<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Amd on Le Site de François</title><link>https://lesitedefrancois.be/en/tags/amd/</link><description>Recent content in Amd on Le Site de François</description><generator>Hugo -- gohugo.io</generator><language>en</language><copyright>© 2026 François</copyright><atom:link href="https://lesitedefrancois.be/en/tags/amd/index.xml" rel="self" type="application/rss+xml"/><item><title>ROCm (Radeon Open Compute)</title><link>https://lesitedefrancois.be/en/ai/rocm/</link><pubDate>Mon, 01 Jan 0001 00:00:00 +0000</pubDate><guid>https://lesitedefrancois.be/en/ai/rocm/</guid><description>&lt;p&gt;&lt;strong&gt;ROCm (Radeon Open Compute)&lt;/strong&gt; is AMD’s software stack for &lt;strong&gt;GPU compute&lt;/strong&gt; on datacenter &lt;strong&gt;Instinct&lt;/strong&gt; accelerators (and select consumer GPUs in community setups). Its objective mirrors &lt;strong&gt;CUDA&lt;/strong&gt; for NVIDIA: provide kernel compilers (&lt;strong&gt;HIP&lt;/strong&gt;), math libraries (rocBLAS, rocFFT), collective communication (&lt;strong&gt;RCCL&lt;/strong&gt;, analogous to &lt;strong&gt;NCCL&lt;/strong&gt;), and framework integrations so &lt;strong&gt;PyTorch&lt;/strong&gt; and inference runtimes can execute training and &lt;strong&gt;inference&lt;/strong&gt; on AMD hardware. ROCm is positioned as an open platform (Linux-first) for customers who want accelerator choice or who standardize on AMD in HPC and AI clusters.&lt;/p&gt;</description></item><item><title>SEV-SNP (AMD Secure Encrypted Virtualization – Secure Nested Paging)</title><link>https://lesitedefrancois.be/en/security/sev-snp/</link><pubDate>Mon, 01 Jan 0001 00:00:00 +0000</pubDate><guid>https://lesitedefrancois.be/en/security/sev-snp/</guid><description>&lt;p&gt;&lt;strong&gt;SEV-SNP&lt;/strong&gt; is AMD&amp;rsquo;s third-generation confidential computing technology for EPYC processors, and the generation in production use across major cloud providers (AWS, Google Cloud) and Linux distributions today. It builds on two predecessors: &lt;strong&gt;SEV&lt;/strong&gt; (2016), which encrypted each VM&amp;rsquo;s memory with a per-VM AES key managed by the AMD Secure Processor, and &lt;strong&gt;SEV-ES&lt;/strong&gt; (2017), which additionally encrypted CPU register state on VM exit to prevent the hypervisor from reading guest execution state. SEV-SNP&amp;rsquo;s defining addition is &lt;strong&gt;memory integrity&lt;/strong&gt;: using Secure Nested Paging, the firmware enforces that if a guest can read an encrypted memory location, the value returned must be exactly what the guest last wrote there — closing the replay, remap, and memory aliasing attacks that made earlier generations insufficient for a fully untrusted hypervisor threat model.&lt;/p&gt;</description></item></channel></rss>