Add AMD uProf module and enable it in fox

Reviewed-by: Aleix Roca Nonell <aleix.rocanonell@bsc.es>

flake.lock

@@ -10,11 +10,11 @@
         "systems": "systems"
       },
       "locked": {
-        "lastModified": 1723293904,
-        "narHash": "sha256-b+uqzj+Wa6xgMS9aNbX4I+sXeb5biPDi39VgvSFqFvU=",
+        "lastModified": 1750173260,
+        "narHash": "sha256-9P1FziAwl5+3edkfFcr5HeGtQUtrSdk/MksX39GieoA=",
         "owner": "ryantm",
         "repo": "agenix",
-        "rev": "f6291c5935fdc4e0bef208cfc0dcab7e3f7a1c41",
+        "rev": "531beac616433bac6f9e2a19feb8e99a22a66baf",
         "type": "github"
       },
       "original": {
@@ -30,11 +30,11 @@
         ]
       },
       "locked": {
-        "lastModified": 1732868163,
-        "narHash": "sha256-qck4h298AgcNI6BnGhEwl26MTLXjumuJVr+9kak7uPo=",
+        "lastModified": 1749650500,
+        "narHash": "sha256-2MHfVPV6RA7qPSCtXh4+KK0F0UjN+J4z8//+n6NK7Xs=",
         "ref": "refs/heads/master",
-        "rev": "6782fc6c5b5a29e84a7f2c2d1064f4bcb1288c0f",
-        "revCount": 952,
+        "rev": "9d1944c658929b6f98b3f3803fead4d1b91c4405",
+        "revCount": 961,
         "type": "git",
         "url": "https://git.sr.ht/~rodarima/bscpkgs"
       },
@@ -51,11 +51,11 @@
         ]
       },
       "locked": {
-        "lastModified": 1700795494,
-        "narHash": "sha256-gzGLZSiOhf155FW7262kdHo2YDeugp3VuIFb4/GGng0=",
+        "lastModified": 1744478979,
+        "narHash": "sha256-dyN+teG9G82G+m+PX/aSAagkC+vUv0SgUw3XkPhQodQ=",
         "owner": "lnl7",
         "repo": "nix-darwin",
-        "rev": "4b9b83d5a92e8c1fbfd8eb27eda375908c11ec4d",
+        "rev": "43975d782b418ebf4969e9ccba82466728c2851b",
         "type": "github"
       },
       "original": {
@@ -73,11 +73,11 @@
         ]
       },
       "locked": {
-        "lastModified": 1703113217,
-        "narHash": "sha256-7ulcXOk63TIT2lVDSExj7XzFx09LpdSAPtvgtM7yQPE=",
+        "lastModified": 1745494811,
+        "narHash": "sha256-YZCh2o9Ua1n9uCvrvi5pRxtuVNml8X2a03qIFfRKpFs=",
         "owner": "nix-community",
         "repo": "home-manager",
-        "rev": "3bfaacf46133c037bb356193bd2f1765d9dc82c1",
+        "rev": "abfad3d2958c9e6300a883bd443512c55dfeb1be",
         "type": "github"
       },
       "original": {
@@ -88,16 +88,16 @@
     },
     "nixpkgs": {
       "locked": {
-        "lastModified": 1736867362,
-        "narHash": "sha256-i/UJ5I7HoqmFMwZEH6vAvBxOrjjOJNU739lnZnhUln8=",
+        "lastModified": 1752436162,
+        "narHash": "sha256-Kt1UIPi7kZqkSc5HVj6UY5YLHHEzPBkgpNUByuyxtlw=",
         "owner": "NixOS",
         "repo": "nixpkgs",
-        "rev": "9c6b49aeac36e2ed73a8c472f1546f6d9cf1addc",
+        "rev": "dfcd5b901dbab46c9c6e80b265648481aafb01f8",
         "type": "github"
       },
       "original": {
         "owner": "NixOS",
-        "ref": "nixos-24.11",
+        "ref": "nixos-25.05",
         "repo": "nixpkgs",
         "type": "github"
       }

flake.nix

@@ -1,6 +1,6 @@
 {
   inputs = {
-    nixpkgs.url = "github:NixOS/nixpkgs/nixos-24.11";
+    nixpkgs.url = "github:NixOS/nixpkgs/nixos-25.05";
     agenix.url = "github:ryantm/agenix";
     agenix.inputs.nixpkgs.follows = "nixpkgs";
     bscpkgs.url = "git+https://git.sr.ht/~rodarima/bscpkgs";
@@ -27,6 +27,8 @@ in
       lake2   = mkConf "lake2";
       raccoon = mkConf "raccoon";
       fox     = mkConf "fox";
+      apex    = mkConf "apex";
+      weasel  = mkConf "weasel";
     };
 
     packages.x86_64-linux = self.nixosConfigurations.hut.pkgs // {

keys.nix

@@ -2,25 +2,27 @@
 # here all the public keys
 rec {
   hosts = {
-    hut   = "ssh-ed25519 AAAAC3NzaC1lZDI1NTE5AAAAICO7jIp6JRnRWTMDsTB/aiaICJCl4x8qmKMPSs4lCqP1 hut";
-    owl1  = "ssh-ed25519 AAAAC3NzaC1lZDI1NTE5AAAAIMqMEXO0ApVsBA6yjmb0xP2kWyoPDIWxBB0Q3+QbHVhv owl1";
-    owl2  = "ssh-ed25519 AAAAC3NzaC1lZDI1NTE5AAAAIHurEYpQzNHqWYF6B9Pd7W8UPgF3BxEg0BvSbsA7BAdK owl2";
-    eudy  = "ssh-ed25519 AAAAC3NzaC1lZDI1NTE5AAAAIL+WYPRRvZupqLAG0USKmd/juEPmisyyJaP8hAgYwXsG eudy";
-    koro  = "ssh-ed25519 AAAAC3NzaC1lZDI1NTE5AAAAIImiTFDbxyUYPumvm8C4mEnHfuvtBY1H8undtd6oDd67 koro";
-    bay   = "ssh-ed25519 AAAAC3NzaC1lZDI1NTE5AAAAICvGBzpRQKuQYHdlUQeAk6jmdbkrhmdLwTBqf3el7IgU bay";
-    lake2 = "ssh-ed25519 AAAAC3NzaC1lZDI1NTE5AAAAINo66//S1yatpQHE/BuYD/Gfq64TY7ZN5XOGXmNchiO0 lake2";
-    fox   = "ssh-ed25519 AAAAC3NzaC1lZDI1NTE5AAAAIDwItIk5uOJcQEVPoy/CVGRzfmE1ojrdDcI06FrU4NFT fox";
-    tent  = "ssh-ed25519 AAAAC3NzaC1lZDI1NTE5AAAAIFAtTpHtdYoelbknD/IcfBlThwLKJv/dSmylOgpg3FRM tent";
+    hut    = "ssh-ed25519 AAAAC3NzaC1lZDI1NTE5AAAAICO7jIp6JRnRWTMDsTB/aiaICJCl4x8qmKMPSs4lCqP1 hut";
+    owl1   = "ssh-ed25519 AAAAC3NzaC1lZDI1NTE5AAAAIMqMEXO0ApVsBA6yjmb0xP2kWyoPDIWxBB0Q3+QbHVhv owl1";
+    owl2   = "ssh-ed25519 AAAAC3NzaC1lZDI1NTE5AAAAIHurEYpQzNHqWYF6B9Pd7W8UPgF3BxEg0BvSbsA7BAdK owl2";
+    eudy   = "ssh-ed25519 AAAAC3NzaC1lZDI1NTE5AAAAIL+WYPRRvZupqLAG0USKmd/juEPmisyyJaP8hAgYwXsG eudy";
+    koro   = "ssh-ed25519 AAAAC3NzaC1lZDI1NTE5AAAAIImiTFDbxyUYPumvm8C4mEnHfuvtBY1H8undtd6oDd67 koro";
+    bay    = "ssh-ed25519 AAAAC3NzaC1lZDI1NTE5AAAAICvGBzpRQKuQYHdlUQeAk6jmdbkrhmdLwTBqf3el7IgU bay";
+    lake2  = "ssh-ed25519 AAAAC3NzaC1lZDI1NTE5AAAAINo66//S1yatpQHE/BuYD/Gfq64TY7ZN5XOGXmNchiO0 lake2";
+    fox    = "ssh-ed25519 AAAAC3NzaC1lZDI1NTE5AAAAIDwItIk5uOJcQEVPoy/CVGRzfmE1ojrdDcI06FrU4NFT fox";
+    tent   = "ssh-ed25519 AAAAC3NzaC1lZDI1NTE5AAAAIFAtTpHtdYoelbknD/IcfBlThwLKJv/dSmylOgpg3FRM tent";
+    apex   = "ssh-ed25519 AAAAC3NzaC1lZDI1NTE5AAAAIBvUFjSfoxXnKwXhEFXx5ckRKJ0oewJ82mRitSMNMKjh apex";
+    weasel = "ssh-ed25519 AAAAC3NzaC1lZDI1NTE5AAAAIFLJrQ8BF6KcweQV8pLkSbFT+tbDxSG9qxrdQE65zJZp weasel";
   };
 
   hostGroup = with hosts; rec {
-    untrusted  = [ fox ];
-    compute    = [ owl1 owl2 ];
-    playground = [ eudy koro ];
+    compute    = [ owl1 owl2 fox ];
+    playground = [ eudy koro weasel ];
     storage    = [ bay lake2 ];
     monitor    = [ hut ];
+    login      = [ apex ];
 
-    system     = storage ++ monitor;
+    system     = storage ++ monitor ++ login;
     safe       = system ++ compute;
     all        = safe ++ playground;
   };
@@ -28,6 +30,7 @@ rec {
   admins = {
     "rarias@hut"  = "ssh-ed25519 AAAAC3NzaC1lZDI1NTE5AAAAIE1oZTPtlEXdGt0Ak+upeCIiBdaDQtcmuWoTUCVuSVIR rarias@hut";
     "rarias@tent" = "ssh-ed25519 AAAAC3NzaC1lZDI1NTE5AAAAIIwlWSBTZi74WTz5xn6gBvTmCoVltmtIAeM3RMmkh4QZ rarias@tent";
+    "rarias@fox"  = "ssh-ed25519 AAAAC3NzaC1lZDI1NTE5AAAAIDSbw3REAKECV7E2c/e2XJITudJQWq2qDSe2N1JHqHZd rarias@fox";
     root          = "ssh-ed25519 AAAAC3NzaC1lZDI1NTE5AAAAIII/1TNArcwA6D47mgW4TArwlxQRpwmIGiZDysah40Gb root@hut";
   };
 }

m/apex/configuration.nix

@@ -0,0 +1,85 @@
+{ lib, config, pkgs, ... }:
+
+{
+  imports = [
+    ../common/xeon.nix
+    ../common/ssf/hosts.nix
+    ../module/ceph.nix
+    ../module/slurm-server.nix
+    ./nfs.nix
+    ./wireguard.nix
+  ];
+
+  # Don't install grub MBR for now
+  boot.loader.grub.device = "nodev";
+
+  boot.initrd.kernelModules = [
+    "megaraid_sas" # For HW RAID
+  ];
+
+  environment.systemPackages = with pkgs; [
+    storcli # To manage HW RAID
+  ];
+
+  fileSystems."/home" = {
+    device = "/dev/disk/by-label/home";
+    fsType = "ext4";
+  };
+
+  # No swap, there is plenty of RAM
+  swapDevices = lib.mkForce [];
+
+  networking = {
+    hostName = "apex";
+    defaultGateway = "84.88.53.233";
+    nameservers = [ "8.8.8.8" ];
+
+    # Public facing interface
+    interfaces.eno1.ipv4.addresses = [ {
+      address = "84.88.53.236";
+      prefixLength = 29;
+    } ];
+
+    # Internal LAN to our Ethernet switch
+    interfaces.eno2.ipv4.addresses = [ {
+      address = "10.0.40.30";
+      prefixLength = 24;
+    } ];
+
+    # Infiniband over Omnipath switch (disconnected for now)
+    # interfaces.ibp5s0 = {};
+
+    nat = {
+      enable = true;
+      internalInterfaces = [ "eno2" ];
+      externalInterface = "eno1";
+    };
+  };
+
+  # Use SSH tunnel to reach internal hosts
+  programs.ssh.extraConfig = ''
+    Host bscpm04.bsc.es gitlab-internal.bsc.es knights3.bsc.es
+      ProxyCommand nc -X connect -x localhost:23080 %h %p
+    Host raccoon
+      HostName knights3.bsc.es
+      ProxyCommand nc -X connect -x localhost:23080 %h %p
+    Host tent
+      ProxyJump raccoon
+  '';
+
+  networking.firewall = {
+    extraCommands = ''
+      # Blackhole BSC vulnerability scanner (OpenVAS) as it is spamming our
+      # logs. Insert as first position so we also protect SSH.
+      iptables -I nixos-fw 1 -p tcp -s 192.168.8.16 -j nixos-fw-refuse
+      # Same with opsmonweb01.bsc.es which seems to be trying to access via SSH
+      iptables -I nixos-fw 2 -p tcp -s 84.88.52.176 -j nixos-fw-refuse
+    '';
+  };
+
+  # Use tent for cache
+  nix.settings = {
+    extra-substituters = [ "https://jungle.bsc.es/cache" ];
+    extra-trusted-public-keys = [ "jungle.bsc.es:pEc7MlAT0HEwLQYPtpkPLwRsGf80ZI26aj29zMw/HH0=" ];
+  };
+}

m/apex/nfs.nix

@@ -0,0 +1,48 @@
+{ ... }:
+
+{
+  services.nfs.server = {
+    enable = true;
+    lockdPort = 4001;
+    mountdPort = 4002;
+    statdPort = 4000;
+    exports = ''
+      /home 10.0.40.0/24(rw,async,no_subtree_check,no_root_squash)
+      /home 10.106.0.0/24(rw,async,no_subtree_check,no_root_squash)
+    '';
+  };
+  networking.firewall = {
+    # Check with `rpcinfo -p`
+    extraCommands = ''
+      # Accept NFS traffic from compute nodes but not from the outside
+      iptables -A nixos-fw -p tcp -s 10.0.40.0/24 --dport 111   -j nixos-fw-accept
+      iptables -A nixos-fw -p tcp -s 10.0.40.0/24 --dport 2049  -j nixos-fw-accept
+      iptables -A nixos-fw -p tcp -s 10.0.40.0/24 --dport 4000  -j nixos-fw-accept
+      iptables -A nixos-fw -p tcp -s 10.0.40.0/24 --dport 4001  -j nixos-fw-accept
+      iptables -A nixos-fw -p tcp -s 10.0.40.0/24 --dport 4002  -j nixos-fw-accept
+      iptables -A nixos-fw -p tcp -s 10.0.40.0/24 --dport 20048 -j nixos-fw-accept
+      # Same but UDP
+      iptables -A nixos-fw -p udp -s 10.0.40.0/24 --dport 111   -j nixos-fw-accept
+      iptables -A nixos-fw -p udp -s 10.0.40.0/24 --dport 2049  -j nixos-fw-accept
+      iptables -A nixos-fw -p udp -s 10.0.40.0/24 --dport 4000  -j nixos-fw-accept
+      iptables -A nixos-fw -p udp -s 10.0.40.0/24 --dport 4001  -j nixos-fw-accept
+      iptables -A nixos-fw -p udp -s 10.0.40.0/24 --dport 4002  -j nixos-fw-accept
+      iptables -A nixos-fw -p udp -s 10.0.40.0/24 --dport 20048 -j nixos-fw-accept
+
+      # Accept NFS traffic from wg0
+      iptables -A nixos-fw -p tcp -i wg0 -s 10.106.0.0/24 --dport 111   -j nixos-fw-accept
+      iptables -A nixos-fw -p tcp -i wg0 -s 10.106.0.0/24 --dport 2049  -j nixos-fw-accept
+      iptables -A nixos-fw -p tcp -i wg0 -s 10.106.0.0/24 --dport 4000  -j nixos-fw-accept
+      iptables -A nixos-fw -p tcp -i wg0 -s 10.106.0.0/24 --dport 4001  -j nixos-fw-accept
+      iptables -A nixos-fw -p tcp -i wg0 -s 10.106.0.0/24 --dport 4002  -j nixos-fw-accept
+      iptables -A nixos-fw -p tcp -i wg0 -s 10.106.0.0/24 --dport 20048 -j nixos-fw-accept
+      # Same but UDP
+      iptables -A nixos-fw -p udp -i wg0 -s 10.106.0.0/24 --dport 111   -j nixos-fw-accept
+      iptables -A nixos-fw -p udp -i wg0 -s 10.106.0.0/24 --dport 2049  -j nixos-fw-accept
+      iptables -A nixos-fw -p udp -i wg0 -s 10.106.0.0/24 --dport 4000  -j nixos-fw-accept
+      iptables -A nixos-fw -p udp -i wg0 -s 10.106.0.0/24 --dport 4001  -j nixos-fw-accept
+      iptables -A nixos-fw -p udp -i wg0 -s 10.106.0.0/24 --dport 4002  -j nixos-fw-accept
+      iptables -A nixos-fw -p udp -i wg0 -s 10.106.0.0/24 --dport 20048 -j nixos-fw-accept
+    '';
+  };
+}

m/apex/wireguard.nix

@@ -0,0 +1,35 @@
+{ config, ... }:
+
+{
+  networking.firewall = {
+    allowedUDPPorts = [ 666 ];
+  };
+
+  age.secrets.wgApex.file = ../../secrets/wg-apex.age;
+
+  # Enable WireGuard
+  networking.wireguard.enable = true;
+  networking.wireguard.interfaces = {
+    # "wg0" is the network interface name. You can name the interface arbitrarily.
+    wg0 = {
+      ips = [ "10.106.0.30/24" ];
+      listenPort = 666;
+      privateKeyFile = config.age.secrets.wgApex.path;
+      # Public key: VwhcN8vSOzdJEotQTpmPHBC52x3Hbv1lkFIyKubrnUA=
+      peers = [
+        {
+          name = "Fox";
+          publicKey = "VfMPBQLQTKeyXJSwv8wBhc6OV0j2qAxUpX3kLHunK2Y=";
+          allowedIPs = [ "10.106.0.0/24" ];
+          endpoint = "fox.ac.upc.edu:666";
+          # Send keepalives every 25 seconds. Important to keep NAT tables alive.
+          persistentKeepalive = 25;
+        }
+      ];
+    };
+  };
+
+  networking.hosts = {
+    "10.106.0.1" = [ "fox" ];
+  };
+}

m/common/base.nix

@@ -3,6 +3,7 @@
   # Includes the basic configuration for an Intel server.
   imports = [
     ./base/agenix.nix
+    ./base/always-power-on.nix
     ./base/august-shutdown.nix
     ./base/boot.nix
     ./base/env.nix

m/common/base/always-power-on.nix

@@ -0,0 +1,8 @@
+{
+  imports = [
+    ../../module/power-policy.nix
+  ];
+
+  # Turn on as soon as we have power
+  power.policy = "always-on";
+}

m/common/base/august-shutdown.nix

@@ -1,12 +1,12 @@
 {
-  # Shutdown all machines on August 2nd at 11:00 AM, so we can protect the
+  # Shutdown all machines on August 3rd at 22:00, so we can protect the
   # hardware from spurious electrical peaks on the yearly electrical cut for
   # manteinance that starts on August 4th.
   systemd.timers.august-shutdown = {
-    description = "Shutdown on August 2nd for maintenance";
+    description = "Shutdown on August 3rd for maintenance";
     wantedBy = [ "timers.target" ];
     timerConfig = {
-      OnCalendar = "*-08-02 11:00:00";
+      OnCalendar = "*-08-03 22:00:00";
       RandomizedDelaySec = "10min";
       Unit = "systemd-poweroff.service";
     };

m/common/base/env.nix

@@ -3,8 +3,8 @@
 {
   environment.systemPackages = with pkgs; [
     vim wget git htop tmux pciutils tcpdump ripgrep nix-index nixos-option
-    nix-diff ipmitool freeipmi ethtool lm_sensors ix cmake gnumake file tree
-    ncdu config.boot.kernelPackages.perf ldns
+    nix-diff ipmitool freeipmi ethtool lm_sensors cmake gnumake file tree
+    ncdu config.boot.kernelPackages.perf ldns pv
     # From bsckgs overlay
     osumb
   ];
@@ -21,6 +21,8 @@
     }
   ];
 
+  environment.enableAllTerminfo = true;
+
   environment.variables = {
     EDITOR = "vim";
     VISUAL = "vim";

m/common/base/net.nix

@@ -1,4 +1,4 @@
-{ pkgs, ... }:
+{ pkgs, lib, ... }:
 
 {
   networking = {
@@ -10,6 +10,9 @@
       allowedTCPPorts = [ 22 ];
     };
 
+    # Make sure we use iptables
+    nftables.enable = lib.mkForce false;
+
     hosts = {
       "84.88.53.236" = [ "ssfhead.bsc.es" "ssfhead" ];
       "84.88.51.152" = [ "raccoon" ];

m/common/base/nix.nix

@@ -6,6 +6,8 @@
     (import ../../../pkgs/overlay.nix)
   ];
 
+  nixpkgs.config.allowUnfree = true;
+
   nix = {
     nixPath = [
       "nixpkgs=${nixpkgs}"

m/common/base/users.nix

@@ -56,7 +56,7 @@
         home = "/home/Computational/rpenacob";
         description = "Raúl Peñacoba";
         group = "Computational";
-        hosts = [ "owl1" "owl2" "hut" "tent" ];
+        hosts = [ "apex" "owl1" "owl2" "hut" "tent" "fox" ];
         hashedPassword = "$6$TZm3bDIFyPrMhj1E$uEDXoYYd1z2Wd5mMPfh3DZAjP7ztVjJ4ezIcn82C0ImqafPA.AnTmcVftHEzLB3tbe2O4SxDyPSDEQgJ4GOtj/";
         openssh.authorizedKeys.keys = [
           "ssh-ed25519 AAAAC3NzaC1lZDI1NTE5AAAAIFYfXg37mauGeurqsLpedgA2XQ9d4Nm0ZGo/hI1f7wwH rpenacob@bsc"
@@ -69,10 +69,10 @@
         home = "/home/Computational/anavarro";
         description = "Antoni Navarro";
         group = "Computational";
-        hosts = [ "hut" "tent" "raccoon" ];
-        hashedPassword = "$6$QdNDsuLehoZTYZlb$CDhCouYDPrhoiB7/seu7RF.Gqg4zMQz0n5sA4U1KDgHaZOxy2as9pbIGeF8tOHJKRoZajk5GiaZv0rZMn7Oq31";
+        hosts = [ "apex" "hut" "tent" "raccoon" "fox" "weasel" ];
+        hashedPassword = "$6$EgturvVYXlKgP43g$gTN78LLHIhaF8hsrCXD.O6mKnZSASWSJmCyndTX8QBWT6wTlUhcWVAKz65lFJPXjlJA4u7G1ydYQ0GG6Wk07b1";
         openssh.authorizedKeys.keys = [
-          "ssh-ed25519 AAAAC3NzaC1lZDI1NTE5AAAAILWjRSlKgzBPZQhIeEtk6Lvws2XNcYwHcwPv4osSgst5 anavarro@ssfhead"
+          "ssh-ed25519 AAAAC3NzaC1lZDI1NTE5AAAAIMsbM21uepnJwPrRe6jYFz8zrZ6AYMtSEvvt4c9spmFP toni@delltoni"
         ];
       };
 
@@ -82,7 +82,7 @@
         home = "/home/Computational/abonerib";
         description = "Aleix Boné";
         group = "Computational";
-        hosts = [ "owl1" "owl2" "hut" "tent" "raccoon" ];
+        hosts = [ "apex" "owl1" "owl2" "hut" "tent" "raccoon" "fox" "weasel" ];
         hashedPassword = "$6$V1EQWJr474whv7XJ$OfJ0wueM2l.dgiJiiah0Tip9ITcJ7S7qDvtSycsiQ43QBFyP4lU0e0HaXWps85nqB4TypttYR4hNLoz3bz662/";
         openssh.authorizedKeys.keys = [
           "ssh-ed25519 AAAAC3NzaC1lZDI1NTE5AAAAIIIFiqXqt88VuUfyANkZyLJNiuroIITaGlOOTMhVDKjf abonerib@bsc"
@@ -95,7 +95,7 @@
         home = "/home/Computational/vlopez";
         description = "Victor López";
         group = "Computational";
-        hosts = [ "koro" ];
+        hosts = [ "apex" "koro" ];
         hashedPassword = "$6$0ZBkgIYE/renVqtt$1uWlJsb0FEezRVNoETTzZMx4X2SvWiOsKvi0ppWCRqI66S6TqMBXBdP4fcQyvRRBt0e4Z7opZIvvITBsEtO0f0";
         openssh.authorizedKeys.keys = [
           "ssh-ed25519 AAAAC3NzaC1lZDI1NTE5AAAAIGMwlUZRf9jfG666Qa5Sb+KtEhXqkiMlBV2su3x/dXHq victor@arch"
@@ -108,7 +108,7 @@
         home = "/home/Computational/dbautist";
         description = "Dylan Bautista Cases";
         group = "Computational";
-        hosts = [ "hut" "tent" "raccoon" ];
+        hosts = [ "apex" "hut" "tent" "raccoon" ];
         hashedPassword = "$6$a2lpzMRVkG9nSgIm$12G6.ka0sFX1YimqJkBAjbvhRKZ.Hl090B27pdbnQOW0wzyxVWySWhyDDCILjQELky.HKYl9gqOeVXW49nW7q/";
         openssh.authorizedKeys.keys = [
           "ssh-ed25519 AAAAC3NzaC1lZDI1NTE5AAAAIAb+EQBoS98zrCwnGKkHKwMLdYABMTqv7q9E0+T0QmkS dbautist@bsc-848818791"
@@ -121,7 +121,7 @@
         home = "/home/Computational/dalvare1";
         description = "David Álvarez";
         group = "Computational";
-        hosts = [ "hut" "tent" ];
+        hosts = [ "apex" "hut" "tent" "fox" ];
         hashedPassword = "$6$mpyIsV3mdq.rK8$FvfZdRH5OcEkUt5PnIUijWyUYZvB1SgeqxpJ2p91TTe.3eQIDTcLEQ5rxeg.e5IEXAZHHQ/aMsR5kPEujEghx0";
         openssh.authorizedKeys.keys = [
           "ssh-ed25519 AAAAC3NzaC1lZDI1NTE5AAAAIGEfy6F4rF80r4Cpo2H5xaWqhuUZzUsVsILSKGJzt5jF dalvare1@ssfhead"
@@ -134,12 +134,40 @@
         home = "/home/Computational/varcila";
         description = "Vincent Arcila";
         group = "Computational";
-        hosts = [ "hut" "tent" "fox" ];
+        hosts = [ "apex" "hut" "tent" "fox" ];
         hashedPassword = "$6$oB0Tcn99DcM4Ch$Vn1A0ulLTn/8B2oFPi9wWl/NOsJzaFAWjqekwcuC9sMC7cgxEVb.Nk5XSzQ2xzYcNe5MLtmzkVYnRS1CqP39Y0";
         openssh.authorizedKeys.keys = [
           "ssh-ed25519 AAAAC3NzaC1lZDI1NTE5AAAAIKGt0ESYxekBiHJQowmKpfdouw0hVm3N7tUMtAaeLejK vincent@varch"
         ];
       };
+
+      pmartin1 = {
+        # Arbitrary UID but large so it doesn't collide with other users on ssfhead.
+        uid = 9652;
+        isNormalUser = true;
+        home = "/home/Computational/pmartin1";
+        description = "Pedro J. Martinez-Ferrer";
+        group = "Computational";
+        hosts = [ "fox" ];
+        hashedPassword = "$6$nIgDMGnt4YIZl3G.$.JQ2jXLtDPRKsbsJfJAXdSvjDIzRrg7tNNjPkLPq3KJQhMjfDXRUvzagUHUU2TrE2hHM8/6uq8ex0UdxQ0ysl.";
+        openssh.authorizedKeys.keys = [
+          "ssh-ed25519 AAAAC3NzaC1lZDI1NTE5AAAAIIV5LEAII5rfe1hYqDYIIrhb1gOw7RcS1p2mhOTqG+zc pedro@pedro-ThinkPad-P14s-Gen-2a"
+        ];
+      };
+
+      csiringo = {
+        # Arbitrary UID but large so it doesn't collide with other users on ssfhead.
+        uid = 9653;
+        isNormalUser = true;
+        home = "/home/Computational/csiringo";
+        description = "Cesare Siringo";
+        group = "Computational";
+        hosts = [ "apex" "weasel" ];
+        hashedPassword = "$6$0IsZlju8jFukLlAw$VKm0FUXbS.mVmPm3rcJeizTNU4IM5Nmmy21BvzFL.cQwvlGwFI1YWRQm6gsbd4nbg47mPDvYkr/ar0SlgF6GO1";
+        openssh.authorizedKeys.keys = [
+          "ssh-ed25519 AAAAC3NzaC1lZDI1NTE5AAAAIHA65zvvG50iuFEMf+guRwZB65jlGXfGLF4HO+THFaed csiringo@bsc.es"
+        ];
+      };
     };
 
     groups = {

m/common/ssf.nix

@@ -3,6 +3,7 @@
   imports = [
     ./xeon.nix
     ./ssf/fs.nix
+    ./ssf/hosts.nix
     ./ssf/net.nix
     ./ssf/ssh.nix
   ];

m/common/ssf/hosts.nix

@@ -0,0 +1,23 @@
+{ pkgs, ... }:
+
+{
+  networking.hosts = {
+    # Login
+    "10.0.40.30" = [ "apex" ];
+
+    # Storage
+    "10.0.40.40" = [ "bay" ];   "10.0.42.40" = [ "bay-ib" ];    "10.0.40.141" = [ "bay-ipmi" ];
+    "10.0.40.41" = [ "oss01" ]; "10.0.42.41" = [ "oss01-ib0" ]; "10.0.40.142" = [ "oss01-ipmi" ];
+    "10.0.40.42" = [ "lake2" ]; "10.0.42.42" = [ "lake2-ib" ];  "10.0.40.143" = [ "lake2-ipmi" ];
+
+    # Xeon compute
+    "10.0.40.1" = [ "owl1" ];   "10.0.42.1" = [ "owl1-ib" ];   "10.0.40.101" = [ "owl1-ipmi" ];
+    "10.0.40.2" = [ "owl2" ];   "10.0.42.2" = [ "owl2-ib" ];   "10.0.40.102" = [ "owl2-ipmi" ];
+    "10.0.40.3" = [ "xeon03" ]; "10.0.42.3" = [ "xeon03-ib" ]; "10.0.40.103" = [ "xeon03-ipmi" ];
+    #"10.0.40.4" = [ "tent" ];   "10.0.42.4" = [ "tent-ib" ];   "10.0.40.104" = [ "tent-ipmi" ];
+    "10.0.40.5" = [ "koro" ];   "10.0.42.5" = [ "koro-ib" ];   "10.0.40.105" = [ "koro-ipmi" ];
+    "10.0.40.6" = [ "weasel" ]; "10.0.42.6" = [ "weasel-ib" ]; "10.0.40.106" = [ "weasel-ipmi" ];
+    "10.0.40.7" = [ "hut" ];    "10.0.42.7" = [ "hut-ib" ];    "10.0.40.107" = [ "hut-ipmi" ];
+    "10.0.40.8" = [ "eudy" ];   "10.0.42.8" = [ "eudy-ib" ];   "10.0.40.108" = [ "eudy-ipmi" ];
+  };
+}

m/common/ssf/net.nix

@@ -9,14 +9,6 @@
     defaultGateway = "10.0.40.30";
     nameservers = ["8.8.8.8"];
 
-    proxy = {
-      default = "http://hut:23080/";
-      noProxy = "127.0.0.1,localhost,internal.domain,10.0.40.40,hut";
-      # Don't set all_proxy as go complains and breaks the gitlab runner, see:
-      # https://github.com/golang/go/issues/16715
-      allProxy = null;
-    };
-
     firewall = {
       extraCommands = ''
         # Prevent ssfhead from contacting our slurmd daemon
@@ -27,64 +19,5 @@
         iptables -A nixos-fw -p tcp -s 10.0.40.0/24 --dport 60000:61000 -j nixos-fw-accept
       '';
     };
-
-    extraHosts = ''
-      10.0.40.30              ssfhead
-      
-      # Node Entry for node: mds01 (ID=72)
-      10.0.40.40              bay mds01 mds01-eth0
-      10.0.42.40              bay-ib mds01-ib0
-      10.0.40.141             bay-ipmi mds01-ipmi0 mds01-ipmi
-      
-      # Node Entry for node: oss01 (ID=73)
-      10.0.40.41              oss01 oss01-eth0
-      10.0.42.41              oss01-ib0
-      10.0.40.142             oss01-ipmi0 oss01-ipmi
-      
-      # Node Entry for node: oss02 (ID=74)
-      10.0.40.42              lake2 oss02 oss02-eth0
-      10.0.42.42              lake2-ib oss02-ib0
-      10.0.40.143             lake2-ipmi oss02-ipmi0 oss02-ipmi
-      
-      # Node Entry for node: xeon01 (ID=15)
-      10.0.40.1               owl1 xeon01 xeon01-eth0
-      10.0.42.1               owl1-ib xeon01-ib0
-      10.0.40.101             owl1-ipmi xeon01-ipmi0 xeon01-ipmi
-      
-      # Node Entry for node: xeon02 (ID=16)
-      10.0.40.2               owl2 xeon02 xeon02-eth0
-      10.0.42.2               owl2-ib xeon02-ib0
-      10.0.40.102             owl2-ipmi xeon02-ipmi0 xeon02-ipmi
-      
-      # Node Entry for node: xeon03 (ID=17)
-      10.0.40.3               xeon03 xeon03-eth0
-      10.0.42.3               xeon03-ib0
-      10.0.40.103             xeon03-ipmi0 xeon03-ipmi
-      
-      # Node Entry for node: xeon04 (ID=18)
-      10.0.40.4               xeon04 xeon04-eth0
-      10.0.42.4               xeon04-ib0
-      10.0.40.104             xeon04-ipmi0 xeon04-ipmi
-      
-      # Node Entry for node: xeon05 (ID=19)
-      10.0.40.5               koro xeon05 xeon05-eth0
-      10.0.42.5               koro-ib xeon05-ib0
-      10.0.40.105             koro-ipmi xeon05-ipmi0
-      
-      # Node Entry for node: xeon06 (ID=20)
-      10.0.40.6               xeon06 xeon06-eth0
-      10.0.42.6               xeon06-ib0
-      10.0.40.106             xeon06-ipmi0 xeon06-ipmi
-      
-      # Node Entry for node: xeon07 (ID=21)
-      10.0.40.7               hut xeon07 xeon07-eth0
-      10.0.42.7               hut-ib xeon07-ib0
-      10.0.40.107             hut-ipmi xeon07-ipmi0 xeon07-ipmi
-      
-      # Node Entry for node: xeon08 (ID=22)
-      10.0.40.8               eudy xeon08 xeon08-eth0
-      10.0.42.8               eudy-ib xeon08-ib0
-      10.0.40.108             eudy-ipmi xeon08-ipmi0 xeon08-ipmi
-    '';
   };
 }

m/common/ssf/ssh.nix

@@ -1,8 +1,16 @@
 {
-  # Connect to intranet git hosts via proxy
+  # Use SSH tunnel to apex to reach internal hosts
   programs.ssh.extraConfig = ''
-    # Connect to BSC machines via hut proxy too
-    Host amdlogin1.bsc.es armlogin1.bsc.es hualogin1.bsc.es glogin1.bsc.es glogin2.bsc.es fpgalogin1.bsc.es
-      ProxyCommand nc -X connect -x hut:23080 %h %p
+    Host tent
+      ProxyJump raccoon
+
+    # Access raccoon via the HTTP proxy
+    Host raccoon knights3.bsc.es
+      HostName knights3.bsc.es
+      ProxyCommand=ssh apex 'nc -X connect -x localhost:23080 %h %p'
+
+    # Make sure we can reach gitlab even if we don't have SSH access to raccoon
+    Host bscpm04.bsc.es gitlab-internal.bsc.es
+      ProxyCommand=ssh apex 'nc -X connect -x localhost:23080 %h %p'
   '';
 }

m/fox/configuration.nix

@@ -4,10 +4,17 @@
   imports = [
     ../common/base.nix
     ../common/xeon/console.nix
-    ../module/emulation.nix
     ../module/amd-uprof.nix
+    ../module/emulation.nix
+    ../module/nvidia.nix
+    ../module/slurm-client.nix
+    ./wireguard.nix
   ];
 
+  # Don't turn off on August as UPC has different dates.
+  # Fox works fine on power cuts.
+  systemd.timers.august-shutdown.enable = false;
+
   # Select the this using the ID to avoid mismatches
   boot.loader.grub.device = "/dev/disk/by-id/wwn-0x500a07514b0c1103";
 
@@ -15,7 +22,7 @@
   swapDevices = lib.mkForce [];
 
   boot.initrd.availableKernelModules = [ "xhci_pci" "ahci" "nvme" "usbhid" "usb_storage" "sd_mod" ];
-  boot.kernelModules = [ "kvm-amd" ];
+  boot.kernelModules = [ "kvm-amd" "amd_uncore" ];
 
   hardware.cpu.amd.updateMicrocode = lib.mkDefault config.hardware.enableRedistributableFirmware;
   hardware.cpu.intel.updateMicrocode = lib.mkForce false;
@@ -25,6 +32,26 @@
 
   services.amd-uprof.enable = true;
 
+  # Disable NUMA balancing
+  boot.kernel.sysctl."kernel.numa_balancing" = 0;
+
+  # Expose kernel addresses
+  boot.kernel.sysctl."kernel.kptr_restrict" = 0;
+
+  services.openssh.settings.X11Forwarding = true;
+
+  services.fail2ban.enable = true;
+
+  # Use SSH tunnel to reach internal hosts
+  programs.ssh.extraConfig = ''
+    Host bscpm04.bsc.es gitlab-internal.bsc.es tent
+      ProxyJump raccoon
+    Host raccoon
+      ProxyJump apex
+      HostName 127.0.0.1
+      Port 22022
+  '';
+
   networking = {
     timeServers = [ "ntp1.upc.edu" "ntp2.upc.edu" ];
     hostName = "fox";
@@ -48,17 +75,20 @@
     extra-trusted-public-keys = [ "jungle.bsc.es:pEc7MlAT0HEwLQYPtpkPLwRsGf80ZI26aj29zMw/HH0=" ];
   };
 
-  # Configure Nvidia driver to use with CUDA
-  hardware.nvidia.package = config.boot.kernelPackages.nvidiaPackages.production;
-  hardware.graphics.enable = true;
-  nixpkgs.config.allowUnfree = true;
-  nixpkgs.config.nvidia.acceptLicense = true;
-  services.xserver.videoDrivers = [ "nvidia" ];
+  # Recommended for new graphics cards
+  hardware.nvidia.open = true;
 
   # Mount NVME disks
   fileSystems."/nvme0" = { device = "/dev/disk/by-label/nvme0"; fsType = "ext4"; };
   fileSystems."/nvme1" = { device = "/dev/disk/by-label/nvme1"; fsType = "ext4"; };
 
+  # Mount the NFS home
+  fileSystems."/nfs/home" = {
+    device = "10.106.0.30:/home";
+    fsType = "nfs";
+    options = [ "nfsvers=3" "rsize=1024" "wsize=1024" "cto" "nofail" ];
+  };
+
   # Make a /nvme{0,1}/$USER directory for each user.
   systemd.services.create-nvme-dirs = let
     # Take only normal users in fox
@@ -75,4 +105,20 @@
     wantedBy = [ "multi-user.target" ];
     serviceConfig.ExecStart = script;
   };
+
+  # Only allow SSH connections from users who have a SLURM allocation
+  # See: https://slurm.schedmd.com/pam_slurm_adopt.html
+  security.pam.services.sshd.rules.account.slurm = {
+    control = "required";
+    enable = true;
+    modulePath = "${pkgs.slurm}/lib/security/pam_slurm_adopt.so";
+    args = [ "log_level=debug5" ];
+    order = 999999; # Make it last one
+  };
+
+  # Disable systemd session (pam_systemd.so) as it will conflict with the
+  # pam_slurm_adopt.so module. What happens is that the shell is first adopted
+  # into the slurmstepd task and then into the systemd session, which is not
+  # what we want, otherwise it will linger even if all jobs are gone.
+  security.pam.services.sshd.startSession = lib.mkForce false;
 }

m/fox/wireguard.nix

@@ -0,0 +1,46 @@
+{ config, ... }:
+
+{
+  networking.firewall = {
+    allowedUDPPorts = [ 666 ];
+  };
+
+  age.secrets.wgFox.file = ../../secrets/wg-fox.age;
+
+  networking.wireguard.enable = true;
+  networking.wireguard.interfaces = {
+    # "wg0" is the network interface name. You can name the interface arbitrarily.
+    wg0 = {
+      # Determines the IP address and subnet of the server's end of the tunnel interface.
+      ips = [ "10.106.0.1/24" ];
+
+      # The port that WireGuard listens to. Must be accessible by the client.
+      listenPort = 666;
+
+      # Path to the private key file.
+      privateKeyFile = config.age.secrets.wgFox.path;
+      # Public key: VfMPBQLQTKeyXJSwv8wBhc6OV0j2qAxUpX3kLHunK2Y=
+
+      peers = [
+        # List of allowed peers.
+        { 
+          name = "Apex";
+          publicKey = "VwhcN8vSOzdJEotQTpmPHBC52x3Hbv1lkFIyKubrnUA=";
+          # List of IPs assigned to this peer within the tunnel subnet. Used to configure routing.
+          allowedIPs = [ "10.106.0.30/32" ];
+        }
+      ];
+    };
+  };
+
+  networking.hosts = {
+    "10.106.0.30" = [ "apex" ];
+  };
+
+  networking.firewall = {
+    extraCommands = ''
+      # Accept slurm connections to slurmd from apex (via wireguard)
+      iptables -A nixos-fw -p tcp -i wg0 -s 10.106.0.30/32 -d 10.106.0.1/32 --dport 6818 -j nixos-fw-accept
+    '';
+  };
+}

m/hut/blackbox.yml

@@ -3,160 +3,12 @@ modules:
     prober: http
     timeout: 5s
     http:
-      proxy_url: "http://127.0.0.1:23080"
-      skip_resolve_phase_with_proxy: true
-      follow_redirects: true
-      valid_status_codes: []  # Defaults to 2xx
-      method: GET
-  http_with_proxy:
-    prober: http
-    http:
-      proxy_url: "http://127.0.0.1:3128"
-      skip_resolve_phase_with_proxy: true
-  http_with_proxy_and_headers:
-    prober: http
-    http:
-      proxy_url: "http://127.0.0.1:3128"
-      proxy_connect_header:
-        Proxy-Authorization:
-          - Bearer token
-  http_post_2xx:
-    prober: http
-    timeout: 5s
-    http:
-      method: POST
-      headers:
-        Content-Type: application/json
-      body: '{}'
-  http_post_body_file:
-    prober: http
-    timeout: 5s
-    http:
-      method: POST
-      body_file: "/files/body.txt"
-  http_basic_auth_example:
-    prober: http
-    timeout: 5s
-    http:
-      method: POST
-      headers:
-        Host: "login.example.com"
-      basic_auth:
-        username: "username"
-        password: "mysecret"
-  http_2xx_oauth_client_credentials:
-    prober: http
-    timeout: 5s
-    http:
-      valid_http_versions: ["HTTP/1.1", "HTTP/2"]
       follow_redirects: true
       preferred_ip_protocol: "ip4"
-      valid_status_codes:
-        - 200
-        - 201
-      oauth2:
-        client_id: "client_id"
-        client_secret: "client_secret"
-        token_url: "https://api.example.com/token"
-        endpoint_params:
-          grant_type: "client_credentials"
-  http_custom_ca_example:
-    prober: http
-    http:
+      valid_status_codes: []  # Defaults to 2xx
       method: GET
-      tls_config:
-        ca_file: "/certs/my_cert.crt"
-  http_gzip:
-    prober: http
-    http:
-      method: GET
-      compression: gzip
-  http_gzip_with_accept_encoding:
-    prober: http
-    http:
-      method: GET
-      compression: gzip
-      headers:
-        Accept-Encoding: gzip
-  tls_connect:
-    prober: tcp
-    timeout: 5s
-    tcp:
-      tls: true
-  tcp_connect_example:
-    prober: tcp
-    timeout: 5s
-  imap_starttls:
-    prober: tcp
-    timeout: 5s
-    tcp:
-      query_response:
-        - expect: "OK.*STARTTLS"
-        - send: ". STARTTLS"
-        - expect: "OK"
-        - starttls: true
-        - send: ". capability"
-        - expect: "CAPABILITY IMAP4rev1"
-  smtp_starttls:
-    prober: tcp
-    timeout: 5s
-    tcp:
-      query_response:
-        - expect: "^220 ([^ ]+) ESMTP (.+)$"
-        - send: "EHLO prober\r"
-        - expect: "^250-STARTTLS"
-        - send: "STARTTLS\r"
-        - expect: "^220"
-        - starttls: true
-        - send: "EHLO prober\r"
-        - expect: "^250-AUTH"
-        - send: "QUIT\r"
-  irc_banner_example:
-    prober: tcp
-    timeout: 5s
-    tcp:
-      query_response:
-        - send: "NICK prober"
-        - send: "USER prober prober prober :prober"
-        - expect: "PING :([^ ]+)"
-          send: "PONG ${1}"
-        - expect: "^:[^ ]+ 001"
   icmp:
     prober: icmp
     timeout: 5s
     icmp:
       preferred_ip_protocol: "ip4"
-  dns_udp_example:
-    prober: dns
-    timeout: 5s
-    dns:
-      query_name: "www.prometheus.io"
-      query_type: "A"
-      valid_rcodes:
-        - NOERROR
-      validate_answer_rrs:
-        fail_if_matches_regexp:
-          - ".*127.0.0.1"
-        fail_if_all_match_regexp:
-          - ".*127.0.0.1"
-        fail_if_not_matches_regexp:
-          - "www.prometheus.io.\t300\tIN\tA\t127.0.0.1"
-        fail_if_none_matches_regexp:
-          - "127.0.0.1"
-      validate_authority_rrs:
-        fail_if_matches_regexp:
-          - ".*127.0.0.1"
-      validate_additional_rrs:
-        fail_if_matches_regexp:
-          - ".*127.0.0.1"
-  dns_soa:
-    prober: dns
-    dns:
-      query_name: "prometheus.io"
-      query_type: "SOA"
-  dns_tcp_example:
-    prober: dns
-    dns:
-      transport_protocol: "tcp" # defaults to "udp"
-      preferred_ip_protocol: "ip4" # defaults to "ip6"
-      query_name: "www.prometheus.io"

m/hut/configuration.nix

@@ -7,11 +7,9 @@
     ../module/ceph.nix
     ../module/debuginfod.nix
     ../module/emulation.nix
-    ../module/slurm-client.nix
     ./gitlab-runner.nix
     ./monitoring.nix
     ./nfs.nix
-    ./slurm-server.nix
     ./nix-serve.nix
     ./public-inbox.nix
     ./gitea.nix

m/hut/slurm-server.nix

@@ -1,7 +0,0 @@
-{ ... }:
-
-{
-  services.slurm = {
-    server.enable = true;
-  };
-}

m/hut/targets.yml

@@ -4,7 +4,7 @@
   - xeon03-ipmi
   - xeon04-ipmi
   - koro-ipmi
-  - xeon06-ipmi
+  - weasel-ipmi
   - hut-ipmi
   - eudy-ipmi
   # Storage

m/map.nix

@@ -6,7 +6,7 @@
     switch-opa = { pos=41; size=1; };
 
     # SSF login
-    ssfhead = { pos=39; size=2; label="SSFHEAD"; board="R2208WTTYSR"; contact="operations@bsc.es"; };
+    apex = { pos=39; size=2; label="SSFHEAD"; board="R2208WTTYSR"; contact="rodrigo.arias@bsc.es"; };
 
     # Storage
     bay   = { pos=38; size=1; label="MDS01"; board="S2600WT2R"; sn="BQWL64850303"; contact="rodrigo.arias@bsc.es"; };
@@ -19,7 +19,7 @@
     xeon03 = { pos=33; size=1; label="SSF-XEON03"; board="S2600WTTR"; sn="BQWL64750826"; contact="rodrigo.arias@bsc.es"; };
     # Slot 34 empty
     koro   = { pos=31; size=1; label="SSF-XEON05"; board="S2600WTTR"; sn="BQWL64954293"; contact="rodrigo.arias@bsc.es"; };
-    xeon06 = { pos=30; size=1; label="SSF-XEON06"; board="S2600WTTR"; sn="BQWL64750846"; contact="antoni.navarro@bsc.es"; };
+    weasel = { pos=30; size=1; label="SSF-XEON06"; board="S2600WTTR"; sn="BQWL64750846"; contact="antoni.navarro@bsc.es"; };
     hut    = { pos=29; size=1; label="SSF-XEON07"; board="S2600WTTR"; sn="BQWL64751184"; contact="rodrigo.arias@bsc.es"; };
     eudy   = { pos=28; size=1; label="SSF-XEON08"; board="S2600WTTR"; sn="BQWL64756586"; contact="aleix.rocanonell@bsc.es"; };
 

m/module/nvidia.nix

@@ -0,0 +1,20 @@
+{ lib, config, pkgs, ... }:
+{
+  # Configure Nvidia driver to use with CUDA
+  hardware.nvidia.package = config.boot.kernelPackages.nvidiaPackages.production;
+  hardware.nvidia.open = lib.mkDefault (builtins.abort "hardware.nvidia.open not set");
+  hardware.graphics.enable = true;
+  nixpkgs.config.nvidia.acceptLicense = true;
+  services.xserver.videoDrivers = [ "nvidia" ];
+
+  # enable support for derivations which require nvidia-gpu to be available
+  # > requiredSystemFeatures = [ "cuda" ];
+  programs.nix-required-mounts.enable = true;
+  programs.nix-required-mounts.presets.nvidia-gpu.enable = true;
+  # They forgot to add the symlink
+  programs.nix-required-mounts.allowedPatterns.nvidia-gpu.paths = [
+    config.systemd.tmpfiles.settings.graphics-driver."/run/opengl-driver"."L+".argument
+  ];
+
+  environment.systemPackages = [ pkgs.cudainfo ];
+}

m/module/power-policy.nix

@@ -0,0 +1,33 @@
+{ config, lib, pkgs, ... }:
+
+with lib;
+
+let
+  cfg = config.power.policy;
+in
+{
+  options = {
+    power.policy = mkOption {
+      type = types.nullOr (types.enum [ "always-on" "previous" "always-off" ]);
+      default = null;
+      description = "Set power policy to use via IPMI.";
+    };
+  };
+
+  config = mkIf (cfg != null) {
+    systemd.services."power-policy" = {
+      description = "Set power policy to use via IPMI";
+      wantedBy = [ "multi-user.target" ];
+      unitConfig = {
+        StartLimitBurst = "10";
+        StartLimitIntervalSec = "10m";
+      };
+      serviceConfig = {
+        ExecStart = "${pkgs.ipmitool}/bin/ipmitool chassis policy ${cfg}";
+        Type = "oneshot";
+        Restart = "on-failure";
+        RestartSec = "5s";
+      };
+    };
+  };
+}

m/module/slurm-client.nix

@@ -1,33 +1,10 @@
-{ config, pkgs, lib, ... }:
+{ lib, ... }:
 
-let
-  suspendProgram = pkgs.writeScript "suspend.sh" ''
-    #!/usr/bin/env bash
-    exec 1>>/var/log/power_save.log 2>>/var/log/power_save.log
-    set -x
-    export "PATH=/run/current-system/sw/bin:$PATH"
-    echo "$(date) Suspend invoked $0 $*" >> /var/log/power_save.log
-    hosts=$(scontrol show hostnames $1)
-    for host in $hosts; do
-      echo Shutting down host: $host
-      ipmitool -I lanplus -H ''${host}-ipmi -P "" -U "" chassis power off
-    done
-  '';
+{
+  imports = [
+    ./slurm-common.nix
+  ];
 
-  resumeProgram = pkgs.writeScript "resume.sh" ''
-    #!/usr/bin/env bash
-    exec 1>>/var/log/power_save.log 2>>/var/log/power_save.log
-    set -x
-    export "PATH=/run/current-system/sw/bin:$PATH"
-    echo "$(date) Suspend invoked $0 $*" >> /var/log/power_save.log
-    hosts=$(scontrol show hostnames $1)
-    for host in $hosts; do
-      echo Starting host: $host
-      ipmitool -I lanplus -H ''${host}-ipmi -P "" -U "" chassis power on
-    done
-  '';
-
-in {
   systemd.services.slurmd.serviceConfig = {
     # Kill all processes in the control group on stop/restart. This will kill
     # all the jobs running, so ensure that we only upgrade when the nodes are
@@ -37,90 +14,5 @@ in {
     KillMode = lib.mkForce "control-group";
   };
 
-  services.slurm = {
-    client.enable = true;
-    controlMachine = "hut";
-    clusterName = "jungle";
-    nodeName = [
-      "owl[1,2]  Sockets=2 CoresPerSocket=14 ThreadsPerCore=2 Feature=owl"
-      "hut       Sockets=2 CoresPerSocket=14 ThreadsPerCore=2"
-    ];
-
-    partitionName = [
-      "owl Nodes=owl[1-2]     Default=YES DefaultTime=01:00:00 MaxTime=INFINITE State=UP"
-    ];
-
-    # See slurm.conf(5) for more details about these options.
-    extraConfig = ''
-      # Use PMIx for MPI by default. It works okay with MPICH and OpenMPI, but
-      # not with Intel MPI. For that use the compatibility shim libpmi.so
-      # setting I_MPI_PMI_LIBRARY=$pmix/lib/libpmi.so while maintaining the PMIx
-      # library in SLURM (--mpi=pmix). See more details here:
-      # https://pm.bsc.es/gitlab/rarias/jungle/-/issues/16
-      MpiDefault=pmix
-
-      # When a node reboots return that node to the slurm queue as soon as it
-      # becomes operative again.
-      ReturnToService=2
-
-      # Track all processes by using a cgroup
-      ProctrackType=proctrack/cgroup
-
-      # Enable task/affinity to allow the jobs to run in a specified subset of
-      # the resources. Use the task/cgroup plugin to enable process containment.
-      TaskPlugin=task/affinity,task/cgroup
-
-      # Power off unused nodes until they are requested
-      SuspendProgram=${suspendProgram}
-      SuspendTimeout=60
-      ResumeProgram=${resumeProgram}
-      ResumeTimeout=300
-      SuspendExcNodes=hut
-
-      # Turn the nodes off after 1 hour of inactivity
-      SuspendTime=3600
-
-      # Reduce port range so we can allow only this range in the firewall
-      SrunPortRange=60000-61000
-
-      # Use cores as consumable resources. In SLURM terms, a core may have
-      # multiple hardware threads (or CPUs).
-      SelectType=select/cons_tres
-
-      # Ignore memory constraints and only use unused cores to share a node with
-      # other jobs.
-      SelectTypeParameters=CR_Core
-
-      # Required for pam_slurm_adopt, see https://slurm.schedmd.com/pam_slurm_adopt.html
-      # This sets up the "extern" step into which ssh-launched processes will be
-      # adopted. Alloc runs the prolog at job allocation (salloc) rather than
-      # when a task runs (srun) so we can ssh early.
-      PrologFlags=Alloc,Contain,X11
-
-      # LaunchParameters=ulimit_pam_adopt will set RLIMIT_RSS in processes
-      # adopted by the external step, similar to tasks running in regular steps
-      # LaunchParameters=ulimit_pam_adopt
-      SlurmdDebug=debug5
-      #DebugFlags=Protocol,Cgroup
-    '';
-
-    extraCgroupConfig = ''
-      CgroupPlugin=cgroup/v2
-      #ConstrainCores=yes
-    '';
-  };
-
-  # Place the slurm config in /etc as this will be required by PAM
-  environment.etc.slurm.source = config.services.slurm.etcSlurm;
-
-  age.secrets.mungeKey = {
-    file = ../../secrets/munge-key.age;
-    owner = "munge";
-    group = "munge";
-  };
-
-  services.munge = {
-    enable = true;
-    password = config.age.secrets.mungeKey.path;
-  };
+  services.slurm.client.enable = true;
 }

m/module/slurm-common.nix

@@ -0,0 +1,115 @@
+{ config, pkgs, ... }:
+
+let
+  suspendProgram = pkgs.writeShellScript "suspend.sh" ''
+    exec 1>>/var/log/power_save.log 2>>/var/log/power_save.log
+    set -x
+    export "PATH=/run/current-system/sw/bin:$PATH"
+    echo "$(date) Suspend invoked $0 $*" >> /var/log/power_save.log
+    hosts=$(scontrol show hostnames $1)
+    for host in $hosts; do
+      echo Shutting down host: $host
+      ipmitool -I lanplus -H ''${host}-ipmi -P "" -U "" chassis power off
+    done
+  '';
+
+  resumeProgram = pkgs.writeShellScript "resume.sh" ''
+    exec 1>>/var/log/power_save.log 2>>/var/log/power_save.log
+    set -x
+    export "PATH=/run/current-system/sw/bin:$PATH"
+    echo "$(date) Suspend invoked $0 $*" >> /var/log/power_save.log
+    hosts=$(scontrol show hostnames $1)
+    for host in $hosts; do
+      echo Starting host: $host
+      ipmitool -I lanplus -H ''${host}-ipmi -P "" -U "" chassis power on
+    done
+  '';
+
+in {
+  services.slurm = {
+    controlMachine = "apex";
+    clusterName = "jungle";
+    nodeName = [
+      "owl[1,2]  Sockets=2 CoresPerSocket=14 ThreadsPerCore=2 Feature=owl"
+      "fox       Sockets=8 CoresPerSocket=24 ThreadsPerCore=1"
+    ];
+
+    partitionName = [
+      "owl Nodes=owl[1-2]     Default=YES DefaultTime=01:00:00 MaxTime=INFINITE State=UP"
+      "fox Nodes=fox          Default=NO  DefaultTime=01:00:00 MaxTime=INFINITE State=UP"
+    ];
+
+    # See slurm.conf(5) for more details about these options.
+    extraConfig = ''
+      # Use PMIx for MPI by default. It works okay with MPICH and OpenMPI, but
+      # not with Intel MPI. For that use the compatibility shim libpmi.so
+      # setting I_MPI_PMI_LIBRARY=$pmix/lib/libpmi.so while maintaining the PMIx
+      # library in SLURM (--mpi=pmix). See more details here:
+      # https://pm.bsc.es/gitlab/rarias/jungle/-/issues/16
+      MpiDefault=pmix
+
+      # When a node reboots return that node to the slurm queue as soon as it
+      # becomes operative again.
+      ReturnToService=2
+
+      # Track all processes by using a cgroup
+      ProctrackType=proctrack/cgroup
+
+      # Enable task/affinity to allow the jobs to run in a specified subset of
+      # the resources. Use the task/cgroup plugin to enable process containment.
+      TaskPlugin=task/affinity,task/cgroup
+
+      # Power off unused nodes until they are requested
+      SuspendProgram=${suspendProgram}
+      SuspendTimeout=60
+      ResumeProgram=${resumeProgram}
+      ResumeTimeout=300
+      SuspendExcNodes=fox
+
+      # Turn the nodes off after 1 hour of inactivity
+      SuspendTime=3600
+
+      # Reduce port range so we can allow only this range in the firewall
+      SrunPortRange=60000-61000
+
+      # Use cores as consumable resources. In SLURM terms, a core may have
+      # multiple hardware threads (or CPUs).
+      SelectType=select/cons_tres
+
+      # Ignore memory constraints and only use unused cores to share a node with
+      # other jobs.
+      SelectTypeParameters=CR_Core
+
+      # Required for pam_slurm_adopt, see https://slurm.schedmd.com/pam_slurm_adopt.html
+      # This sets up the "extern" step into which ssh-launched processes will be
+      # adopted. Alloc runs the prolog at job allocation (salloc) rather than
+      # when a task runs (srun) so we can ssh early.
+      PrologFlags=Alloc,Contain,X11
+
+      # LaunchParameters=ulimit_pam_adopt will set RLIMIT_RSS in processes
+      # adopted by the external step, similar to tasks running in regular steps
+      # LaunchParameters=ulimit_pam_adopt
+      SlurmdDebug=debug5
+      #DebugFlags=Protocol,Cgroup
+    '';
+
+    extraCgroupConfig = ''
+      CgroupPlugin=cgroup/v2
+      #ConstrainCores=yes
+    '';
+  };
+
+  # Place the slurm config in /etc as this will be required by PAM
+  environment.etc.slurm.source = config.services.slurm.etcSlurm;
+
+  age.secrets.mungeKey = {
+    file = ../../secrets/munge-key.age;
+    owner = "munge";
+    group = "munge";
+  };
+
+  services.munge = {
+    enable = true;
+    password = config.age.secrets.mungeKey.path;
+  };
+}

m/module/slurm-server.nix

@@ -0,0 +1,23 @@
+{ ... }:
+
+{
+  imports = [
+    ./slurm-common.nix
+  ];
+
+  services.slurm.server.enable = true;
+
+  networking.firewall = {
+    extraCommands = ''
+      # Accept slurm connections to controller from compute nodes
+      iptables -A nixos-fw -p tcp -s 10.0.40.0/24 --dport 6817 -j nixos-fw-accept
+      # Accept slurm connections from compute nodes for srun
+      iptables -A nixos-fw -p tcp -s 10.0.40.0/24 --dport 60000:61000 -j nixos-fw-accept
+
+      # Accept slurm connections to controller from fox (via wireguard)
+      iptables -A nixos-fw -p tcp -i wg0 -s 10.106.0.1/32 --dport 6817 -j nixos-fw-accept
+      # Accept slurm connections from fox for srun (via wireguard)
+      iptables -A nixos-fw -p tcp -i wg0 -s 10.106.0.1/32 --dport 60000:61000 -j nixos-fw-accept
+    '';
+  };
+}

m/module/ssh-hut-extern.nix

@@ -1,9 +1,8 @@
 {
   programs.ssh.extraConfig = ''
-    Host ssfhead
+    Host apex ssfhead
       HostName ssflogin.bsc.es
     Host hut
-      ProxyJump ssfhead
-      HostName xeon07
+      ProxyJump apex
   '';
 }

m/module/vpn-dac.nix

@@ -0,0 +1,35 @@
+{config, ...}:
+{
+  age.secrets.vpn-dac-login.file = ../../secrets/vpn-dac-login.age;
+  age.secrets.vpn-dac-client-key.file = ../../secrets/vpn-dac-client-key.age;
+
+  services.openvpn.servers = {
+    # systemctl status openvpn-dac.service
+    dac = {
+      config = ''
+        client
+        dev tun
+        proto tcp
+        remote vpn.ac.upc.edu 1194
+        remote vpn.ac.upc.edu 80
+        resolv-retry infinite
+        nobind
+        persist-key
+        persist-tun
+        ca ${./vpn-dac/ca.crt}
+        cert ${./vpn-dac/client.crt}
+        # Only key needs to be secret
+        key ${config.age.secrets.vpn-dac-client-key.path}
+        remote-cert-tls server
+        comp-lzo
+        verb 3
+        auth-user-pass ${config.age.secrets.vpn-dac-login.path}
+        reneg-sec 0
+
+        # Only route fox-ipmi
+        pull-filter ignore "route "
+        route 147.83.35.27 255.255.255.255
+      '';
+    };
+  };
+}

m/module/vpn-dac/ca.crt

@@ -0,0 +1,31 @@
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----

m/module/vpn-dac/client.crt

@@ -0,0 +1,100 @@
+Certificate:
+    Data:
+        Version: 3 (0x2)
+        Serial Number: 2 (0x2)
+    Signature Algorithm: sha256WithRSAEncryption
+        Issuer: C=ES, ST=Barcelona, L=Barcelona, O=Universitat Politecnica de Catalunya, OU=Arquitectura de Computadors, CN=LCAC CA/name=LCAC/emailAddress=lcac@ac.upc.edu
+        Validity
+            Not Before: Jan 12 12:45:41 2016 GMT
+            Not After : Jan 12 12:45:41 2046 GMT
+        Subject: C=ES, ST=Barcelona, L=Barcelona, O=Universitat Politecnica de Catalunya, OU=Arquitectura de Computadors, CN=client/name=LCAC/emailAddress=lcac@ac.upc.edu
+        Subject Public Key Info:
+            Public Key Algorithm: rsaEncryption
+                Public-Key: (2048 bit)
+                Modulus:
+                    00:97:99:fa:7a:0e:4d:e2:1d:a5:b1:a8:14:18:64:
+                    c7:66:bf:de:99:1d:92:3b:86:82:4d:95:39:f7:a6:
+                    56:49:97:14:4f:e3:37:00:6c:f4:d0:1d:56:79:e7:
+                    19:b5:dd:36:15:8e:1d:57:7b:59:29:d2:11:bf:58:
+                    48:e0:f7:41:3d:16:64:8d:a2:0b:4a:ac:fa:c6:83:
+                    dc:10:2a:2c:d9:97:48:ee:11:2a:bc:4b:60:dd:b9:
+                    2e:8f:45:ca:87:0b:38:65:1c:f8:a2:1d:f9:50:aa:
+                    6e:60:f9:48:df:57:12:23:e1:e7:0c:81:5c:9f:c5:
+                    b2:e6:99:99:95:30:6d:57:36:06:8c:fd:fb:f9:4f:
+                    60:d2:3c:ba:ae:28:56:2f:da:58:5c:e8:c5:7b:ec:
+                    76:d9:28:6e:fb:8c:07:f9:d7:23:c3:72:76:3c:fa:
+                    dc:20:67:8f:cc:16:e0:91:07:d5:68:f9:20:4d:7d:
+                    5c:2d:02:04:16:76:52:f3:53:be:a3:dc:0d:d5:fb:
+                    6b:55:29:f3:52:35:c8:7d:99:d1:4a:94:be:b1:8e:
+                    fd:85:18:25:eb:41:e9:56:da:af:62:84:20:0a:00:
+                    17:94:92:94:91:6a:f8:54:37:17:ee:1e:bb:fb:93:
+                    71:91:d9:e4:e9:b8:3b:18:7d:6d:7d:4c:ce:58:55:
+                    f9:41
+                Exponent: 65537 (0x10001)
+        X509v3 extensions:
+            X509v3 Basic Constraints: 
+                CA:FALSE
+            Netscape Comment: 
+                Easy-RSA Generated Certificate
+            X509v3 Subject Key Identifier: 
+                1B:88:06:D5:33:1D:5C:48:46:B5:DE:78:89:36:96:91:3A:74:43:18
+            X509v3 Authority Key Identifier: 
+                keyid:DC:9E:CD:7D:5C:1F:53:78:41:1D:78:79:B0:69:96:3A:A6:B4:4E:EA
+                DirName:/C=ES/ST=Barcelona/L=Barcelona/O=Universitat Politecnica de Catalunya/OU=Arquitectura de Computadors/CN=LCAC CA/name=LCAC/emailAddress=lcac@ac.upc.edu
+                serial:91:F5:D7:C3:C0:A6:4E:61
+
+            X509v3 Extended Key Usage: 
+                TLS Web Client Authentication
+            X509v3 Key Usage: 
+                Digital Signature
+            X509v3 Subject Alternative Name: 
+                DNS:client
+    Signature Algorithm: sha256WithRSAEncryption
+         42:e8:50:b2:e7:88:75:86:0b:bb:29:e3:aa:c6:0e:4c:e8:ea:
+         3d:0c:02:31:7f:3b:80:0c:3f:80:af:45:d6:62:27:a0:0e:e7:
+         26:09:12:97:95:f8:d9:9b:89:b5:ef:56:64:f1:de:82:74:e0:
+         31:0a:cc:90:0a:bd:50:b8:54:95:0a:ae:3b:40:df:76:b6:d1:
+         01:2e:f3:96:9f:52:d4:e9:14:6d:b7:14:9d:45:99:33:36:2a:
+         01:0b:15:1a:ed:55:dc:64:83:65:1a:06:42:d9:c7:dc:97:d4:
+         02:81:c2:58:2b:ea:e4:b7:ae:84:3a:e4:3f:f1:2e:fa:ec:f3:
+         40:5d:b8:6a:d5:5e:e1:e8:2f:e2:2f:48:a4:38:a1:4f:22:e3:
+         4f:66:94:aa:02:78:9a:2b:7a:5d:aa:aa:51:a5:e3:d0:91:e9:
+         1d:f9:08:ed:8b:51:c9:a6:af:46:85:b5:1c:ed:12:a1:28:33:
+         75:36:00:d8:5c:14:65:96:c0:28:7d:47:50:a4:89:5f:b0:72:
+         1a:4b:13:17:26:0f:f0:b8:65:3c:e9:96:36:f9:bf:90:59:33:
+         87:1f:01:03:25:f8:f0:3a:9b:33:02:d0:0a:43:b5:0a:cf:62:
+         a1:45:38:37:07:9d:9c:94:0b:31:c6:3c:34:b7:fc:5a:0c:e4:
+         bf:23:f6:7d
+-----BEGIN CERTIFICATE-----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+-----END CERTIFICATE-----

m/raccoon/configuration.nix

@@ -6,6 +6,7 @@
     ../module/emulation.nix
     ../module/debuginfod.nix
     ../module/ssh-hut-extern.nix
+    ../module/nvidia.nix
     ../eudy/kernel/perf.nix
   ];
 
@@ -38,6 +39,7 @@
     };
     hosts = {
       "10.0.44.4" = [ "tent" ];
+      "84.88.53.236" = [ "apex" ];
     };
   };
 
@@ -49,15 +51,7 @@
   # Enable performance governor
   powerManagement.cpuFreqGovernor = "performance";
 
-  # Configure Nvidia driver to use with CUDA
-  hardware.nvidia.package = config.boot.kernelPackages.nvidiaPackages.production;
-  hardware.graphics.enable = true;
-  nixpkgs.config.allowUnfree = true;
-  nixpkgs.config.nvidia.acceptLicense = true;
-  services.xserver.videoDrivers = [ "nvidia" ];
-
-  # Disable garbage collection for now
-  nix.gc.automatic = lib.mkForce false;
+  hardware.nvidia.open = false; # Maxwell is older than Turing architecture
 
   services.openssh.settings.X11Forwarding = true;
 

m/tent/configuration.nix

@@ -14,6 +14,7 @@
     ../hut/public-inbox.nix
     ../hut/msmtp.nix
     ../module/p.nix
+    ../module/vpn-dac.nix
   ];
 
   # Select the this using the ID to avoid mismatches
@@ -30,8 +31,11 @@
 
     # Only BSC DNSs seem to be reachable from the office VLAN
     nameservers = [ "84.88.52.35" "84.88.52.36" ];
-    search = [ "bsc.es" ];
+    search = [ "bsc.es" "ac.upc.edu" ];
     defaultGateway = "10.0.44.1";
+    hosts = {
+      "84.88.53.236" = [ "apex" ];
+    };
   };
 
   services.p.enable = true;

m/tent/monitoring.nix

@@ -165,6 +165,7 @@
             "anella-bsc.cesca.cat"
             "upc-anella.cesca.cat"
             "fox.ac.upc.edu"
+            "fox-ipmi.ac.upc.edu"
             "arenys5.ac.upc.edu"
             "arenys0-2.ac.upc.edu"
             "epi01.bsc.es"
@@ -200,6 +201,17 @@
           module = [ "raccoon" ];
         };
       }
+      {
+        job_name = "ipmi-fox";
+        metrics_path = "/ipmi";
+        static_configs = [
+          { targets = [ "127.0.0.1:9290" ]; }
+        ];
+        params = {
+          target = [ "fox-ipmi.ac.upc.edu" ];
+          module = [ "fox" ];
+        };
+      }
     ];
   };
 }

m/weasel/configuration.nix

@@ -0,0 +1,32 @@
+{ lib, ... }:
+
+{
+  imports = [
+    ../common/ssf.nix
+  ];
+
+  # Select this using the ID to avoid mismatches
+  boot.loader.grub.device = "/dev/disk/by-id/wwn-0x55cd2e414d5356ca";
+
+  # No swap, there is plenty of RAM
+  swapDevices = lib.mkForce [];
+
+  # Users with sudo access
+  users.groups.wheel.members = [ "abonerib" "anavarro" ];
+
+  # Run julia installed with juliaup using julia's own libraries:
+  # NIX_LD_LIBRARY_PATH=~/.julia/juliaup/${VERS}/lib/julia ~/.juliaup/bin/julia
+  programs.nix-ld.enable = true;
+
+  networking = {
+    hostName = "weasel";
+    interfaces.eno1.ipv4.addresses = [ {
+      address = "10.0.40.6";
+      prefixLength = 24;
+    } ];
+    interfaces.ibp5s0.ipv4.addresses = [ {
+      address = "10.0.42.6";
+      prefixLength = 24;
+    } ];
+  };
+}

pkgs/cudainfo/Makefile

@@ -0,0 +1,12 @@
+HOSTCXX  ?= g++
+NVCC     := nvcc -ccbin $(HOSTCXX)
+CXXFLAGS := -m64
+
+# Target rules
+all: cudainfo
+
+cudainfo: cudainfo.cpp
+	$(NVCC) $(CXXFLAGS) -o $@ $<
+
+clean:
+	rm -f cudainfo cudainfo.o

pkgs/cudainfo/cudainfo.cpp

@@ -0,0 +1,600 @@
+/*
+ * Copyright 1993-2015 NVIDIA Corporation.  All rights reserved.
+ *
+ * Please refer to the NVIDIA end user license agreement (EULA) associated
+ * with this source code for terms and conditions that govern your use of
+ * this software. Any use, reproduction, disclosure, or distribution of
+ * this software and related documentation outside the terms of the EULA
+ * is strictly prohibited.
+ *
+ */
+/* This sample queries the properties of the CUDA devices present in the system via CUDA Runtime API. */
+
+// Shared Utilities (QA Testing)
+
+// std::system includes
+#include <memory>
+#include <iostream>
+
+#include <cuda_runtime.h>
+
+// This will output the proper CUDA error strings in the event that a CUDA host call returns an error
+#define checkCudaErrors(val)           check ( (val), #val, __FILE__, __LINE__ )
+
+// CUDA Runtime error messages
+#ifdef __DRIVER_TYPES_H__
+static const char *_cudaGetErrorEnum(cudaError_t error)
+{
+    switch (error)
+    {
+        case cudaSuccess:
+            return "cudaSuccess";
+
+        case cudaErrorMissingConfiguration:
+            return "cudaErrorMissingConfiguration";
+
+        case cudaErrorMemoryAllocation:
+            return "cudaErrorMemoryAllocation";
+
+        case cudaErrorInitializationError:
+            return "cudaErrorInitializationError";
+
+        case cudaErrorLaunchFailure:
+            return "cudaErrorLaunchFailure";
+
+        case cudaErrorPriorLaunchFailure:
+            return "cudaErrorPriorLaunchFailure";
+
+        case cudaErrorLaunchTimeout:
+            return "cudaErrorLaunchTimeout";
+
+        case cudaErrorLaunchOutOfResources:
+            return "cudaErrorLaunchOutOfResources";
+
+        case cudaErrorInvalidDeviceFunction:
+            return "cudaErrorInvalidDeviceFunction";
+
+        case cudaErrorInvalidConfiguration:
+            return "cudaErrorInvalidConfiguration";
+
+        case cudaErrorInvalidDevice:
+            return "cudaErrorInvalidDevice";
+
+        case cudaErrorInvalidValue:
+            return "cudaErrorInvalidValue";
+
+        case cudaErrorInvalidPitchValue:
+            return "cudaErrorInvalidPitchValue";
+
+        case cudaErrorInvalidSymbol:
+            return "cudaErrorInvalidSymbol";
+
+        case cudaErrorMapBufferObjectFailed:
+            return "cudaErrorMapBufferObjectFailed";
+
+        case cudaErrorUnmapBufferObjectFailed:
+            return "cudaErrorUnmapBufferObjectFailed";
+
+        case cudaErrorInvalidHostPointer:
+            return "cudaErrorInvalidHostPointer";
+
+        case cudaErrorInvalidDevicePointer:
+            return "cudaErrorInvalidDevicePointer";
+
+        case cudaErrorInvalidTexture:
+            return "cudaErrorInvalidTexture";
+
+        case cudaErrorInvalidTextureBinding:
+            return "cudaErrorInvalidTextureBinding";
+
+        case cudaErrorInvalidChannelDescriptor:
+            return "cudaErrorInvalidChannelDescriptor";
+
+        case cudaErrorInvalidMemcpyDirection:
+            return "cudaErrorInvalidMemcpyDirection";
+
+        case cudaErrorAddressOfConstant:
+            return "cudaErrorAddressOfConstant";
+
+        case cudaErrorTextureFetchFailed:
+            return "cudaErrorTextureFetchFailed";
+
+        case cudaErrorTextureNotBound:
+            return "cudaErrorTextureNotBound";
+
+        case cudaErrorSynchronizationError:
+            return "cudaErrorSynchronizationError";
+
+        case cudaErrorInvalidFilterSetting:
+            return "cudaErrorInvalidFilterSetting";
+
+        case cudaErrorInvalidNormSetting:
+            return "cudaErrorInvalidNormSetting";
+
+        case cudaErrorMixedDeviceExecution:
+            return "cudaErrorMixedDeviceExecution";
+
+        case cudaErrorCudartUnloading:
+            return "cudaErrorCudartUnloading";
+
+        case cudaErrorUnknown:
+            return "cudaErrorUnknown";
+
+        case cudaErrorNotYetImplemented:
+            return "cudaErrorNotYetImplemented";
+
+        case cudaErrorMemoryValueTooLarge:
+            return "cudaErrorMemoryValueTooLarge";
+
+        case cudaErrorInvalidResourceHandle:
+            return "cudaErrorInvalidResourceHandle";
+
+        case cudaErrorNotReady:
+            return "cudaErrorNotReady";
+
+        case cudaErrorInsufficientDriver:
+            return "cudaErrorInsufficientDriver";
+
+        case cudaErrorSetOnActiveProcess:
+            return "cudaErrorSetOnActiveProcess";
+
+        case cudaErrorInvalidSurface:
+            return "cudaErrorInvalidSurface";
+
+        case cudaErrorNoDevice:
+            return "cudaErrorNoDevice";
+
+        case cudaErrorECCUncorrectable:
+            return "cudaErrorECCUncorrectable";
+
+        case cudaErrorSharedObjectSymbolNotFound:
+            return "cudaErrorSharedObjectSymbolNotFound";
+
+        case cudaErrorSharedObjectInitFailed:
+            return "cudaErrorSharedObjectInitFailed";
+
+        case cudaErrorUnsupportedLimit:
+            return "cudaErrorUnsupportedLimit";
+
+        case cudaErrorDuplicateVariableName:
+            return "cudaErrorDuplicateVariableName";
+
+        case cudaErrorDuplicateTextureName:
+            return "cudaErrorDuplicateTextureName";
+
+        case cudaErrorDuplicateSurfaceName:
+            return "cudaErrorDuplicateSurfaceName";
+
+        case cudaErrorDevicesUnavailable:
+            return "cudaErrorDevicesUnavailable";
+
+        case cudaErrorInvalidKernelImage:
+            return "cudaErrorInvalidKernelImage";
+
+        case cudaErrorNoKernelImageForDevice:
+            return "cudaErrorNoKernelImageForDevice";
+
+        case cudaErrorIncompatibleDriverContext:
+            return "cudaErrorIncompatibleDriverContext";
+
+        case cudaErrorPeerAccessAlreadyEnabled:
+            return "cudaErrorPeerAccessAlreadyEnabled";
+
+        case cudaErrorPeerAccessNotEnabled:
+            return "cudaErrorPeerAccessNotEnabled";
+
+        case cudaErrorDeviceAlreadyInUse:
+            return "cudaErrorDeviceAlreadyInUse";
+
+        case cudaErrorProfilerDisabled:
+            return "cudaErrorProfilerDisabled";
+
+        case cudaErrorProfilerNotInitialized:
+            return "cudaErrorProfilerNotInitialized";
+
+        case cudaErrorProfilerAlreadyStarted:
+            return "cudaErrorProfilerAlreadyStarted";
+
+        case cudaErrorProfilerAlreadyStopped:
+            return "cudaErrorProfilerAlreadyStopped";
+
+        /* Since CUDA 4.0*/
+        case cudaErrorAssert:
+            return "cudaErrorAssert";
+
+        case cudaErrorTooManyPeers:
+            return "cudaErrorTooManyPeers";
+
+        case cudaErrorHostMemoryAlreadyRegistered:
+            return "cudaErrorHostMemoryAlreadyRegistered";
+
+        case cudaErrorHostMemoryNotRegistered:
+            return "cudaErrorHostMemoryNotRegistered";
+
+        /* Since CUDA 5.0 */
+        case cudaErrorOperatingSystem:
+            return "cudaErrorOperatingSystem";
+
+        case cudaErrorPeerAccessUnsupported:
+            return "cudaErrorPeerAccessUnsupported";
+
+        case cudaErrorLaunchMaxDepthExceeded:
+            return "cudaErrorLaunchMaxDepthExceeded";
+
+        case cudaErrorLaunchFileScopedTex:
+            return "cudaErrorLaunchFileScopedTex";
+
+        case cudaErrorLaunchFileScopedSurf:
+            return "cudaErrorLaunchFileScopedSurf";
+
+        case cudaErrorSyncDepthExceeded:
+            return "cudaErrorSyncDepthExceeded";
+
+        case cudaErrorLaunchPendingCountExceeded:
+            return "cudaErrorLaunchPendingCountExceeded";
+
+        case cudaErrorNotPermitted:
+            return "cudaErrorNotPermitted";
+
+        case cudaErrorNotSupported:
+            return "cudaErrorNotSupported";
+
+        /* Since CUDA 6.0 */
+        case cudaErrorHardwareStackError:
+            return "cudaErrorHardwareStackError";
+
+        case cudaErrorIllegalInstruction:
+            return "cudaErrorIllegalInstruction";
+
+        case cudaErrorMisalignedAddress:
+            return "cudaErrorMisalignedAddress";
+
+        case cudaErrorInvalidAddressSpace:
+            return "cudaErrorInvalidAddressSpace";
+
+        case cudaErrorInvalidPc:
+            return "cudaErrorInvalidPc";
+
+        case cudaErrorIllegalAddress:
+            return "cudaErrorIllegalAddress";
+
+        /* Since CUDA 6.5*/
+        case cudaErrorInvalidPtx:
+            return "cudaErrorInvalidPtx";
+
+        case cudaErrorInvalidGraphicsContext:
+            return "cudaErrorInvalidGraphicsContext";
+
+        case cudaErrorStartupFailure:
+            return "cudaErrorStartupFailure";
+
+        case cudaErrorApiFailureBase:
+            return "cudaErrorApiFailureBase";
+    }
+
+    return "<unknown>";
+}
+#endif
+
+template< typename T >
+void check(T result, char const *const func, const char *const file, int const line)
+{
+    if (result)
+    {
+        fprintf(stderr, "CUDA error at %s:%d code=%d(%s) \"%s\" \n",
+                file, line, static_cast<unsigned int>(result), _cudaGetErrorEnum(result), func);
+        cudaDeviceReset();
+        // Make sure we call CUDA Device Reset before exiting
+        exit(EXIT_FAILURE);
+    }
+}
+
+int *pArgc = NULL;
+char **pArgv = NULL;
+
+#if CUDART_VERSION < 5000
+
+// CUDA-C includes
+#include <cuda.h>
+
+// This function wraps the CUDA Driver API into a template function
+template <class T>
+inline void getCudaAttribute(T *attribute, CUdevice_attribute device_attribute, int device)
+{
+    CUresult error =    cuDeviceGetAttribute(attribute, device_attribute, device);
+
+    if (CUDA_SUCCESS != error) {
+        fprintf(stderr, "cuSafeCallNoSync() Driver API error = %04d from file <%s>, line %i.\n",
+                error, __FILE__, __LINE__);
+
+        // cudaDeviceReset causes the driver to clean up all state. While
+        // not mandatory in normal operation, it is good practice.  It is also
+        // needed to ensure correct operation when the application is being
+        // profiled. Calling cudaDeviceReset causes all profile data to be
+        // flushed before the application exits
+        cudaDeviceReset();
+        exit(EXIT_FAILURE);
+    }
+}
+
+#endif /* CUDART_VERSION < 5000 */
+
+// Beginning of GPU Architecture definitions
+inline int ConvertSMVer2Cores(int major, int minor)
+{
+    // Defines for GPU Architecture types (using the SM version to determine the # of cores per SM
+    typedef struct {
+        int SM; // 0xMm (hexidecimal notation), M = SM Major version, and m = SM minor version
+        int Cores;
+    } sSMtoCores;
+
+    sSMtoCores nGpuArchCoresPerSM[] = {
+        { 0x20, 32 }, // Fermi Generation (SM 2.0) GF100 class
+        { 0x21, 48 }, // Fermi Generation (SM 2.1) GF10x class
+        { 0x30, 192}, // Kepler Generation (SM 3.0) GK10x class
+        { 0x32, 192}, // Kepler Generation (SM 3.2) GK10x class
+        { 0x35, 192}, // Kepler Generation (SM 3.5) GK11x class
+        { 0x37, 192}, // Kepler Generation (SM 3.7) GK21x class
+        { 0x50, 128}, // Maxwell Generation (SM 5.0) GM10x class
+        { 0x52, 128}, // Maxwell Generation (SM 5.2) GM20x class
+        {   -1, -1 }
+    };
+
+    int index = 0;
+
+    while (nGpuArchCoresPerSM[index].SM != -1) {
+        if (nGpuArchCoresPerSM[index].SM == ((major << 4) + minor)) {
+            return nGpuArchCoresPerSM[index].Cores;
+        }
+
+        index++;
+    }
+
+    // If we don't find the values, we default use the previous one to run properly
+    printf("MapSMtoCores for SM %d.%d is undefined.  Default to use %d Cores/SM\n", major, minor, nGpuArchCoresPerSM[index-1].Cores);
+    return nGpuArchCoresPerSM[index-1].Cores;
+}
+
+////////////////////////////////////////////////////////////////////////////////
+// Program main
+////////////////////////////////////////////////////////////////////////////////
+int
+main(int argc, char **argv)
+{
+    pArgc = &argc;
+    pArgv = argv;
+
+    printf("%s Starting...\n\n", argv[0]);
+    printf(" CUDA Device Query (Runtime API) version (CUDART static linking)\n\n");
+
+    int deviceCount = 0;
+    cudaError_t error_id = cudaGetDeviceCount(&deviceCount);
+
+    if (error_id != cudaSuccess) {
+        printf("cudaGetDeviceCount failed: %s (%d)\n",
+			cudaGetErrorString(error_id), (int) error_id);
+        printf("Result = FAIL\n");
+        exit(EXIT_FAILURE);
+    }
+
+    // This function call returns 0 if there are no CUDA capable devices.
+    if (deviceCount == 0)
+        printf("There are no available device(s) that support CUDA\n");
+    else
+        printf("Detected %d CUDA Capable device(s)\n", deviceCount);
+
+    int dev, driverVersion = 0, runtimeVersion = 0;
+
+    for (dev = 0; dev < deviceCount; ++dev) {
+        cudaSetDevice(dev);
+        cudaDeviceProp deviceProp;
+        cudaGetDeviceProperties(&deviceProp, dev);
+
+        printf("\nDevice %d: \"%s\"\n", dev, deviceProp.name);
+
+        // Console log
+        cudaDriverGetVersion(&driverVersion);
+        cudaRuntimeGetVersion(&runtimeVersion);
+        printf("  CUDA Driver Version / Runtime Version          %d.%d / %d.%d\n", driverVersion/1000, (driverVersion%100)/10, runtimeVersion/1000, (runtimeVersion%100)/10);
+        printf("  CUDA Capability Major/Minor version number:    %d.%d\n", deviceProp.major, deviceProp.minor);
+
+        printf("  Total amount of global memory:                 %.0f MBytes (%llu bytes)\n",
+                (float)deviceProp.totalGlobalMem/1048576.0f, (unsigned long long) deviceProp.totalGlobalMem);
+
+        printf("  (%2d) Multiprocessors, (%3d) CUDA Cores/MP:     %d CUDA Cores\n",
+               deviceProp.multiProcessorCount,
+               ConvertSMVer2Cores(deviceProp.major, deviceProp.minor),
+               ConvertSMVer2Cores(deviceProp.major, deviceProp.minor) * deviceProp.multiProcessorCount);
+        printf("  GPU Max Clock rate:                            %.0f MHz (%0.2f GHz)\n", deviceProp.clockRate * 1e-3f, deviceProp.clockRate * 1e-6f);
+
+
+#if CUDART_VERSION >= 5000
+        // This is supported in CUDA 5.0 (runtime API device properties)
+        printf("  Memory Clock rate:                             %.0f Mhz\n", deviceProp.memoryClockRate * 1e-3f);
+        printf("  Memory Bus Width:                              %d-bit\n",   deviceProp.memoryBusWidth);
+
+        if (deviceProp.l2CacheSize) {
+            printf("  L2 Cache Size:                                 %d bytes\n", deviceProp.l2CacheSize);
+        }
+
+#else
+        // This only available in CUDA 4.0-4.2 (but these were only exposed in the CUDA Driver API)
+        int memoryClock;
+        getCudaAttribute<int>(&memoryClock, CU_DEVICE_ATTRIBUTE_MEMORY_CLOCK_RATE, dev);
+        printf("  Memory Clock rate:                             %.0f Mhz\n", memoryClock * 1e-3f);
+        int memBusWidth;
+        getCudaAttribute<int>(&memBusWidth, CU_DEVICE_ATTRIBUTE_GLOBAL_MEMORY_BUS_WIDTH, dev);
+        printf("  Memory Bus Width:                              %d-bit\n", memBusWidth);
+        int L2CacheSize;
+        getCudaAttribute<int>(&L2CacheSize, CU_DEVICE_ATTRIBUTE_L2_CACHE_SIZE, dev);
+
+        if (L2CacheSize) {
+            printf("  L2 Cache Size:                                 %d bytes\n", L2CacheSize);
+        }
+
+#endif
+
+        printf("  Maximum Texture Dimension Size (x,y,z)         1D=(%d), 2D=(%d, %d), 3D=(%d, %d, %d)\n",
+               deviceProp.maxTexture1D   , deviceProp.maxTexture2D[0], deviceProp.maxTexture2D[1],
+               deviceProp.maxTexture3D[0], deviceProp.maxTexture3D[1], deviceProp.maxTexture3D[2]);
+        printf("  Maximum Layered 1D Texture Size, (num) layers  1D=(%d), %d layers\n",
+               deviceProp.maxTexture1DLayered[0], deviceProp.maxTexture1DLayered[1]);
+        printf("  Maximum Layered 2D Texture Size, (num) layers  2D=(%d, %d), %d layers\n",
+               deviceProp.maxTexture2DLayered[0], deviceProp.maxTexture2DLayered[1], deviceProp.maxTexture2DLayered[2]);
+
+
+        printf("  Total amount of constant memory:               %lu bytes\n", deviceProp.totalConstMem);
+        printf("  Total amount of shared memory per block:       %lu bytes\n", deviceProp.sharedMemPerBlock);
+        printf("  Total number of registers available per block: %d\n", deviceProp.regsPerBlock);
+        printf("  Warp size:                                     %d\n", deviceProp.warpSize);
+        printf("  Maximum number of threads per multiprocessor:  %d\n", deviceProp.maxThreadsPerMultiProcessor);
+        printf("  Maximum number of threads per block:           %d\n", deviceProp.maxThreadsPerBlock);
+        printf("  Max dimension size of a thread block (x,y,z): (%d, %d, %d)\n",
+               deviceProp.maxThreadsDim[0],
+               deviceProp.maxThreadsDim[1],
+               deviceProp.maxThreadsDim[2]);
+        printf("  Max dimension size of a grid size    (x,y,z): (%d, %d, %d)\n",
+               deviceProp.maxGridSize[0],
+               deviceProp.maxGridSize[1],
+               deviceProp.maxGridSize[2]);
+        printf("  Maximum memory pitch:                          %lu bytes\n", deviceProp.memPitch);
+        printf("  Texture alignment:                             %lu bytes\n", deviceProp.textureAlignment);
+        printf("  Concurrent copy and kernel execution:          %s with %d copy engine(s)\n", (deviceProp.deviceOverlap ? "Yes" : "No"), deviceProp.asyncEngineCount);
+        printf("  Run time limit on kernels:                     %s\n", deviceProp.kernelExecTimeoutEnabled ? "Yes" : "No");
+        printf("  Integrated GPU sharing Host Memory:            %s\n", deviceProp.integrated ? "Yes" : "No");
+        printf("  Support host page-locked memory mapping:       %s\n", deviceProp.canMapHostMemory ? "Yes" : "No");
+        printf("  Alignment requirement for Surfaces:            %s\n", deviceProp.surfaceAlignment ? "Yes" : "No");
+        printf("  Device has ECC support:                        %s\n", deviceProp.ECCEnabled ? "Enabled" : "Disabled");
+#if defined(WIN32) || defined(_WIN32) || defined(WIN64) || defined(_WIN64)
+        printf("  CUDA Device Driver Mode (TCC or WDDM):         %s\n", deviceProp.tccDriver ? "TCC (Tesla Compute Cluster Driver)" : "WDDM (Windows Display Driver Model)");
+#endif
+        printf("  Device supports Unified Addressing (UVA):      %s\n", deviceProp.unifiedAddressing ? "Yes" : "No");
+        printf("  Device PCI Domain ID / Bus ID / location ID:   %d / %d / %d\n", deviceProp.pciDomainID, deviceProp.pciBusID, deviceProp.pciDeviceID);
+
+        const char *sComputeMode[] = {
+            "Default (multiple host threads can use ::cudaSetDevice() with device simultaneously)",
+            "Exclusive (only one host thread in one process is able to use ::cudaSetDevice() with this device)",
+            "Prohibited (no host thread can use ::cudaSetDevice() with this device)",
+            "Exclusive Process (many threads in one process is able to use ::cudaSetDevice() with this device)",
+            "Unknown",
+            NULL
+        };
+        printf("  Compute Mode:\n");
+        printf("     < %s >\n", sComputeMode[deviceProp.computeMode]);
+    }
+
+    // If there are 2 or more GPUs, query to determine whether RDMA is supported
+    if (deviceCount >= 2)
+    {
+        cudaDeviceProp prop[64];
+        int gpuid[64]; // we want to find the first two GPU's that can support P2P
+        int gpu_p2p_count = 0;
+
+        for (int i=0; i < deviceCount; i++)
+        {
+            checkCudaErrors(cudaGetDeviceProperties(&prop[i], i));
+
+            // Only boards based on Fermi or later can support P2P
+            if ((prop[i].major >= 2)
+#if defined(WIN32) || defined(_WIN32) || defined(WIN64) || defined(_WIN64)
+                // on Windows (64-bit), the Tesla Compute Cluster driver for windows must be enabled to supprot this
+                && prop[i].tccDriver
+#endif
+               )
+            {
+                // This is an array of P2P capable GPUs
+                gpuid[gpu_p2p_count++] = i;
+            }
+        }
+
+        // Show all the combinations of support P2P GPUs
+        int can_access_peer_0_1, can_access_peer_1_0;
+
+        if (gpu_p2p_count >= 2)
+        {
+            for (int i = 0; i < gpu_p2p_count-1; i++)
+            {
+                for (int j = 1; j < gpu_p2p_count; j++)
+                {
+                    checkCudaErrors(cudaDeviceCanAccessPeer(&can_access_peer_0_1, gpuid[i], gpuid[j]));
+                    printf("> Peer access from %s (GPU%d) -> %s (GPU%d) : %s\n", prop[gpuid[i]].name, gpuid[i],
+                           prop[gpuid[j]].name, gpuid[j] ,
+                           can_access_peer_0_1 ? "Yes" : "No");
+                }
+            }
+
+            for (int j = 1; j < gpu_p2p_count; j++)
+            {
+                for (int i = 0; i < gpu_p2p_count-1; i++)
+                {
+                    checkCudaErrors(cudaDeviceCanAccessPeer(&can_access_peer_1_0, gpuid[j], gpuid[i]));
+                    printf("> Peer access from %s (GPU%d) -> %s (GPU%d) : %s\n", prop[gpuid[j]].name, gpuid[j],
+                           prop[gpuid[i]].name, gpuid[i] ,
+                           can_access_peer_1_0 ? "Yes" : "No");
+                }
+            }
+        }
+    }
+
+    // csv masterlog info
+    // *****************************
+    // exe and CUDA driver name
+    printf("\n");
+    std::string sProfileString = "deviceQuery, CUDA Driver = CUDART";
+    char cTemp[128];
+
+    // driver version
+    sProfileString += ", CUDA Driver Version = ";
+#if defined(WIN32) || defined(_WIN32) || defined(WIN64) || defined(_WIN64)
+    sprintf_s(cTemp, 10, "%d.%d", driverVersion/1000, (driverVersion%100)/10);
+#else
+    sprintf(cTemp, "%d.%d", driverVersion/1000, (driverVersion%100)/10);
+#endif
+    sProfileString +=  cTemp;
+
+    // Runtime version
+    sProfileString += ", CUDA Runtime Version = ";
+#if defined(WIN32) || defined(_WIN32) || defined(WIN64) || defined(_WIN64)
+    sprintf_s(cTemp, 10, "%d.%d", runtimeVersion/1000, (runtimeVersion%100)/10);
+#else
+    sprintf(cTemp, "%d.%d", runtimeVersion/1000, (runtimeVersion%100)/10);
+#endif
+    sProfileString +=  cTemp;
+
+    // Device count
+    sProfileString += ", NumDevs = ";
+#if defined(WIN32) || defined(_WIN32) || defined(WIN64) || defined(_WIN64)
+    sprintf_s(cTemp, 10, "%d", deviceCount);
+#else
+    sprintf(cTemp, "%d", deviceCount);
+#endif
+    sProfileString += cTemp;
+
+    // Print Out all device Names
+    for (dev = 0; dev < deviceCount; ++dev)
+    {
+#if defined(WIN32) || defined(_WIN32) || defined(WIN64) || defined(_WIN64)
+        sprintf_s(cTemp, 13, ", Device%d = ", dev);
+#else
+        sprintf(cTemp, ", Device%d = ", dev);
+#endif
+        cudaDeviceProp deviceProp;
+        cudaGetDeviceProperties(&deviceProp, dev);
+        sProfileString += cTemp;
+        sProfileString += deviceProp.name;
+    }
+
+    sProfileString += "\n";
+    printf("%s", sProfileString.c_str());
+
+    printf("Result = PASS\n");
+
+    // finish
+    // cudaDeviceReset causes the driver to clean up all state. While
+    // not mandatory in normal operation, it is good practice.  It is also
+    // needed to ensure correct operation when the application is being
+    // profiled. Calling cudaDeviceReset causes all profile data to be
+    // flushed before the application exits
+    cudaDeviceReset();
+    return 0;
+}

pkgs/cudainfo/default.nix

@@ -0,0 +1,43 @@
+{
+  stdenv
+, cudatoolkit
+, cudaPackages
+, autoAddDriverRunpath
+, strace
+}:
+
+stdenv.mkDerivation (finalAttrs: {
+  name = "cudainfo";
+  src = ./.;
+  buildInputs = [
+    cudatoolkit # Required for nvcc
+    cudaPackages.cuda_cudart.static # Required for -lcudart_static
+    autoAddDriverRunpath
+  ];
+  installPhase = ''
+    mkdir -p $out/bin
+    cp -a cudainfo $out/bin
+  '';
+  passthru.gpuCheck = stdenv.mkDerivation {
+    name = "cudainfo-test";
+    requiredSystemFeatures = [ "cuda" ];
+    dontBuild = true;
+    nativeCheckInputs = [
+      finalAttrs.finalPackage # The cudainfo package from above
+      strace # When it fails, it will show the trace
+    ];
+    dontUnpack = true;
+    doCheck = true;
+    checkPhase = ''
+      if ! cudainfo; then
+        set -x
+        cudainfo=$(command -v cudainfo)
+        ldd $cudainfo
+        readelf -d $cudainfo
+        strace -f $cudainfo
+        set +x
+      fi
+    '';
+    installPhase = "touch $out";
+  };
+})

pkgs/mpich-fix-hwtopo.patch

@@ -1,36 +0,0 @@
-diff --git a/src/util/mpir_hwtopo.c b/src/util/mpir_hwtopo.c
-index 33e88bc..ee3641c 100644
---- a/src/util/mpir_hwtopo.c
-+++ b/src/util/mpir_hwtopo.c
-@@ -200,18 +200,6 @@ int MPII_hwtopo_init(void)
- #ifdef HAVE_HWLOC
-     bindset = hwloc_bitmap_alloc();
-     hwloc_topology_init(&hwloc_topology);
--    char *xmlfile = MPIR_pmi_get_jobattr("PMI_hwloc_xmlfile");
--    if (xmlfile != NULL) {
--        int rc;
--        rc = hwloc_topology_set_xml(hwloc_topology, xmlfile);
--        if (rc == 0) {
--            /* To have hwloc still actually call OS-specific hooks, the
--             * HWLOC_TOPOLOGY_FLAG_IS_THISSYSTEM has to be set to assert that the loaded
--             * file is really the underlying system. */
--            hwloc_topology_set_flags(hwloc_topology, HWLOC_TOPOLOGY_FLAG_IS_THISSYSTEM);
--        }
--        MPL_free(xmlfile);
--    }
-
-     hwloc_topology_set_io_types_filter(hwloc_topology, HWLOC_TYPE_FILTER_KEEP_ALL);
-     if (!hwloc_topology_load(hwloc_topology)) 
-
---- a/src/mpi/init/local_proc_attrs.c
-+++ b/src/mpi/init/local_proc_attrs.c
-@@ -79,10 +79,6 @@ int MPII_init_local_proc_attrs(int *p_thread_required)
-     /* Set the number of tag bits. The device may override this value. */
-     MPIR_Process.tag_bits = MPIR_TAG_BITS_DEFAULT;
-
--    char *requested_kinds = MPIR_pmi_get_jobattr("PMI_mpi_memory_alloc_kinds");
--    MPIR_get_supported_memory_kinds(requested_kinds, &MPIR_Process.memory_alloc_kinds);
--    MPL_free(requested_kinds);
--
-     return mpi_errno;
- }

pkgs/overlay.nix

@@ -11,10 +11,6 @@ final: prev:
       paths = [ pmix.dev pmix.out ];
     };
   in prev.mpich.overrideAttrs (old: {
-    patches = (old.patches or []) ++ [
-      # See https://github.com/pmodels/mpich/issues/6946
-      ./mpich-fix-hwtopo.patch
-    ];
     buildInput = old.buildInputs ++ [
       libfabric
       pmixAll
@@ -56,6 +52,7 @@ final: prev:
   prometheus-slurm-exporter = prev.callPackage ./slurm-exporter.nix { };
   meteocat-exporter = prev.callPackage ./meteocat-exporter/default.nix { };
   upc-qaire-exporter = prev.callPackage ./upc-qaire-exporter/default.nix { };
+  cudainfo = prev.callPackage ./cudainfo/default.nix { };
 
   amd-uprof = prev.callPackage ./amd-uprof/default.nix { };
 

secrets/ceph-user.age

@@ -1,19 +1,25 @@
 age-encryption.org/v1
--> ssh-ed25519 AY8zKw xeyzSqfio6SMS9SqywR+7II80D12Oha9T5zOgAIABSQ
-ST26VaF2G1xv9l7d3jWKG32ssOivfwx+p9jLLV7ZFnU
--> ssh-ed25519 sgAamA HrRx+x7NjXKVDaealWFo+Q8zMAdzoj6nTBxw0KMi3jE
-nlcEVTDTe1mPeS16/t9GYRnSSkm5EjpeiBZPIC/2f8U
--> ssh-ed25519 HY2yRg NDp5vUeX35rDV78DFQi9fsc71pQNVE8YQ1StCp+YjTg
-MdUAWHd1k6Jed2pp7Wct/DgF6ShqXFwNxPaXeBOLAcs
--> ssh-ed25519 tcumPQ d0zVVB8t7W9KUapOsnsrvpAj7LgM9zS0yCv8SQnF0g8
-aAPaWRTEBEQgmCkRG69NuWZ/lEva7vH+L8ifQSE0Z1I
--> ssh-ed25519 JJ1LWg 4l8GZNdGOSbqKvmKq1q1aPvjeQIwpgbJj4DBYBse7x4
-rNhTiZlwzyOiCLzYRSzJ5AHebbv94dOgl1UyNmDJD8A
--> ssh-ed25519 CAWG4Q vGhwJDLJIAU8BpV6GP8Dnz2pvTAMufY4v4nvrr2O9yw
-hNZZFDYUMPQNM5+qcc5arIgqQw0PXuqq1WWDTpE+EHo
--> ssh-ed25519 xA739A 8eEi9S5dMWPVR4fKVZdV5eHBOJVf2Ap+3qHSYtYHYgc
-GcgzvJiqsNyZTVk12Z0FEnqB4LgfQ1xjKQwXdto1Hjs
--> ssh-ed25519 MSF3dg oUY9IjDR6hi1qbrCV5z5IcYj85cMppxO94iqkD60Eww
-cBzFGrhh+kWjIi0llw2RqACU1pa7XT9kqWkSeAY8VGI
---- q7AaMOj7ZaS+Mf6trWK56o/1q/c2urrQBPAqk4PtATA
-~ù°k-/¿XwÙåËVÖ(ÈÚZ“d\¶Ñtš'qøõé3œ´R²a\yFWá
	À—Q·+±
ŸRjå…—K³aNe
+-> ssh-ed25519 AY8zKw /gmhFOFqOs8IobAImvQVKeM5Y6k0FpuR61/Cu5drVVI
+g9FXJg2oIoien0zJ70FWHwSTM8SBwbpS188S3Swj7EM
+-> ssh-ed25519 sgAamA opPjlWPhSiI0Rd5l7kd204S5FXFLcQcQftyKb7MDmnU
+3XrRDVnglCP+vBwvfd1rP5gHttsGDHyXwbf10a8/kKY
+-> ssh-ed25519 HY2yRg QKZbubM76C3tobPoyCFDRclA9Pzb2fC7s4WOoIgdORc
+K5kckU0KhQFTE6SikJXFJgM41Tco5+VqOsaG0qLrY1Q
+-> ssh-ed25519 fw2Xhg +ohqts8dLFjvdHxrGHcOGxU0dm+V3N//giljHkobpDM
+jR/UzGrfS9lrJ/VeolKLxfzeJAf2fIB2pdIn/6ukqNk
+-> ssh-ed25519 tcumPQ 3DPkDPIQQSVtXSLzIRETsIyXQ0k1o18Evn6vf+l/6R8
+bLXF62OmJjnOT1vvgq3+AcOKKSG5NonrK5EqCVc0Mwo
+-> ssh-ed25519 JJ1LWg 2Wefc7eLolMU5InEmCNTq21Mf71mI0a2N1HgDrlHvy4
+qXFW9CQBnrzubZ0mzS0Io2WGRrwGBkmeYndBTcZn/fM
+-> ssh-ed25519 cDBabA oiH36AoIt/fFFYgnoxtH7OoetP+2/wjtn8qo3RJDSHc
+qKmkxy1aZGP4ZwC0iH7n7hiJ0+rFQYvjQb5O1a1Z0r4
+-> ssh-ed25519 cK5kHw bX3RtO5StMejUYWAaA37fjHA5nO7Xs1vWDQk3yOjs2o
+Egxmcf8FKAd+E5hMLmhV1yQsCo5rJyUazf1szOvpTAM
+-> ssh-ed25519 CAWG4Q oKqqRDJH0w8lsoQBQk0w8PO+z5gFNmSaGBUSumvDp1I
+m1zWp9MfViAmtpbJhqOHraIokDaPKb0DvvO4vAGCTWI
+-> ssh-ed25519 xA739A G26kPOz6sbFATs+KAr7gbDvji13eA1smFusQAOJXMwA
+Sppvz7A103kZoNxoGsd6eXeCvVh7mBE2MRwLFj9O1dY
+-> ssh-ed25519 MSF3dg 55ekNcp+inbUd+GQ/VZ7BoBASaJ8YDqF74CVXy1PUxQ
+aTHLLAbzQPWWld/OT3BKebc6FcmsqMTaWCPBGm1UHic
+--- mVkAMnI9XQhS3fMiFuuXP/yLR9wEG9+Rr8pA4Uc0avY
+çDU ôÜsÌÝÌÜjùüMî”$Î[†M„°Ú [_™K7síju±v<C2B1>DÊ4¸g‚ÚÜ„3øGnÔä£ É½ÌP§7~rZsˆ­

secrets/gitea-runner-token.age

@@ -1,11 +1,13 @@
 age-encryption.org/v1
--> ssh-ed25519 HY2yRg WUMWvyagPalsy7u1RaEFAwJvFowso1/quNBo+nAkxhQ
-OHcebB7koPKhy58A6qngEVNWckkWChyEK3dwgy8EL5o
--> ssh-ed25519 CAWG4Q Yx/HLIryUNE2BaqTl84FrNRy4XLCY2TRkRgbA9k3qU4
-LZljfuLS5yMVVK6N57iC6cKEaFP6Hh2OkvWJjuFg8q0
--> ssh-ed25519 xA739A DOXjPRttSWz51Sr7KfjgKfAtaIYMo3foB1Ywqw9HYDY
-CA5puXK/1HDOitA2XHBI3OdKmZ7BzHst4DyuWGMC6hE
--> ssh-ed25519 MSF3dg +2LetdIiIZUk7wtHNS1tYsLo4ypwqZ9gpg77RQrnzHU
-yIUu8BVbF3dhUx3531RR50/cJQd9gd8VfKUQzEeT/iQ
---- oY/wQ+RjZO2CmKZtbQ0yOVZ5fv2+AlvvkRu1UDfCNAA
-_8`G©=C7@x&<26>¢\Ft·)ø¹ç„ÙcPe˜Ã%êÖ½[zX-0Ã[¡í¡ɲ½ótz›’;%áÊé•Ð~ÂH0ªØƒ*XDØ;ð€
+-> ssh-ed25519 HY2yRg gKGxsjHfpiRDQ6Tuvcx7pjKgrVUGweotuplLYwCGvik
+DSz9j/stVyB1lXpVP+kg+H+RDgSftREGFFLQZClC3kI
+-> ssh-ed25519 cK5kHw 17DpKekfNVy4V742QSd61r2w6iawtOJR7Ct3UflDXio
+hsqTEPCYjHKvndMWPl4GpG23CzjGgVrS+cLIymISJHU
+-> ssh-ed25519 CAWG4Q oK01d4pbBqEZVsymSiKijPvJo714xsMSRMbzkssJKiw
+hs0tVFkqtIHXg9jtC2iDgCtefFcWvGJkXB+HJUcqXQs
+-> ssh-ed25519 xA739A KxO+AawfLMERHwzt3YnZRwPFlCfGETma7fo8M+ZtsAY
+eSn0+/rhLQxNKt5xKubKck8Nxun2Sh3eJqBU/hwgzZM
+-> ssh-ed25519 MSF3dg OyaZBLB2kO8fU139lXbbC404gT7IzIWk+BMhYzabBDg
+/fiPFfBJcb+e40+fZbwCw7niF2hh+JxUPiKSiwUSOWg
+--- ycZyGX+Li+LsOuweF9OVPl8aoMaRgp/RdFbDrPszkUs
+µ¾îäYMÐÂ:E OìÈ2€r=É&4Ë°CQΣá<C2A3>hCêðïcb¦^SyŸ©%	»Þx-vC`g”þ–îÿW^<5E>ÕwVG¹œƒ

secrets/gitlab-runner-docker-token.age

@@ -1,11 +1,13 @@
 age-encryption.org/v1
--> ssh-ed25519 HY2yRg 6C5Cv7ILdBrpMkCTT/insUY0kyQWbfgU500Ai8ePOXY
-tMw6ehFrsq2dvDEXkLOJwrNZfI28trlr9uy3xW/fzpA
--> ssh-ed25519 CAWG4Q x/j+364IYURgt7fhIPBzabbWMEg08nX8MRrJM/1Q6RU
-AL5Ut2rDr3UXcQXMZJ53ZMf5wMHmT83whx0ntJfW/WU
--> ssh-ed25519 xA739A QjXftBsoGV1rVeHSKcsjp+HMpRVsaHOeeGdDcF6ZWg4
-ovVoYPaPn3liGPAxHWY37CBIUFjAXurv6jMWs2He3HQ
--> ssh-ed25519 MSF3dg FG0CQOj9fRlneW5QrWiy5ksRpicUwHqX9QMpZWhDImw
-L20n1vZRepsRPT4xM6TO6PcI/MJxw4mBLUF0EPv9Uhs
---- DEi7iuzkniq0JPatJ5f2KhrhxWid7ojHpvNfUCGxFtk
-”½%	nœø!;^Q›
rqGµ:‰jC.8l¶|ù”o¥Öñ$LYyþN’b¥«˜:Æ{âã¶fÞ§xTS\ètÕU®ò\Fª)%™™“KL¬ã™‡p¼:>®¾‹ë&ž)­QÕ1ŒH܃V¿SÞ‘£nð
+-> ssh-ed25519 HY2yRg U2KQWviZIVNemm9e8h7H+eOzoYNxXgLLS3hsZLMAuGk
+6n5dH1McNzk3rscP4v2pqZYDWtUFMd15rZsEd/mqIFM
+-> ssh-ed25519 cK5kHw Ebrj/cpz1cFWAYAV9OxgyyH85OEMUnfUIV66p7jaoFY
+6J7hWqODtS/fIF4BpxhxbrxZq5vbolvbLqRKqazT02M
+-> ssh-ed25519 CAWG4Q mXqoQH9ycHF7u0y8mazCgynHxNLxTnrmQHke+2a5QCc
+mq6PdSF+KOqthuXwzTCsOQsi5KG0z1wHUck+bSTyOBY
+-> ssh-ed25519 xA739A TADeswueqDEroZWLjMw3RDNwVQ2xRD+JUMVZENovn0M
+KFlnSjVFbjc+ZsbY8Ed7edC5B01TJGzd/dSryiLArPc
+-> ssh-ed25519 MSF3dg Pq+ZD8AqJGDHDbd4PO1ngNFST8+6C2ghZkO/knKzzEc
+wyiL/u38hdQMokmfTsBrY7CtYwc+31FG4EDaqVEn31U
+--- 1z4cOipayh0zYkvasEVEvGreajegE/dqBV7b6E7aFh0
+³‚ƒ¾R–@×/i–I'Òù<C392>Nxšr"œ`¶O­³Üy£<79>8ÅÜ \/öÔIÂñD¢`Ùß“Šü랦uy¨¢Þ:9Lt¤û”³Ø‹€æú±îû·AUüñ€èÏìú`ë;­q8žGLU#îiãyù•i¼Úœ

secrets/secrets.nix

@@ -2,6 +2,8 @@ let
   keys = import ../keys.nix;
   adminsKeys = builtins.attrValues keys.admins;
   hut = [ keys.hosts.hut ] ++ adminsKeys;
+  fox = [ keys.hosts.fox ] ++ adminsKeys;
+  apex = [ keys.hosts.apex ] ++ adminsKeys;
   mon = [ keys.hosts.hut keys.hosts.tent ] ++ adminsKeys;
   tent = [ keys.hosts.tent ] ++ adminsKeys;
   # Only expose ceph keys to safe nodes and admins
@@ -19,7 +21,12 @@ in
   "tent-gitlab-runner-pm-docker-token.age".publicKeys = tent;
   "tent-gitlab-runner-pm-shell-token.age".publicKeys = tent;
   "tent-gitlab-runner-bsc-docker-token.age".publicKeys = tent;
+  "vpn-dac-login.age".publicKeys = tent;
+  "vpn-dac-client-key.age".publicKeys = tent;
 
   "ceph-user.age".publicKeys = safe;
   "munge-key.age".publicKeys = safe;
+
+  "wg-fox.age".publicKeys = fox;
+  "wg-apex.age".publicKeys = apex;
 }

secrets/tent-gitlab-runner-pm-shell-token.age

@@ -1,13 +1,13 @@
 age-encryption.org/v1
--> ssh-ed25519 G5LX5w V9bHLoGuY4stRwbzVS9Qa0L9yoY+UoCoXc+dJJQW/Ag
-2ut9GfdJ3KBCqZRaloZCQsl8MLfaZAZxqj6JtPJzu2k
--> ssh-ed25519 CAWG4Q OAqnIfMECpKglZ7aF9tv/PQinG1Ou2+IEZ+nf4dtQjg
-dANdMLe4iI0d6Xd/dIMpZK+mgw2+VmJFQScHaIxD7WI
--> ssh-ed25519 xA739A nVNF4Y6VSa5PP6FFBJpVmoFYYseoFx5F2wJU+Pwk+Xk
-A5CiuTSNlX9Y76qhYgblBdJl3zPhtjWho2oL5/sIKu0
--> ssh-ed25519 MSF3dg /WMsGnBGzquIMyw06gHKpSS4OUxheulT59kxi+/pxxU
-ppwcv7RLzUbQUM7j0Tb9rRVT9XyPMhqYr2fr4S0nTJY
---- zOe0Ko0oxArbmxePMPDVAT0pDju7IeOAih7sNrDcoVs
-iÜkªA
-hODV‹w!ËÕØE݈ƒÔ+±§`í¬<C3AD>ÅCî©5<C2A9>L<EFBFBD>A‹t<1A>M^
˜E<ÏHI–_ÍnnàÃÕoÁ?ój-ö
-A³nԔίË>ZÕòzšë…dT½Ìb"‘(@‹§{_Ú<5F>C
+-> ssh-ed25519 G5LX5w 1KfTmTRP3iSdcclf/FuIpFWpy1tgKs5ED+qSYWo7inY
+RX6Q1nLFF/yiVLpkWrl0BI0PpLoBi753+y8l/AXjNE4
+-> ssh-ed25519 cK5kHw TP7+OQpQSNuyArnUo1C97J3P3oB0YtzCEPeVvlzsYHE
+Bsy5KPNHTVNHnF1sxOvlfJq3CNMVFaXdYkRG2vSj7qM
+-> ssh-ed25519 CAWG4Q eQyzwNaH6CfaYIjs8abEuQxt6vxRXsGz69UletMUVDE
+FDcynPO7xg4PWez5Z8gTg5LyE0Wgb3zT9i3Kon67QsU
+-> ssh-ed25519 xA739A 2JuLai2fUu3dZBydS8cMrLrEUIUkz4NNaiupoBOtTwU
+sdM3X+XRzysop7yqa76Z7FAwTHOj91STCtZvfIgCdB0
+-> ssh-ed25519 MSF3dg fSPkiWnpInX1V5p3afPCoPotcGFoWFiOMPThtY927lc
+8v7E/3l0xA2VWZPXzkN4NmnaA0KJutLMurn/ZXZmhxA
+--- MQkyBx9hT4ILYXKoZT18PWny1QbDFymcZr63zjMN/qQ
+-bÉ#ÔÛM.Â@ªtŠ<74>ñŵ}+ÏŒ#@®«“ûïkéy­±Ê?vÙänÅÚT¤+<2B>é<EFBFBD>[´Q<C2B4> gA©<41>ó
"ãqh¯·á§]WÂVox¦D](S%øIU¿“â_f¬•2Á§dç[›Ö<>pS`¿ 

secrets/vpn-dac-login.age

@@ -0,0 +1,14 @@
+age-encryption.org/v1
+-> ssh-ed25519 G5LX5w SRJhNenoQXbT1FgX3TMPnVH5P6oe2eHot+M1YsEjsEk
+hfTSLgKi98Eh7JK5o7x2POpTEtQlQCpEa3keUFYCuME
+-> ssh-ed25519 cK5kHw z5TwWJTkvx7HztjXHJW/aCOtOfPrQaLP0gyIT7rXcyU
+b4NCpHfasgvkLLr+6LcWUl60p59aSNnfp3bl2OFYXo0
+-> ssh-ed25519 CAWG4Q 4VpS1/OnFe8nxcQbRTKNhjsh/ZQ5cbhSMXwK/jjQ+3o
+WF9wvOkqVml4UcEzyzeumKuUwCwwr2zvKLMg+PCB8nk
+-> ssh-ed25519 xA739A 67FhuJ070jBVMt/xbKHWhfri6iIm0FyaFvzQabsvFBM
+1G5/913dDv/r/6p1x/c5YiUnZzrX/LvIj33KW+PN0KU
+-> ssh-ed25519 MSF3dg Bj/yB4N2wkyHCHC22tcjjJAA4ebSamN0Z4UVX3ZnryI
+6D/ZgTs+j+MGDAbPU5zyK0i9zN6tQy68IcOnQZ27mYg
+--- 169erk3ICSYLs4FPEuXCn7QlekWhsmSn0Lr+/R14I5Q
+§¼ã<EFBFBD>¦Ò½3‰sø
+w ‰4DºÏb.ÿ<>Ö"|€…ó)"¹»¹<C2BB>¤½æ¥;Œ.‡É«7)¤LeCù=SØŸ

secrets/wg-fox.age

@@ -0,0 +1,14 @@
+age-encryption.org/v1
+-> ssh-ed25519 cDBabA heyW9/cxgwFX9IexQIXjAQDWGQPNcMXcArQp2Rxsqx4
+o9MQ7EH8PDDjsJdpH9F3Xq2zUoaDAJQlfFmYucSFs6Y
+-> ssh-ed25519 cK5kHw Sza4pos7K3qW3omEeyidI/jszJNf9smemSZnUJfCIww
+D6vazXki7hIYraIuSiGPS+FPbkFUwHhHWDf52OhEIMg
+-> ssh-ed25519 CAWG4Q YexIHueOIMmIN8JIDyNUOKBkyz/k18HqV3hTXh48KlM
+xh8UJzzWT6ByN+Dpn4JrMNsjGC/uc/v6LynwjBDz9NQ
+-> ssh-ed25519 xA739A KySG3TXdqfCMUkVEDGa74B0op745s3XGYxFLyAXSQAc
+5EI/yb5ctW9Qu18bHm3/sK97kwGcKzzmWvPSCWm89XA
+-> ssh-ed25519 MSF3dg MNxnNj0fHmri8ophexXPNjRUBUWrzcuk5S1mucxUMTE
+GVFWXtISEU8ZmlwL4nh4weAgfGrt2GHX0DTzbpS6zg8
+--- UdrqkYG2ZApAuwdZeNhC50NP2rkD/Ol6y8nJa4RHx7Y
+·Ü»¸m(Éó”>æHýY87ª’G¸+*ÿô†²9V™.ÿ²<C3BF>ÿ‰›ÝpÎOo‰=+哇ÐP0±þ{â)<29>¡‹÷Ö>®z3P^
+u

web/content/access/index.md

@@ -5,18 +5,12 @@ description: "Request access to the machines"
 
 ![Cave](./cave.jpg)
 
-Before requesting access to the jungle machines, you must be able to access the
-`ssfhead.bsc.es` node (only available via the intranet or VPN). You can request
-access to the login machine using a resource petition in the BSC intranet.
-
-Then, to request access to the machines we will need some information about you:
+To request access to the machines we will need some information:
 
 1. Which machines you want access to ([hut](/hut), [fox](/fox), owl1, owl2, eudy, koro...)
-1. Your user name and user id (to match the NFS permissions)
+1. Your user name (make sure it matches the one you use for the BSC intranet)
 1. Your real name and surname (for identification purposes)
 1. The salted hash of your login password, generated with `mkpasswd -m sha-512`
 1. An SSH public key of type Ed25519 (can be generated with `ssh-keygen -t ed25519`)
 
-Send an email to <jungle@bsc.es> with the details, or directly open a
-merge request in the [jungle
-repository](https://pm.bsc.es/gitlab/rarias/jungle/).
+Send an email to <jungle@bsc.es> with the details.

web/content/fox/_index.md

@@ -21,17 +21,28 @@ the detailed specifications:
 
 ## Access
 
-To access the machine, request a SLURM session from [hut](/hut) using the `fox`
-partition:
+To access the machine, request a SLURM session from [apex](/apex) using the `fox`
+partition. If you need the machine for performance measurements, use an
+exclusive reservation:
 
-    hut% salloc -p fox
+    apex% salloc -p fox --exclusive
 
-Then connect via ssh:
+Otherwise, specify the CPUs that you need so other users can also use the node
+at the same time:
 
-    hut% ssh fox
+    apex% salloc -p fox -c 8
+
+Then use srun to execute an interactive shell:
+
+    apex% srun --pty $SHELL
     fox%
 
-Follow [these steps](/access) if you don't have access to hut or fox.
+Make sure you get all CPUs you expect:
+
+    fox% grep Cpus_allowed_list /proc/self/status
+    Cpus_allowed_list:	0-191
+
+Follow [these steps](/access) if you don't have access to apex or fox.
 
 ## CUDA
 
@@ -89,9 +100,8 @@ Then just run `nix develop` from the same directory:
 
 The machine has several file systems available.
 
-- `$HOME`: Mounted via NFS across all nodes. It is slow and has low capacity.
-  Don't abuse.
-- `/ceph/home/$USER`: Shared Ceph file system across jungle nodes. Slow but high
-  capacity. Stores three redundant copies of every file.
+- `/nfs/home`: The `/home` from apex via NFS, which is also shared with other
+  xeon machines. It has about 2 ms of latency, so not suitable for quick random
+  access.
 - `/nvme{0,1}/$USER`: The two local NVME disks, very fast and large capacity.
 - `/tmp`: tmpfs, fast but not backed by a disk. Will be erased on reboot.