Aki_A
@aki1029.bsky.social
🚗 New Case Study: Sony Honda Mobility Inc. × Fixstars AIBooster
www.fixstars.com/en/whitepape...
See how SHM improved GPU efficiency in AD/ADAS development with AIBooster—visualizing utilization, spotting bottlenecks, and building a cost-conscious, high-performance culture.
www.fixstars.com/en/whitepape...
See how SHM improved GPU efficiency in AD/ADAS development with AIBooster—visualizing utilization, spotting bottlenecks, and building a cost-conscious, high-performance culture.
November 12, 2025 at 9:13 PM
🚗 New Case Study: Sony Honda Mobility Inc. × Fixstars AIBooster
www.fixstars.com/en/whitepape...
See how SHM improved GPU efficiency in AD/ADAS development with AIBooster—visualizing utilization, spotting bottlenecks, and building a cost-conscious, high-performance culture.
www.fixstars.com/en/whitepape...
See how SHM improved GPU efficiency in AD/ADAS development with AIBooster—visualizing utilization, spotting bottlenecks, and building a cost-conscious, high-performance culture.
Hello from TechEquity Ai Summit ✨
techequity-ai.org
We’re here in Silicon Valley. If you’re attending, come say hi 👋 We'd love to talk about AI acceleration and performance optimization.
techequity-ai.org
We’re here in Silicon Valley. If you’re attending, come say hi 👋 We'd love to talk about AI acceleration and performance optimization.
November 7, 2025 at 5:26 PM
Hello from TechEquity Ai Summit ✨
techequity-ai.org
We’re here in Silicon Valley. If you’re attending, come say hi 👋 We'd love to talk about AI acceleration and performance optimization.
techequity-ai.org
We’re here in Silicon Valley. If you’re attending, come say hi 👋 We'd love to talk about AI acceleration and performance optimization.
We’ll be at the TechEquity Ai Summit in Silicon Valley, November 7–8, 2025.
techequity-ai.org
We’ll be hosting a booth, so if you’ll be attending, we’d love for you to stop by and connect with us 👋
techequity-ai.org
We’ll be hosting a booth, so if you’ll be attending, we’d love for you to stop by and connect with us 👋
November 6, 2025 at 11:00 PM
We’ll be at the TechEquity Ai Summit in Silicon Valley, November 7–8, 2025.
techequity-ai.org
We’ll be hosting a booth, so if you’ll be attending, we’d love for you to stop by and connect with us 👋
techequity-ai.org
We’ll be hosting a booth, so if you’ll be attending, we’d love for you to stop by and connect with us 👋
Fixstars AIBooster just received the AI/Generative AI Excellence Prize at EdgeTech+ AWARD 2025 🎉
www.fixstars.com/en/news/882
Proud of our team pushing performance-engineering forward so organizations can train and deploy AI more efficiently.
www.fixstars.com/en/news/882
Proud of our team pushing performance-engineering forward so organizations can train and deploy AI more efficiently.
November 6, 2025 at 5:00 PM
Fixstars AIBooster just received the AI/Generative AI Excellence Prize at EdgeTech+ AWARD 2025 🎉
www.fixstars.com/en/news/882
Proud of our team pushing performance-engineering forward so organizations can train and deploy AI more efficiently.
www.fixstars.com/en/news/882
Proud of our team pushing performance-engineering forward so organizations can train and deploy AI more efficiently.
🔬 We fine-tuned Llama 4 Scout with LoRA on an 8× H100 server using LLaMA-Factory — and achieved 2.7× faster training just by tuning batch size.
👉 Read more: blog.us.fixstars.com/llama-4-scou...
#Llama4 #AI #LoRA #GPUOptimization
👉 Read more: blog.us.fixstars.com/llama-4-scou...
#Llama4 #AI #LoRA #GPUOptimization
Llama 4 Scout Fine-tuning and Performance Engineering - Fixstars Corporation Tech Blog
Fine-tuning Llama 4 Scout using LLaMA-Factory and DeepSpeed, and implementing speedups and GPU optimization through batch size adjustments. We will explain the procedure in detail.
blog.us.fixstars.com
November 5, 2025 at 6:19 PM
🔬 We fine-tuned Llama 4 Scout with LoRA on an 8× H100 server using LLaMA-Factory — and achieved 2.7× faster training just by tuning batch size.
👉 Read more: blog.us.fixstars.com/llama-4-scou...
#Llama4 #AI #LoRA #GPUOptimization
👉 Read more: blog.us.fixstars.com/llama-4-scou...
#Llama4 #AI #LoRA #GPUOptimization
🧠 “NVIDIA H200: AI Acceleration and Performance Engineering in Practice”
www.fixstars.com/en/whitepape...
Fixstars explores how to turn that extra memory into measurable speedups using real benchmarks and optimization methods.
www.fixstars.com/en/whitepape...
Fixstars explores how to turn that extra memory into measurable speedups using real benchmarks and optimization methods.
November 3, 2025 at 9:45 PM
🧠 “NVIDIA H200: AI Acceleration and Performance Engineering in Practice”
www.fixstars.com/en/whitepape...
Fixstars explores how to turn that extra memory into measurable speedups using real benchmarks and optimization methods.
www.fixstars.com/en/whitepape...
Fixstars explores how to turn that extra memory into measurable speedups using real benchmarks and optimization methods.
Fixstars Amplify launches “Amplify AE Enterprise” — an on-premises solution for its GPU-based optimization solver, delivering high-speed and secure computing within your own infrastructure.
www.fixstars.com/en/news/880
www.fixstars.com/en/news/880
October 30, 2025 at 8:26 PM
Fixstars Amplify launches “Amplify AE Enterprise” — an on-premises solution for its GPU-based optimization solver, delivering high-speed and secure computing within your own infrastructure.
www.fixstars.com/en/news/880
www.fixstars.com/en/news/880
⚙️ New blog: Inside NVIDIA Blackwell Architecture – A CTO’s Guide to Next-Gen AI Infrastructure
blog.us.fixstars.com/nvidia-black...
NVIDIA’s Blackwell isn’t just faster — it redefines how we design AI infrastructure.
CTOs: rethink your TCO, power, and cooling strategies now.
blog.us.fixstars.com/nvidia-black...
NVIDIA’s Blackwell isn’t just faster — it redefines how we design AI infrastructure.
CTOs: rethink your TCO, power, and cooling strategies now.
October 29, 2025 at 1:08 AM
⚙️ New blog: Inside NVIDIA Blackwell Architecture – A CTO’s Guide to Next-Gen AI Infrastructure
blog.us.fixstars.com/nvidia-black...
NVIDIA’s Blackwell isn’t just faster — it redefines how we design AI infrastructure.
CTOs: rethink your TCO, power, and cooling strategies now.
blog.us.fixstars.com/nvidia-black...
NVIDIA’s Blackwell isn’t just faster — it redefines how we design AI infrastructure.
CTOs: rethink your TCO, power, and cooling strategies now.
🌟 Speed Up Your AI Deployment
www.fixstars.com/en/news/877
Join Fixstars’ webinar to learn how to boost GPU/CPU efficiency and cut inference costs by up to 50% with AIBooster.
We’ll also introduce the latest update, featuring autonomous optimization for Edge AI inference on NVIDIA GPUs.
www.fixstars.com/en/news/877
Join Fixstars’ webinar to learn how to boost GPU/CPU efficiency and cut inference costs by up to 50% with AIBooster.
We’ll also introduce the latest update, featuring autonomous optimization for Edge AI inference on NVIDIA GPUs.
October 27, 2025 at 8:46 PM
🌟 Speed Up Your AI Deployment
www.fixstars.com/en/news/877
Join Fixstars’ webinar to learn how to boost GPU/CPU efficiency and cut inference costs by up to 50% with AIBooster.
We’ll also introduce the latest update, featuring autonomous optimization for Edge AI inference on NVIDIA GPUs.
www.fixstars.com/en/news/877
Join Fixstars’ webinar to learn how to boost GPU/CPU efficiency and cut inference costs by up to 50% with AIBooster.
We’ll also introduce the latest update, featuring autonomous optimization for Edge AI inference on NVIDIA GPUs.
🚀 Fixstars Amplify AE v1.0 is now live!
www.fixstars.com/en/news/879
The latest update to our GPU-based Ising machine delivers faster, easier, and more scalable solutions for real-world combinatorial optimization.
Amplify AE v1.0 helps unlock practical optimization at scale.
www.fixstars.com/en/news/879
The latest update to our GPU-based Ising machine delivers faster, easier, and more scalable solutions for real-world combinatorial optimization.
Amplify AE v1.0 helps unlock practical optimization at scale.
October 23, 2025 at 8:52 PM
🚀 Fixstars Amplify AE v1.0 is now live!
www.fixstars.com/en/news/879
The latest update to our GPU-based Ising machine delivers faster, easier, and more scalable solutions for real-world combinatorial optimization.
Amplify AE v1.0 helps unlock practical optimization at scale.
www.fixstars.com/en/news/879
The latest update to our GPU-based Ising machine delivers faster, easier, and more scalable solutions for real-world combinatorial optimization.
Amplify AE v1.0 helps unlock practical optimization at scale.
What are Unikernels and why do they matter? blog.us.fixstars.com/current-stat...
Our blog explores their potential in cloud & embedded systems, surveys MirageOS, Unikraft, UKL, and shares hands-on trials.
#Unikernels #CloudComputing #EmbeddedSystems #Fixstars
Our blog explores their potential in cloud & embedded systems, surveys MirageOS, Unikraft, UKL, and shares hands-on trials.
#Unikernels #CloudComputing #EmbeddedSystems #Fixstars
Current State Survey of Unikernels - Fixstars Corporation Tech Blog
* This blog post is an English translation of an article originally published in Japanese on March 12, 2020. * The content of this article is the result of an approximately one-month internship in…
blog.us.fixstars.com
October 22, 2025 at 5:19 PM
What are Unikernels and why do they matter? blog.us.fixstars.com/current-stat...
Our blog explores their potential in cloud & embedded systems, surveys MirageOS, Unikraft, UKL, and shares hands-on trials.
#Unikernels #CloudComputing #EmbeddedSystems #Fixstars
Our blog explores their potential in cloud & embedded systems, surveys MirageOS, Unikraft, UKL, and shares hands-on trials.
#Unikernels #CloudComputing #EmbeddedSystems #Fixstars
🚗 Case Study: Sony Honda Mobility x Fixstars AIBooster
www.fixstars.com/en/cases/shm...
Sony Honda Mobility adopted Fixstars AIBooster to visualize GPU utilization and optimize AI training performance for their AFEELA brand. 💡 Learn how AIBooster became the first step in SHM’s AI improvement cycle.
www.fixstars.com/en/cases/shm...
Sony Honda Mobility adopted Fixstars AIBooster to visualize GPU utilization and optimize AI training performance for their AFEELA brand. 💡 Learn how AIBooster became the first step in SHM’s AI improvement cycle.
October 20, 2025 at 6:26 PM
🚗 Case Study: Sony Honda Mobility x Fixstars AIBooster
www.fixstars.com/en/cases/shm...
Sony Honda Mobility adopted Fixstars AIBooster to visualize GPU utilization and optimize AI training performance for their AFEELA brand. 💡 Learn how AIBooster became the first step in SHM’s AI improvement cycle.
www.fixstars.com/en/cases/shm...
Sony Honda Mobility adopted Fixstars AIBooster to visualize GPU utilization and optimize AI training performance for their AFEELA brand. 💡 Learn how AIBooster became the first step in SHM’s AI improvement cycle.
Join Fixstars’ webinar to learn practical ways to boost GPU/CPU efficiency and cut inference costs by 50%.
www.fixstars.com/en/news/877
We’ll also share updates from the latest AIBooster release — now featuring Autonomous Optimization for Edge AI and AI training.
🗓️ Oct 30 | 12:00 PM PDT | Free
www.fixstars.com/en/news/877
We’ll also share updates from the latest AIBooster release — now featuring Autonomous Optimization for Edge AI and AI training.
🗓️ Oct 30 | 12:00 PM PDT | Free
October 16, 2025 at 7:25 PM
Join Fixstars’ webinar to learn practical ways to boost GPU/CPU efficiency and cut inference costs by 50%.
www.fixstars.com/en/news/877
We’ll also share updates from the latest AIBooster release — now featuring Autonomous Optimization for Edge AI and AI training.
🗓️ Oct 30 | 12:00 PM PDT | Free
www.fixstars.com/en/news/877
We’ll also share updates from the latest AIBooster release — now featuring Autonomous Optimization for Edge AI and AI training.
🗓️ Oct 30 | 12:00 PM PDT | Free
📘 “Revolutionizing AI Development Efficiency Through GPU Optimization with Fixstars AIBooster.” www.fixstars.com/en/whitepape...
93% of companies underutilize GPUs. Learn how to uncover hidden inefficiencies and boost AI performance. Includes case studies from Sony Honda Mobility & Sakura Internet.
93% of companies underutilize GPUs. Learn how to uncover hidden inefficiencies and boost AI performance. Includes case studies from Sony Honda Mobility & Sakura Internet.
October 14, 2025 at 9:45 PM
📘 “Revolutionizing AI Development Efficiency Through GPU Optimization with Fixstars AIBooster.” www.fixstars.com/en/whitepape...
93% of companies underutilize GPUs. Learn how to uncover hidden inefficiencies and boost AI performance. Includes case studies from Sony Honda Mobility & Sakura Internet.
93% of companies underutilize GPUs. Learn how to uncover hidden inefficiencies and boost AI performance. Includes case studies from Sony Honda Mobility & Sakura Internet.
TEE basics in action: 🛡️ Kinibi on ARM TrustZone. blog.us.fixstars.com/introduction...
From Secure Boot to TA execution, Kinibi provides a closed-source TrustZone control stack with certifications like EAL5+ & FIPS140-2.
Our blog explains its operational flow step by step.
From Secure Boot to TA execution, Kinibi provides a closed-source TrustZone control stack with certifications like EAL5+ & FIPS140-2.
Our blog explains its operational flow step by step.
Introduction to TrustZone Kinibi Operational Overview - Fixstars Corporation Tech Blog
TEEs (Trusted Execution Environments), which enable programs to be executed "securely," are used for various applications such as DRM. While multiple TEE realization methods and implementations…
blog.us.fixstars.com
October 10, 2025 at 5:19 PM
TEE basics in action: 🛡️ Kinibi on ARM TrustZone. blog.us.fixstars.com/introduction...
From Secure Boot to TA execution, Kinibi provides a closed-source TrustZone control stack with certifications like EAL5+ & FIPS140-2.
Our blog explains its operational flow step by step.
From Secure Boot to TA execution, Kinibi provides a closed-source TrustZone control stack with certifications like EAL5+ & FIPS140-2.
Our blog explains its operational flow step by step.
🚀 Fixstars AIBooster just got smarter. www.fixstars.com/en/news/876
🔹 Edge AI Optimization – Converts PyTorch models for edge devices, boosting speed.
🔹 Training Optimization – Automates hyperparameter tuning for better performance.
🔹 SaaS Dashboard – Monitors GPU & AI workloads in real time.
🔹 Edge AI Optimization – Converts PyTorch models for edge devices, boosting speed.
🔹 Training Optimization – Automates hyperparameter tuning for better performance.
🔹 SaaS Dashboard – Monitors GPU & AI workloads in real time.
October 9, 2025 at 7:31 PM
🚀 Fixstars AIBooster just got smarter. www.fixstars.com/en/news/876
🔹 Edge AI Optimization – Converts PyTorch models for edge devices, boosting speed.
🔹 Training Optimization – Automates hyperparameter tuning for better performance.
🔹 SaaS Dashboard – Monitors GPU & AI workloads in real time.
🔹 Edge AI Optimization – Converts PyTorch models for edge devices, boosting speed.
🔹 Training Optimization – Automates hyperparameter tuning for better performance.
🔹 SaaS Dashboard – Monitors GPU & AI workloads in real time.
Bottlenecks → Breakthroughs ⚡
Join Fixstars’ encore webinar to see real B200 case studies, 5× faster training, and 50% lower inference costs.
📅 Oct 9 | 12 PM PDT | Free
🔗 www.fixstars.com/en/news/875?...
Join Fixstars’ encore webinar to see real B200 case studies, 5× faster training, and 50% lower inference costs.
📅 Oct 9 | 12 PM PDT | Free
🔗 www.fixstars.com/en/news/875?...
October 1, 2025 at 8:25 PM
Bottlenecks → Breakthroughs ⚡
Join Fixstars’ encore webinar to see real B200 case studies, 5× faster training, and 50% lower inference costs.
📅 Oct 9 | 12 PM PDT | Free
🔗 www.fixstars.com/en/news/875?...
Join Fixstars’ encore webinar to see real B200 case studies, 5× faster training, and 50% lower inference costs.
📅 Oct 9 | 12 PM PDT | Free
🔗 www.fixstars.com/en/news/875?...
What is 3D Gaussian Splatting and how can it be made faster?
blog.us.fixstars.com/introduction...
Our blog explains the technique, its role in novel-view synthesis, and how a simple optimization boosted rendering speed by 15%.
#3DGaussianSplatting #NeRF #ComputerGraphics #Fixstars
blog.us.fixstars.com/introduction...
Our blog explains the technique, its role in novel-view synthesis, and how a simple optimization boosted rendering speed by 15%.
#3DGaussianSplatting #NeRF #ComputerGraphics #Fixstars
Introduction to 3D Gaussian Splatting and Its Speedup - Fixstars Corporation Tech Blog
I will introduce 3D Gaussian Splatting and explain the work I did on its acceleration as an internship project.
blog.us.fixstars.com
September 30, 2025 at 5:19 PM
What is 3D Gaussian Splatting and how can it be made faster?
blog.us.fixstars.com/introduction...
Our blog explains the technique, its role in novel-view synthesis, and how a simple optimization boosted rendering speed by 15%.
#3DGaussianSplatting #NeRF #ComputerGraphics #Fixstars
blog.us.fixstars.com/introduction...
Our blog explains the technique, its role in novel-view synthesis, and how a simple optimization boosted rendering speed by 15%.
#3DGaussianSplatting #NeRF #ComputerGraphics #Fixstars
From hello world to matrix multiplication, our SYCL guide 👉 blog.us.fixstars.com/introduction...
This blog shows step by step how to build and run code across multiple devices, and compares performance on NVIDIA vs Intel GPUs — all from the same program. m the same program.
This blog shows step by step how to build and run code across multiple devices, and compares performance on NVIDIA vs Intel GPUs — all from the same program. m the same program.
Introduction to SYCL Programming: A Beginner's Guide - Fixstars Corporation Tech Blog
I will explain SYCL, starting with an overview, then covering the basic features and terminology essential for a fundamental understanding, and finally, how to create and run simple programs.
blog.us.fixstars.com
September 26, 2025 at 5:19 PM
From hello world to matrix multiplication, our SYCL guide 👉 blog.us.fixstars.com/introduction...
This blog shows step by step how to build and run code across multiple devices, and compares performance on NVIDIA vs Intel GPUs — all from the same program. m the same program.
This blog shows step by step how to build and run code across multiple devices, and compares performance on NVIDIA vs Intel GPUs — all from the same program. m the same program.
💾 Cut your ROS 2 bag file size from 7.9GB → 2.8GB with MCAP + Zstd compression.
blog.us.fixstars.com/using-rosbag...
Our blog shows how to use rosbag2_storage_mcap and why MCAP is becoming the default format for ROS 2.
blog.us.fixstars.com/using-rosbag...
Our blog shows how to use rosbag2_storage_mcap and why MCAP is becoming the default format for ROS 2.
Using rosbag2_storage_mcap: Integrating MCAP for Efficient ROS 2 Data Storage - Fixstars Corporation Tech Blog
The rosbag2_storage_mcap package has been added, allowing MCAP to be selected as the storage for rosbag2. This article introduces how to use rosbag2_storage_mcap.
blog.us.fixstars.com
September 22, 2025 at 11:50 PM
💾 Cut your ROS 2 bag file size from 7.9GB → 2.8GB with MCAP + Zstd compression.
blog.us.fixstars.com/using-rosbag...
Our blog shows how to use rosbag2_storage_mcap and why MCAP is becoming the default format for ROS 2.
blog.us.fixstars.com/using-rosbag...
Our blog shows how to use rosbag2_storage_mcap and why MCAP is becoming the default format for ROS 2.
We pushed Llama 4 on vLLM to its limits.
blog.us.fixstars.com/verifying-th...
With 8 H100s in FP8, it reached 1.5M tokens, while Maverick capped at 300k. Not yet 10M, but a major step toward large-scale deployment.
#AI #Llama4 #vLLM #Fixstars
blog.us.fixstars.com/verifying-th...
With 8 H100s in FP8, it reached 1.5M tokens, while Maverick capped at 300k. Not yet 10M, but a major step toward large-scale deployment.
#AI #Llama4 #vLLM #Fixstars
Verifying the Optimal Context Length for Deploying Llama 4 with vLLM - Fixstars Corporation Tech Blog
Llama 4 supports a long context of up to 10 million tokens; however, for actual deployment, it's important to be mindful of library and hardware constraints. In this article, through experiments…
blog.us.fixstars.com
September 18, 2025 at 5:19 PM
We pushed Llama 4 on vLLM to its limits.
blog.us.fixstars.com/verifying-th...
With 8 H100s in FP8, it reached 1.5M tokens, while Maverick capped at 300k. Not yet 10M, but a major step toward large-scale deployment.
#AI #Llama4 #vLLM #Fixstars
blog.us.fixstars.com/verifying-th...
With 8 H100s in FP8, it reached 1.5M tokens, while Maverick capped at 300k. Not yet 10M, but a major step toward large-scale deployment.
#AI #Llama4 #vLLM #Fixstars
[Webinar Tomorrow] Is your AI infrastructure keeping up with demand? Join us tomorrow as we dive into real-world performance engineering to help you eliminate bottlenecks and maximize efficiency.
🗓 Sept 16 | 12:00–12:40 PM PDT
🔗 Register now: fixstars.com/en/news/874?...
🗓 Sept 16 | 12:00–12:40 PM PDT
🔗 Register now: fixstars.com/en/news/874?...
Announcing Our Upcoming Seminar: “Optimize AI Infrastructure: From Bottlenecks to Breakthroughs” -Real-World Performance Engineering with Fixstars AIBooster- - Fixstars Corporation
In this session, Fixstars engineers will introduce key concepts in AI performance engineering and demonstrate how to identify and fix inefficiencies across training and inference workflows.
fixstars.com
September 15, 2025 at 3:56 PM
[Webinar Tomorrow] Is your AI infrastructure keeping up with demand? Join us tomorrow as we dive into real-world performance engineering to help you eliminate bottlenecks and maximize efficiency.
🗓 Sept 16 | 12:00–12:40 PM PDT
🔗 Register now: fixstars.com/en/news/874?...
🗓 Sept 16 | 12:00–12:40 PM PDT
🔗 Register now: fixstars.com/en/news/874?...
Reposted by Aki_A
❓ Struggling with AI bottlenecks?
Join our free webinar to learn how Fixstars AIBooster delivers 5× faster training & 50% lower inference costs with the new NVIDIA B200.
www.fixstars.com/en/news/874?...
📅 Sept 16 | ⏰ 12–12:40 PM PDT | 📍 Zoom
Join our free webinar to learn how Fixstars AIBooster delivers 5× faster training & 50% lower inference costs with the new NVIDIA B200.
www.fixstars.com/en/news/874?...
📅 Sept 16 | ⏰ 12–12:40 PM PDT | 📍 Zoom
Announcing Our Upcoming Seminar: “Optimize AI Infrastructure: From Bottlenecks to Breakthroughs” -Real-World Performance Engineering with Fixstars AIBooster- - Fixstars Corporation
In this session, Fixstars engineers will introduce key concepts in AI performance engineering and demonstrate how to identify and fix inefficiencies across training and inference workflows.
www.fixstars.com
September 10, 2025 at 9:01 PM
❓ Struggling with AI bottlenecks?
Join our free webinar to learn how Fixstars AIBooster delivers 5× faster training & 50% lower inference costs with the new NVIDIA B200.
www.fixstars.com/en/news/874?...
📅 Sept 16 | ⏰ 12–12:40 PM PDT | 📍 Zoom
Join our free webinar to learn how Fixstars AIBooster delivers 5× faster training & 50% lower inference costs with the new NVIDIA B200.
www.fixstars.com/en/news/874?...
📅 Sept 16 | ⏰ 12–12:40 PM PDT | 📍 Zoom
⚡ New: Intel AMX (Advanced Matrix Extension) explained!
blog.us.fixstars.com/intel-amx-ad...
🟩 Tile ops for Xeon matrix math
💻 Sample C++ code (clang/gcc)
📈 11.7× faster (1-thread) & ~2× (112-threads) vs AVX512
blog.us.fixstars.com/intel-amx-ad...
🟩 Tile ops for Xeon matrix math
💻 Sample C++ code (clang/gcc)
📈 11.7× faster (1-thread) & ~2× (112-threads) vs AVX512
Intel AMX (Advanced Matrix Extension) Explained (Introduction) - Fixstars Corporation Tech Blog
We introduced the instruction set, operation, and simple sample code for Intel's AMX (Advanced Matrix Extension), and confirmed that it can calculate matrix multiplication 11.7 times faster than AVX…
blog.us.fixstars.com
September 11, 2025 at 6:27 PM
⚡ New: Intel AMX (Advanced Matrix Extension) explained!
blog.us.fixstars.com/intel-amx-ad...
🟩 Tile ops for Xeon matrix math
💻 Sample C++ code (clang/gcc)
📈 11.7× faster (1-thread) & ~2× (112-threads) vs AVX512
blog.us.fixstars.com/intel-amx-ad...
🟩 Tile ops for Xeon matrix math
💻 Sample C++ code (clang/gcc)
📈 11.7× faster (1-thread) & ~2× (112-threads) vs AVX512
❓ Struggling with AI bottlenecks?
Join our free webinar to learn how Fixstars AIBooster delivers 5× faster training & 50% lower inference costs with the new NVIDIA B200.
www.fixstars.com/en/news/874?...
📅 Sept 16 | ⏰ 12–12:40 PM PDT | 📍 Zoom
Join our free webinar to learn how Fixstars AIBooster delivers 5× faster training & 50% lower inference costs with the new NVIDIA B200.
www.fixstars.com/en/news/874?...
📅 Sept 16 | ⏰ 12–12:40 PM PDT | 📍 Zoom
Announcing Our Upcoming Seminar: “Optimize AI Infrastructure: From Bottlenecks to Breakthroughs” -Real-World Performance Engineering with Fixstars AIBooster- - Fixstars Corporation
In this session, Fixstars engineers will introduce key concepts in AI performance engineering and demonstrate how to identify and fix inefficiencies across training and inference workflows.
www.fixstars.com
September 10, 2025 at 9:01 PM
❓ Struggling with AI bottlenecks?
Join our free webinar to learn how Fixstars AIBooster delivers 5× faster training & 50% lower inference costs with the new NVIDIA B200.
www.fixstars.com/en/news/874?...
📅 Sept 16 | ⏰ 12–12:40 PM PDT | 📍 Zoom
Join our free webinar to learn how Fixstars AIBooster delivers 5× faster training & 50% lower inference costs with the new NVIDIA B200.
www.fixstars.com/en/news/874?...
📅 Sept 16 | ⏰ 12–12:40 PM PDT | 📍 Zoom