aws-bootstrap-g4dn 0.1.0__tar.gz → 0.3.0__tar.gz

{aws_bootstrap_g4dn-0.1.0 → aws_bootstrap_g4dn-0.3.0}/.github/workflows/ci.yml

@@ -20,7 +20,7 @@ jobs:
     strategy:
       fail-fast: false
       matrix:
-        python-version: ["3.14"]
+        python-version: ["3.12", "3.13", "3.14"]
 
     steps:
     - uses: actions/checkout@v4

{aws_bootstrap_g4dn-0.1.0 → aws_bootstrap_g4dn-0.3.0}/.github/workflows/publish-to-pypi.yml

@@ -7,6 +7,8 @@ jobs:
   build:
     name: Build distribution
     runs-on: ubuntu-latest
+    permissions:
+      contents: read
 
     steps:
     - uses: actions/checkout@v4

{aws_bootstrap_g4dn-0.1.0 → aws_bootstrap_g4dn-0.3.0}/CLAUDE.md

@@ -10,7 +10,7 @@ Target workflows: Jupyter server-client, VSCode Remote SSH, and NVIDIA Nsight re
 
 ## Tech Stack & Requirements
 
-- **Python 3.14+** with **uv** package manager (astral-sh/uv) — used for venv creation, dependency management, and running the project
+- **Python 3.12+** with **uv** package manager (astral-sh/uv) — used for venv creation, dependency management, and running the project
 - **boto3** — AWS SDK for EC2 provisioning (AMI lookup, security groups, instance launch, waiters)
 - **click** — CLI framework with built-in color support (`click.secho`, `click.style`)
 - **setuptools + setuptools-scm** — build backend with git-tag-based versioning (configured in pyproject.toml)
@@ -33,7 +33,8 @@ aws_bootstrap/
     cli.py               # Click CLI entry point (launch, status, terminate commands)
     config.py            # LaunchConfig dataclass with defaults
     ec2.py               # AMI lookup, security group, instance launch/find/terminate, polling, spot pricing
-    ssh.py               # SSH key pair import, SSH readiness check, remote setup, ~/.ssh/config management
+    gpu.py               # GPU architecture mapping and GpuInfo dataclass
+    ssh.py               # SSH key pair import, SSH readiness check, remote setup, ~/.ssh/config management, GPU queries
     resources/           # Non-Python artifacts SCP'd to remote instances
         __init__.py
         gpu_benchmark.py       # GPU throughput benchmark (CNN + Transformer), copied to ~/gpu_benchmark.py on instance
@@ -44,6 +45,7 @@ aws_bootstrap/
         test_config.py
         test_cli.py
         test_ec2.py
+        test_gpu.py
         test_ssh_config.py
         test_ssh_gpu.py
 docs/
@@ -54,8 +56,8 @@ Entry point: `aws-bootstrap = "aws_bootstrap.cli:main"` (installed via `uv sync`
 
 ## CLI Commands
 
-- **`launch`** — provisions an EC2 instance (spot by default, falls back to on-demand on capacity errors); adds SSH config alias (e.g. `aws-gpu1`) to `~/.ssh/config`
-- **`status`** — lists all non-terminated instances (including `shutting-down`) with type, IP, SSH alias, pricing (spot price/hr or on-demand), uptime, and estimated cost for running spot instances; `--gpu` flag queries GPU info via SSH, reporting both CUDA toolkit version (from `nvcc`) and driver-supported max (from `nvidia-smi`)
+- **`launch`** — provisions an EC2 instance (spot by default, falls back to on-demand on capacity errors); adds SSH config alias (e.g. `aws-gpu1`) to `~/.ssh/config`; `--python-version` controls which Python `uv` installs in the remote venv; `--ssh-port` overrides the default SSH port (22) for security group ingress, connection checks, and SSH config
+- **`status`** — lists all non-terminated instances (including `shutting-down`) with type, IP, SSH alias, pricing (spot price/hr or on-demand), uptime, and estimated cost for running spot instances; `--gpu` flag queries GPU info via SSH, reporting both CUDA toolkit version (from `nvcc`) and driver-supported max (from `nvidia-smi`); `--instructions` (default: on) prints connection commands (SSH, Jupyter tunnel, VSCode Remote SSH, GPU benchmark) for each running instance; suppress with `--no-instructions`
 - **`terminate`** — terminates instances by ID or all aws-bootstrap instances in the region; removes SSH config aliases
 - **`list instance-types`** — lists EC2 instance types matching a family prefix (default: `g4dn`), showing vCPUs, memory, and GPU info
 - **`list amis`** — lists available AMIs matching a name pattern (default: Deep Learning Base OSS Nvidia Driver GPU AMIs), sorted newest-first
@@ -96,7 +98,7 @@ The `KNOWN_CUDA_TAGS` array in `remote_setup.sh` lists the CUDA wheel tags publi
 ## Remote Setup Details
 
 `remote_setup.sh` also:
-- Creates `~/venv` and appends `source ~/venv/bin/activate` to `~/.bashrc` so the venv is auto-activated on SSH login
+- Creates `~/venv` and appends `source ~/venv/bin/activate` to `~/.bashrc` so the venv is auto-activated on SSH login. When `--python-version` is passed to `launch`, the CLI sets `PYTHON_VERSION` as an inline env var on the SSH command; `remote_setup.sh` reads it to run `uv python install` and `uv venv --python` with the requested version
 - Runs a quick CUDA smoke test (`torch.cuda.is_available()` + GPU matmul) after PyTorch installation to verify the GPU stack; prints a WARNING on failure but does not abort
 - Copies `gpu_benchmark.py` to `~/gpu_benchmark.py` and `gpu_smoke_test.ipynb` to `~/gpu_smoke_test.ipynb`
 

{aws_bootstrap_g4dn-0.1.0/aws_bootstrap_g4dn.egg-info → aws_bootstrap_g4dn-0.3.0}/PKG-INFO

@@ -1,13 +1,16 @@
 Metadata-Version: 2.4
 Name: aws-bootstrap-g4dn
-Version: 0.1.0
+Version: 0.3.0
 Summary: Bootstrap AWS EC2 GPU instances for hybrid local-remote development
 Author: Adam Ever-Hadani
 License-Expression: MIT
 Project-URL: Homepage, https://github.com/promptromp/aws-bootstrap-g4dn
 Project-URL: Issues, https://github.com/promptromp/aws-bootstrap-g4dn/issues
 Keywords: aws,ec2,gpu,cuda,deep-learning,spot-instances,cli
-Requires-Python: >=3.14
+Classifier: Programming Language :: Python :: 3.12
+Classifier: Programming Language :: Python :: 3.13
+Classifier: Programming Language :: Python :: 3.14
+Requires-Python: >=3.12
 Description-Content-Type: text/markdown
 License-File: LICENSE
 Requires-Dist: boto3>=1.35
@@ -55,7 +58,7 @@ ssh aws-gpu1                  # You're in, venv activated, PyTorch works
 
 1. AWS profile configured with relevant permissions (profile name can be passed via `--profile` or read from `AWS_PROFILE` env var)
 2. AWS CLI v2 — see [here](https://docs.aws.amazon.com/cli/latest/userguide/getting-started-install.html)
-3. Python 3.14+ and [uv](https://github.com/astral-sh/uv)
+3. Python 3.12+ and [uv](https://github.com/astral-sh/uv)
 4. An SSH key pair (see below)
 
 ## Installation
@@ -66,6 +69,16 @@ ssh aws-gpu1                  # You're in, venv activated, PyTorch works
 pip install aws-bootstrap-g4dn
 ```
 
+### With uvx (no install needed)
+
+[uvx](https://docs.astral.sh/uv/guides/tools/) runs the CLI directly in a temporary environment — no global install required:
+
+```bash
+uvx --from aws-bootstrap-g4dn aws-bootstrap launch
+uvx --from aws-bootstrap-g4dn aws-bootstrap status
+uvx --from aws-bootstrap-g4dn aws-bootstrap terminate
+```
+
 ### From source (development)
 
 ```bash
@@ -75,7 +88,7 @@ uv venv
 uv sync
 ```
 
-Both methods install the `aws-bootstrap` CLI.
+All methods install the `aws-bootstrap` CLI.
 
 ## SSH Key Setup
 
@@ -113,6 +126,12 @@ aws-bootstrap launch --on-demand --instance-type g5.xlarge --region us-east-1
 # Launch without running the remote setup script
 aws-bootstrap launch --no-setup
 
+# Use a specific Python version in the remote venv
+aws-bootstrap launch --python-version 3.13
+
+# Use a non-default SSH port
+aws-bootstrap launch --ssh-port 2222
+
 # Use a specific AWS profile
 aws-bootstrap launch --profile my-aws-profile
 ```
@@ -136,7 +155,7 @@ The setup script runs automatically on the instance after SSH becomes available:
 |------|------|
 | **GPU verify** | Confirms `nvidia-smi` and `nvcc` are working |
 | **Utilities** | Installs `htop`, `tmux`, `tree`, `jq` |
-| **Python venv** | Creates `~/venv` with `uv`, auto-activates in `~/.bashrc` |
+| **Python venv** | Creates `~/venv` with `uv`, auto-activates in `~/.bashrc`. Use `--python-version` to pin a specific Python (e.g. `3.13`) |
 | **CUDA-aware PyTorch** | Detects CUDA toolkit version → installs PyTorch from the matching `cu{TAG}` wheel index |
 | **CUDA smoke test** | Runs `torch.cuda.is_available()` + GPU matmul to verify the stack |
 | **GPU benchmark** | Copies `gpu_benchmark.py` to `~/gpu_benchmark.py` |
@@ -210,6 +229,9 @@ aws-bootstrap status
 # Include GPU info (CUDA toolkit + driver version, GPU name, architecture) via SSH
 aws-bootstrap status --gpu
 
+# Hide connection commands (shown by default for each running instance)
+aws-bootstrap status --no-instructions
+
 # List instances in a specific region
 aws-bootstrap status --region us-east-1
 

aws_bootstrap_g4dn-0.1.0/PKG-INFO → aws_bootstrap_g4dn-0.3.0/README.md

@@ -1,19 +1,3 @@
-Metadata-Version: 2.4
-Name: aws-bootstrap-g4dn
-Version: 0.1.0
-Summary: Bootstrap AWS EC2 GPU instances for hybrid local-remote development
-Author: Adam Ever-Hadani
-License-Expression: MIT
-Project-URL: Homepage, https://github.com/promptromp/aws-bootstrap-g4dn
-Project-URL: Issues, https://github.com/promptromp/aws-bootstrap-g4dn/issues
-Keywords: aws,ec2,gpu,cuda,deep-learning,spot-instances,cli
-Requires-Python: >=3.14
-Description-Content-Type: text/markdown
-License-File: LICENSE
-Requires-Dist: boto3>=1.35
-Requires-Dist: click>=8.1
-Dynamic: license-file
-
 # aws-bootstrap-g4dn
 
 --------------------------------------------------------------------------------
@@ -55,7 +39,7 @@ ssh aws-gpu1                  # You're in, venv activated, PyTorch works
 
 1. AWS profile configured with relevant permissions (profile name can be passed via `--profile` or read from `AWS_PROFILE` env var)
 2. AWS CLI v2 — see [here](https://docs.aws.amazon.com/cli/latest/userguide/getting-started-install.html)
-3. Python 3.14+ and [uv](https://github.com/astral-sh/uv)
+3. Python 3.12+ and [uv](https://github.com/astral-sh/uv)
 4. An SSH key pair (see below)
 
 ## Installation
@@ -66,6 +50,16 @@ ssh aws-gpu1                  # You're in, venv activated, PyTorch works
 pip install aws-bootstrap-g4dn
 ```
 
+### With uvx (no install needed)
+
+[uvx](https://docs.astral.sh/uv/guides/tools/) runs the CLI directly in a temporary environment — no global install required:
+
+```bash
+uvx --from aws-bootstrap-g4dn aws-bootstrap launch
+uvx --from aws-bootstrap-g4dn aws-bootstrap status
+uvx --from aws-bootstrap-g4dn aws-bootstrap terminate
+```
+
 ### From source (development)
 
 ```bash
@@ -75,7 +69,7 @@ uv venv
 uv sync
 ```
 
-Both methods install the `aws-bootstrap` CLI.
+All methods install the `aws-bootstrap` CLI.
 
 ## SSH Key Setup
 
@@ -113,6 +107,12 @@ aws-bootstrap launch --on-demand --instance-type g5.xlarge --region us-east-1
 # Launch without running the remote setup script
 aws-bootstrap launch --no-setup
 
+# Use a specific Python version in the remote venv
+aws-bootstrap launch --python-version 3.13
+
+# Use a non-default SSH port
+aws-bootstrap launch --ssh-port 2222
+
 # Use a specific AWS profile
 aws-bootstrap launch --profile my-aws-profile
 ```
@@ -136,7 +136,7 @@ The setup script runs automatically on the instance after SSH becomes available:
 |------|------|
 | **GPU verify** | Confirms `nvidia-smi` and `nvcc` are working |
 | **Utilities** | Installs `htop`, `tmux`, `tree`, `jq` |
-| **Python venv** | Creates `~/venv` with `uv`, auto-activates in `~/.bashrc` |
+| **Python venv** | Creates `~/venv` with `uv`, auto-activates in `~/.bashrc`. Use `--python-version` to pin a specific Python (e.g. `3.13`) |
 | **CUDA-aware PyTorch** | Detects CUDA toolkit version → installs PyTorch from the matching `cu{TAG}` wheel index |
 | **CUDA smoke test** | Runs `torch.cuda.is_available()` + GPU matmul to verify the stack |
 | **GPU benchmark** | Copies `gpu_benchmark.py` to `~/gpu_benchmark.py` |
@@ -210,6 +210,9 @@ aws-bootstrap status
 # Include GPU info (CUDA toolkit + driver version, GPU name, architecture) via SSH
 aws-bootstrap status --gpu
 
+# Hide connection commands (shown by default for each running instance)
+aws-bootstrap status --no-instructions
+
 # List instances in a specific region
 aws-bootstrap status --region us-east-1
 

{aws_bootstrap_g4dn-0.1.0 → aws_bootstrap_g4dn-0.3.0}/aws_bootstrap/cli.py

@@ -5,6 +5,7 @@ from datetime import UTC, datetime
 from pathlib import Path
 
 import boto3
+import botocore.exceptions
 import click
 
 from .config import LaunchConfig
@@ -56,7 +57,39 @@ def warn(msg: str) -> None:
     click.secho(f"  WARNING: {msg}", fg="yellow", err=True)
 
 
-@click.group()
+class _AWSGroup(click.Group):
+    """Click group that catches common AWS credential/auth errors."""
+
+    def invoke(self, ctx):
+        try:
+            return super().invoke(ctx)
+        except botocore.exceptions.NoCredentialsError:
+            raise CLIError(
+                "Unable to locate AWS credentials.\n\n"
+                "  Make sure you have configured AWS credentials using one of:\n"
+                "    - Set the AWS_PROFILE environment variable:  export AWS_PROFILE=<profile-name>\n"
+                "    - Pass --profile to the command:  aws-bootstrap <command> --profile <profile-name>\n"
+                "    - Configure a default profile:  aws configure\n\n"
+                "  See: https://docs.aws.amazon.com/cli/latest/userguide/cli-configure-files.html"
+            ) from None
+        except botocore.exceptions.ProfileNotFound as e:
+            raise CLIError(f"{e}\n\n  List available profiles with:  aws configure list-profiles") from None
+        except botocore.exceptions.PartialCredentialsError as e:
+            raise CLIError(
+                f"Incomplete AWS credentials: {e}\n\n  Check your AWS configuration with:  aws configure list"
+            ) from None
+        except botocore.exceptions.ClientError as e:
+            code = e.response["Error"]["Code"]
+            if code in ("AuthFailure", "UnauthorizedOperation", "ExpiredTokenException", "ExpiredToken"):
+                raise CLIError(
+                    f"AWS authorization failed: {e.response['Error']['Message']}\n\n"
+                    "  Your credentials may be expired or lack the required permissions.\n"
+                    "  Check your AWS configuration with:  aws configure list"
+                ) from None
+            raise
+
+
+@click.group(cls=_AWSGroup)
 @click.version_option(package_name="aws-bootstrap-g4dn")
 def main():
     """Bootstrap AWS EC2 GPU instances for hybrid local-remote development."""
@@ -80,6 +113,12 @@ def main():
 @click.option("--no-setup", is_flag=True, default=False, help="Skip running the remote setup script.")
 @click.option("--dry-run", is_flag=True, default=False, help="Show what would be done without executing.")
 @click.option("--profile", default=None, help="AWS profile override (defaults to AWS_PROFILE env var).")
+@click.option(
+    "--python-version",
+    default=None,
+    help="Python version for the remote venv (e.g. 3.13, 3.14.2). Passed to uv during setup.",
+)
+@click.option("--ssh-port", default=22, show_default=True, type=int, help="SSH port on the remote instance.")
 def launch(
     instance_type,
     ami_filter,
@@ -92,6 +131,8 @@ def launch(
     no_setup,
     dry_run,
     profile,
+    python_version,
+    ssh_port,
 ):
     """Launch a GPU-accelerated EC2 instance."""
     config = LaunchConfig(
@@ -104,6 +145,8 @@ def launch(
         volume_size=volume_size,
         run_setup=not no_setup,
         dry_run=dry_run,
+        ssh_port=ssh_port,
+        python_version=python_version,
     )
     if ami_filter:
         config.ami_filter = ami_filter
@@ -130,7 +173,7 @@ def launch(
 
     # Step 3: Security group
     step(3, 6, "Ensuring security group...")
-    sg_id = ensure_security_group(ec2, config.security_group, config.tag_value)
+    sg_id = ensure_security_group(ec2, config.security_group, config.tag_value, ssh_port=config.ssh_port)
 
     pricing = "spot" if config.spot else "on-demand"
 
@@ -145,6 +188,10 @@ def launch(
         val("Volume", f"{config.volume_size} GB gp3")
         val("Region", config.region)
         val("Remote setup", "yes" if config.run_setup else "no")
+        if config.ssh_port != 22:
+            val("SSH port", str(config.ssh_port))
+        if config.python_version:
+            val("Python version", config.python_version)
         click.echo()
         click.secho("No resources launched (dry-run mode).", fg="yellow")
         return
@@ -169,9 +216,13 @@ def launch(
     # Step 6: SSH and remote setup
     step(6, 6, "Waiting for SSH access...")
     private_key = private_key_path(config.key_path)
-    if not wait_for_ssh(public_ip, config.ssh_user, config.key_path):
+    if not wait_for_ssh(public_ip, config.ssh_user, config.key_path, port=config.ssh_port):
         warn("SSH did not become available within the timeout.")
-        info(f"Instance is running — try connecting manually: ssh -i {private_key} {config.ssh_user}@{public_ip}")
+        port_flag = f" -p {config.ssh_port}" if config.ssh_port != 22 else ""
+        info(
+            f"Instance is running — try connecting manually:"
+            f" ssh -i {private_key}{port_flag} {config.ssh_user}@{public_ip}"
+        )
         return
 
     if config.run_setup:
@@ -179,7 +230,9 @@ def launch(
             warn(f"Setup script not found at {SETUP_SCRIPT}, skipping.")
         else:
             info("Running remote setup...")
-            if run_remote_setup(public_ip, config.ssh_user, config.key_path, SETUP_SCRIPT):
+            if run_remote_setup(
+                public_ip, config.ssh_user, config.key_path, SETUP_SCRIPT, config.python_version, port=config.ssh_port
+            ):
                 success("Remote setup completed successfully.")
             else:
                 warn("Remote setup failed. Instance is still running.")
@@ -191,6 +244,7 @@ def launch(
         user=config.ssh_user,
         key_path=config.key_path,
         alias_prefix=config.alias_prefix,
+        port=config.ssh_port,
     )
     success(f"Added SSH config alias: {alias}")
 
@@ -206,18 +260,27 @@ def launch(
     val("Pricing", pricing)
     val("SSH alias", alias)
 
+    port_flag = f" -p {config.ssh_port}" if config.ssh_port != 22 else ""
+
     click.echo()
     click.secho("  SSH:", fg="cyan")
-    click.secho(f"    ssh {alias}", bold=True)
-    info(f"or: ssh -i {private_key} {config.ssh_user}@{public_ip}")
+    click.secho(f"    ssh{port_flag} {alias}", bold=True)
+    info(f"or: ssh -i {private_key}{port_flag} {config.ssh_user}@{public_ip}")
 
     click.echo()
     click.secho("  Jupyter (via SSH tunnel):", fg="cyan")
-    click.secho(f"    ssh -NL 8888:localhost:8888 {alias}", bold=True)
-    info(f"or: ssh -i {private_key} -NL 8888:localhost:8888 {config.ssh_user}@{public_ip}")
+    click.secho(f"    ssh -NL 8888:localhost:8888{port_flag} {alias}", bold=True)
+    info(f"or: ssh -i {private_key} -NL 8888:localhost:8888{port_flag} {config.ssh_user}@{public_ip}")
     info("Then open: http://localhost:8888")
     info("Notebook: ~/gpu_smoke_test.ipynb (GPU smoke test)")
 
+    click.echo()
+    click.secho("  VSCode Remote SSH:", fg="cyan")
+    click.secho(
+        f"    code --folder-uri vscode-remote://ssh-remote+{alias}/home/{config.ssh_user}",
+        bold=True,
+    )
+
     click.echo()
     click.secho("  GPU Benchmark:", fg="cyan")
     click.secho(f"    ssh {alias} 'python ~/gpu_benchmark.py'", bold=True)
@@ -233,7 +296,14 @@ def launch(
 @click.option("--region", default="us-west-2", show_default=True, help="AWS region.")
 @click.option("--profile", default=None, help="AWS profile override.")
 @click.option("--gpu", is_flag=True, default=False, help="Query GPU info (CUDA, driver) via SSH.")
-def status(region, profile, gpu):
+@click.option(
+    "--instructions/--no-instructions",
+    "-I",
+    default=True,
+    show_default=True,
+    help="Show connection commands (SSH, Jupyter, VSCode) for each running instance.",
+)
+def status(region, profile, gpu, instructions):
     """Show running instances created by aws-bootstrap."""
     session = boto3.Session(profile_name=profile, region_name=region)
     ec2 = session.client("ec2")
@@ -272,11 +342,15 @@ def status(region, profile, gpu):
         if inst["PublicIp"]:
             val("    IP", inst["PublicIp"])
 
+        # Look up SSH config details once (used by --gpu and --with-instructions)
+        details = None
+        if (gpu or instructions) and state == "running" and inst["PublicIp"]:
+            details = get_ssh_host_details(inst["InstanceId"])
+
         # GPU info (opt-in, only for running instances with a public IP)
         if gpu and state == "running" and inst["PublicIp"]:
-            details = get_ssh_host_details(inst["InstanceId"])
             if details:
-                gpu_info = query_gpu_info(details.hostname, details.user, details.identity_file)
+                gpu_info = query_gpu_info(details.hostname, details.user, details.identity_file, port=details.port)
             else:
                 gpu_info = query_gpu_info(
                     inst["PublicIp"],
@@ -320,6 +394,29 @@ def status(region, profile, gpu):
                 val("    Est. cost", f"~${est_cost:.4f}")
 
         val("    Launched", str(inst["LaunchTime"]))
+
+        # Connection instructions (opt-in, only for running instances with a public IP and alias)
+        if instructions and state == "running" and inst["PublicIp"] and alias:
+            user = details.user if details else "ubuntu"
+            port = details.port if details else 22
+            port_flag = f" -p {port}" if port != 22 else ""
+
+            click.echo()
+            click.secho("    SSH:", fg="cyan")
+            click.secho(f"      ssh{port_flag} {alias}", bold=True)
+
+            click.secho("    Jupyter (via SSH tunnel):", fg="cyan")
+            click.secho(f"      ssh -NL 8888:localhost:8888{port_flag} {alias}", bold=True)
+
+            click.secho("    VSCode Remote SSH:", fg="cyan")
+            click.secho(
+                f"      code --folder-uri vscode-remote://ssh-remote+{alias}/home/{user}",
+                bold=True,
+            )
+
+            click.secho("    GPU Benchmark:", fg="cyan")
+            click.secho(f"      ssh {alias} 'python ~/gpu_benchmark.py'", bold=True)
+
     click.echo()
     first_id = instances[0]["InstanceId"]
     click.echo("  To terminate:  " + click.style(f"aws-bootstrap terminate {first_id}", bold=True))

{aws_bootstrap_g4dn-0.1.0 → aws_bootstrap_g4dn-0.3.0}/aws_bootstrap/config.py

@@ -22,3 +22,5 @@ class LaunchConfig:
     ssh_user: str = "ubuntu"
     tag_value: str = "aws-bootstrap-g4dn"
     alias_prefix: str = "aws-gpu"
+    ssh_port: int = 22
+    python_version: str | None = None

{aws_bootstrap_g4dn-0.1.0 → aws_bootstrap_g4dn-0.3.0}/aws_bootstrap/ec2.py

@@ -59,7 +59,7 @@ def get_latest_ami(ec2_client, ami_filter: str) -> dict:
     return images[0]
 
 
-def ensure_security_group(ec2_client, name: str, tag_value: str) -> str:
+def ensure_security_group(ec2_client, name: str, tag_value: str, ssh_port: int = 22) -> str:
     """Find or create a security group with SSH ingress in the default VPC."""
     # Find default VPC
     vpcs = ec2_client.describe_vpcs(Filters=[{"Name": "isDefault", "Values": ["true"]}])
@@ -103,8 +103,8 @@ def ensure_security_group(ec2_client, name: str, tag_value: str) -> str:
         IpPermissions=[
             {
                 "IpProtocol": "tcp",
-                "FromPort": 22,
-                "ToPort": 22,
+                "FromPort": ssh_port,
+                "ToPort": ssh_port,
                 "IpRanges": [{"CidrIp": "0.0.0.0/0", "Description": "SSH access"}],
             }
         ],

aws_bootstrap_g4dn-0.3.0/aws_bootstrap/gpu.py

@@ -0,0 +1,27 @@
+"""GPU architecture mapping and GPU info dataclass."""
+
+from __future__ import annotations
+from dataclasses import dataclass
+
+
+_GPU_ARCHITECTURES: dict[str, str] = {
+    "7.0": "Volta",
+    "7.5": "Turing",
+    "8.0": "Ampere",
+    "8.6": "Ampere",
+    "8.7": "Ampere",
+    "8.9": "Ada Lovelace",
+    "9.0": "Hopper",
+}
+
+
+@dataclass
+class GpuInfo:
+    """GPU information retrieved via nvidia-smi and nvcc."""
+
+    driver_version: str
+    cuda_driver_version: str  # max CUDA version supported by driver (from nvidia-smi)
+    cuda_toolkit_version: str | None  # actual CUDA toolkit installed (from nvcc), None if unavailable
+    gpu_name: str
+    compute_capability: str
+    architecture: str

{aws_bootstrap_g4dn-0.1.0 → aws_bootstrap_g4dn-0.3.0}/aws_bootstrap/resources/remote_setup.sh

@@ -34,7 +34,13 @@ if ! command -v uv &>/dev/null; then
 fi
 export PATH="$HOME/.local/bin:$PATH"
 
-uv venv ~/venv
+if [ -n "${PYTHON_VERSION:-}" ]; then
+    echo "  Installing Python ${PYTHON_VERSION}..."
+    uv python install "$PYTHON_VERSION"
+    uv venv --python "$PYTHON_VERSION" ~/venv
+else
+    uv venv ~/venv
+fi
 
 # --- CUDA-aware PyTorch installation ---
 # Known PyTorch CUDA wheel tags (ascending order).