docs(gpu): wording

Author: bene2k1

pages/gpu/reference-content/migration-h100.mdx

@@ -16,7 +16,7 @@ For optimal availability and performance, we recommend switching from **H100-2-8
 There are two primary scenarios: migrating **Kubernetes (Kapsule)** workloads or **standalone** workloads.
 
 <Message type="important">
-  Always ensure that your **data is backed up** before performing any operations that could affect it. Keep in mind that **scratch storage** is ephemeral and does not survive once the Instance is stopped: doing a full stop/start cycle will **erase the scratch data**. However, doing a simple reboot or using the **stop in place** function will keep the data.
+  Always make sure your **data is backed up** before performing any operation that could affect it. Remember that **scratch storage** is ephemeral and will not persist after an Instance is fully stopped. A full stop/start cycle—such as during an Instance server migration—will **erase all scratch data**. However, outside of server-type migrations, a simple reboot or using **stop in place** will preserve the data stored on the Instance’s scratch storage.
 </Message>
 
 ### Migrating Kubernetes workloads (Kubernetes Kapsule)
@@ -86,11 +86,9 @@ H100 PCIe-based GPU Instances are not End-of-Life (EOL), but due to limited avai
 #### Is H100-SXM-2-80G compatible with my current setup?
 Yes — it runs the same CUDA toolchain and supports standard frameworks (PyTorch, TensorFlow, etc.). No changes in your code base are required when upgrading to a SXM-based GPU Instance.
 
-#### Why is H100-SXM better for multi-GPU?
-The NVIDIA H100-SXM outperforms the H100-PCIe in multi-GPU configurations due to its superior interconnect and higher power capacity.
-It leverages fourth-generation NVLink and NVSwitch, providing up to 900 GB/s of bidirectional bandwidth for rapid GPU-to-GPU communication, compared to the H100-PCIe's 128 GB/s via PCIe Gen 5, which creates bottlenecks in demanding workloads like large-scale AI training and HPC. 
-Additionally, the H100-SXM’s 700W TDP enables higher clock speeds and sustained performance, while the H100-PCIe’s 300-350W TDP limits its throughput.
-For high-communication, multi-GPU tasks, the H100-SXM is the optimal choice, while the H100-PCIe suits less intensive applications with greater flexibility.
+### Why is the H100-SXM better for multi-GPU workloads?
 
-#### What if my workload needs more CPU or RAM?
-Let us know via [support ticket](https://console.scaleway.com/support/tickets/create) what your specific requoirements are. Currently we are evaluating options for compute-optimized configurations to complement our GPU offerings.
+The NVIDIA H100-SXM outperforms the H100-PCIe in multi-GPU configurations primarily due to its higher interconnect bandwidth and greater power capacity. It uses fourth-generation NVLink and NVSwitch, delivering up to **900 GB/s of bidirectional bandwidth** for fast GPU-to-GPU communication. In contrast, the H100-PCIe is limited to a **theoretical maximum of 128 GB/s** via PCIe Gen 5, which becomes a bottleneck in communication-heavy workloads such as large-scale AI training and HPC.
+The H100-SXM also provides **HBM3e memory** with up to **3.35 TB/s of bandwidth**, compared to **2 TB/s** with the H100-PCIe’s HBM3, improving performance in memory-bound tasks.
+Additionally, the H100-SXM’s **700W TDP** allows higher sustained clock speeds and throughput, while the H100-PCIe’s **300–350W TDP** imposes stricter performance limits.
+Overall, the H100-SXM is the optimal choice for high-communication, multi-GPU workloads, whereas the H100-PCIe offers more flexibility for less communication-intensive applications.