.com/9550pro/status/1844596653216690387 +[def2]: https://www.techpowerup.com/327573/amd-ryzen-7-9800x3d-zen-5-spied-in-cinebench-boosts-up-to-5-20-ghz-all-core +[def3]: https://www.bilibili.com/video/BV1QS2MYeEVs diff --git a/source/_posts/Eliyan-Corp-delivers-highest-performing-Chiplet-PHY-at-64Gbps-in-3nm-process.md b/source/_posts/Eliyan-Corp-delivers-highest-performing-Chiplet-PHY-at-64Gbps-in-3nm-process.md new file mode 100644 index 0000000..7127a97 --- /dev/null +++ b/source/_posts/Eliyan-Corp-delivers-highest-performing-Chiplet-PHY-at-64Gbps-in-3nm-process.md @@ -0,0 +1,27 @@ +--- +title: Eliyan Corp delivers highest performing Chiplet PHY at 64Gbps in 3nm process +date: 2024-10-12 23:03:53 +tags: + - eliyan + - chiplet + - performance + - hbm + - 3nm +--- + +### Quick Report + +Eliyan corp credited for the invention of semiconductor industry's highest performing and most power efficient chiplet PHY at 64Gbps in 3nm process named **NuLink 2.0 PHY**. The device achieved 64Gbps/bump the highest performing die-to-die PHY solution for multi-die architectures. It's compatible with UCIe standard with 2x bandwidth on standard as well as advanced packaging. + + +NuLink 2.0 supports Universal Memory Interconnect (UMI), a chiplet interconnect technology improving die to memory bandwidth efficiency by 2x. UMI also leverages bi-directional PHY which is currently in draft stage and will be finalized Open Compute Project (OCP) BoW 2.1. + +You can read more about NuLink 2.0 PHY [here][def]. + +### Source(s) + +- [TPU][def2] +- [NuLink 2.0][def] + +[def]: https://eliyan.com/eliyan-news/eliyan-delivers-industrys-highest-performing-chiplet-interconnect-phy-at-64gbps-in-3nm-process/ +[def2]: https://www.techpowerup.com/327599/eliyan-delivers-highest-performing-chiplet-interconnect-phy-at-64gbps-in-3nm-process diff --git a/source/_posts/HP-Enterprise-introduces-Direct-Fanless-Cooling-System-Architecture.md b/source/_posts/HP-Enterprise-introduces-Direct-Fanless-Cooling-System-Architecture.md new file mode 100644 index 0000000..6e963eb --- /dev/null +++ b/source/_posts/HP-Enterprise-introduces-Direct-Fanless-Cooling-System-Architecture.md @@ -0,0 +1,27 @@ +--- +title: HP Enterprise introduces Direct Fanless Cooling System Architecture +date: 2024-10-12 20:10:57 +tags: + - HP + - HPE + - HP Enterprise division + - AI + - direct liquid cooling + - liquid cooling + - direct fanless cooling +--- + +### Quick Report + +In a industry first move for Direct Fanless Liquid Cooling that improved the cooling efficiency by 90% and helped enhance the cost and energy efficiency for large scale AI deployments according to its **AI event**. + +This cooling helped to deliver 7 of Top 10 Green500 Supercomputer list. For more info you can check their [blog][def] + + +### Source(s) + +- [TPU][def2] +- [HPE][def] + +[def]: https://www.hpe.com/us/en/newsroom/press-release/2024/10/hpe-announces-industrys-first-100-fanless-direct-liquid-cooling-systems-architecture.html +[def2]: https://www.techpowerup.com/327586/hpe-announces-industrys-first-100-fanless-direct-liquid-cooling-systems-architecture diff --git a/source/_posts/Nvidia-Blackwell-has-1-year-of-waitlist-due-to-huge-demand-from-cloud-providers-and-enterprise-customers.md b/source/_posts/Nvidia-Blackwell-has-1-year-of-waitlist-due-to-huge-demand-from-cloud-providers-and-enterprise-customers.md new file mode 100644 index 0000000..69dbaf1 --- /dev/null +++ b/source/_posts/Nvidia-Blackwell-has-1-year-of-waitlist-due-to-huge-demand-from-cloud-providers-and-enterprise-customers.md @@ -0,0 +1,34 @@ +--- +title: >- + Nvidia Blackwell has 1 year of waitlist due to huge demand from cloud + providers and enterprise customers +date: 2024-10-12 21:01:05 +tags: + - nvidia + - blackwell + - ai + - cloud + - enterprise + - google + - meta + - microsoft + - aws + - oracle + - gpu + - waitlist + - demand +--- + +### Quick Report + +In a shocking news regarding Nvidia's Blackwell GPU B100, B200 and GB200 are sold out for the next 12 months. This is due to the huge demand from cloud providers and enterprise customers. This news was confirmed by an analyst named Joe Moore working at Morgan Stanley in a meeting with NVIDIA and its investors. Nvidia executives also confirmed the news and they are still getting orders from Google, Meta, Microsoft, Oracle, AWS and others in excess quantities of 100,000 to drive their AI cluster deployments. + + +The news states earlier generation called Hopper was ordered in 10's of thousands by Google, Meta, Microsoft, Oracle, AWS and others while Blackwell continues to break records both in market share and AI gen-on-gen improvements. +Nvidia is very pleased with this news and they are working hard to meet the demand working with TSMC to satisfy the demands ASAP. + +### Source(s) + +- [TPU][def] + +[def]: https://www.techpowerup.com/327588/nvidia-blackwell-gpus-are-sold-out-for-12-months-customers-ordering-in-100k-gpu-quantities