[PATCH v4 7/8] aws: add GPU instance defconfigs for AI/ML workloads

[Luis Chamberlain <mcgrof@kernel.org>]

Add two AWS GPU instance defconfigs to support GPU-accelerated
workloads:

1. aws-gpu-p5-48xlarge: High-end configuration with 8x NVIDIA H100 80GB
   GPUs for large-scale AI/ML training and inference. Includes 500GB
   storage for models and datasets.

2. aws-gpu-g5-xlarge: Cost-effective single NVIDIA A10G 24GB GPU for
   development and smaller workloads. Includes 200GB storage.

Both configurations:
- Use Debian 12 for stability
- Enable kernel development workflows (KOTD)
- Configure 9P filesystem for host-guest development
- Set up in us-east-1 for best GPU availability
- Use GP3 EBS volumes for better performance

These defconfigs leverage the new dynamic Kconfig generation to
automatically include all available P5 and G5 instance sizes.

Generated-by: Claude AI
Signed-off-by: Luis Chamberlain <mcgrof@kernel.org>
---
 defconfigs/aws-gpu-g5-xlarge   | 36 ++++++++++++++++++++++++++++++++++
 defconfigs/aws-gpu-p5-48xlarge | 36 ++++++++++++++++++++++++++++++++++
 2 files changed, 72 insertions(+)
 create mode 100644 defconfigs/aws-gpu-g5-xlarge
 create mode 100644 defconfigs/aws-gpu-p5-48xlarge

diff --git a/defconfigs/aws-gpu-g5-xlarge b/defconfigs/aws-gpu-g5-xlarge
new file mode 100644
index 00000000..a2f0a21f
--- /dev/null
+++ b/defconfigs/aws-gpu-g5-xlarge
@@ -0,0 +1,36 @@
+# AWS G5.xlarge GPU instance configuration
+# 1x NVIDIA A10G 24GB GPU for cost-effective GPU development
+
+CONFIG_KDEVOPS_FIRST_RUN=y
+CONFIG_TERRAFORM=y
+CONFIG_TERRAFORM_AWS=y
+CONFIG_TERRAFORM_AWS_DATA_ENABLE=y
+
+# Instance configuration
+CONFIG_TERRAFORM_AWS_INSTANCE_FAMILY_G5=y
+CONFIG_TERRAFORM_AWS_INSTANCE_G5_XLARGE=y
+
+# Moderate root volume
+CONFIG_TERRAFORM_AWS_EBS_SIZE=200
+CONFIG_TERRAFORM_AWS_EBS_VOLUME_TYPE_GP3=y
+
+# US East 1 for availability
+CONFIG_TERRAFORM_AWS_REGION_US_EAST_1=y
+
+# Debian 12 for stability
+CONFIG_VAGRANT_DEBIAN=y
+CONFIG_VAGRANT_DEBIAN12=y
+
+# Use kdevops user
+CONFIG_KDEVOPS_SSH_CONFIG_USER_KDEVOPS=y
+
+# Workflows
+CONFIG_KDEVOPS_WORKFLOW_ENABLE_KOTD=y
+CONFIG_KDEVOPS_WORKFLOW_KOTD_ENABLE_LATEST=y
+
+# Enable git for kernel development
+CONFIG_BOOTLINUX_9P=y
+CONFIG_BOOTLINUX_TREE_LINUS=y
+
+# Networking
+CONFIG_KDEVOPS_ENABLE_TERRAFORM_SSH_COMPLETE_SETUP=y
\ No newline at end of file
diff --git a/defconfigs/aws-gpu-p5-48xlarge b/defconfigs/aws-gpu-p5-48xlarge
new file mode 100644
index 00000000..741b98e3
--- /dev/null
+++ b/defconfigs/aws-gpu-p5-48xlarge
@@ -0,0 +1,36 @@
+# AWS P5.48xlarge GPU instance configuration
+# 8x NVIDIA H100 80GB GPUs for AI/ML workloads
+
+CONFIG_KDEVOPS_FIRST_RUN=y
+CONFIG_TERRAFORM=y
+CONFIG_TERRAFORM_AWS=y
+CONFIG_TERRAFORM_AWS_DATA_ENABLE=y
+
+# Instance configuration
+CONFIG_TERRAFORM_AWS_INSTANCE_FAMILY_P5=y
+CONFIG_TERRAFORM_AWS_INSTANCE_P5_48XLARGE=y
+
+# Large root volume for models and datasets
+CONFIG_TERRAFORM_AWS_EBS_SIZE=500
+CONFIG_TERRAFORM_AWS_EBS_VOLUME_TYPE_GP3=y
+
+# US East 1 typically has best GPU availability
+CONFIG_TERRAFORM_AWS_REGION_US_EAST_1=y
+
+# Debian 12 for stability
+CONFIG_VAGRANT_DEBIAN=y
+CONFIG_VAGRANT_DEBIAN12=y
+
+# Use kdevops user
+CONFIG_KDEVOPS_SSH_CONFIG_USER_KDEVOPS=y
+
+# Workflows
+CONFIG_KDEVOPS_WORKFLOW_ENABLE_KOTD=y
+CONFIG_KDEVOPS_WORKFLOW_KOTD_ENABLE_LATEST=y
+
+# Enable git for kernel development
+CONFIG_BOOTLINUX_9P=y
+CONFIG_BOOTLINUX_TREE_LINUS=y
+
+# Networking
+CONFIG_KDEVOPS_ENABLE_TERRAFORM_SSH_COMPLETE_SETUP=y
\ No newline at end of file
-- 
2.51.0