From d74c0d0b82edc8881df79caa169dd65d8239e1bd Mon Sep 17 00:00:00 2001 From: Carlos Gavidia-Calderon Date: Wed, 13 Nov 2024 08:28:34 +0000 Subject: [PATCH] Changing suggested SKU to Standard_D8s_v5 --- docs/source/deployment/deploy_sre.md | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) diff --git a/docs/source/deployment/deploy_sre.md b/docs/source/deployment/deploy_sre.md index 5157de2267..5a5a5b4166 100644 --- a/docs/source/deployment/deploy_sre.md +++ b/docs/source/deployment/deploy_sre.md @@ -167,7 +167,7 @@ As some general recommendations, - For general purpose use, the D family gives decent performance and a good balance of CPU and memory. The [Dsv6 series](https://learn.microsoft.com/en-us/azure/virtual-machines/sizes/general-purpose/dsv6-series#sizes-in-series) is a good starting point and can be scaled from 2 CPUs and 8 GB RAM to 128 CPUs and 512 GB RAM. - - `Standard_D8s_v6` should give reasonable performance for a single concurrent user. + - `Standard_D8s_v5` should give reasonable performance for a single concurrent user. - For GPU accelerated work, the NC family provides Nvidia GPUs and a good balance of CPU and memory. In order of increasing throughput, the `NCv3` series features Nvidia V100 GPUs, the `NC_A100_v4` series features Nvidia A100 GPUs, and the `NCads_H100_v5` series features Nvidia H100 GPUs. - `Stanard_NC6s_v3` should give reasonable performance for a single concurrent user with AI/ML workloads.