From dafbe79ae265bb3e276bd757da5a7fba2cef783d Mon Sep 17 00:00:00 2001 From: KeithMnemonic Date: Wed, 4 Sep 2024 14:34:41 -0400 Subject: [PATCH] Add H200 with PCI ID 0x233510DE This PR updates the config-default.yaml to include the GPU ID for the NVIDIA H200. In addition, other changes were added to sync this file with the assets/state-mig-manager/0400_configmap.yaml in https://github.com/NVIDIA/gpu-operator/ These files should contain the same information but have drifted over time. Signed-off-by: KeithMnemonic --- deployments/systemd/config-default.yaml | 132 ++++++++++++++++++------ 1 file changed, 98 insertions(+), 34 deletions(-) diff --git a/deployments/systemd/config-default.yaml b/deployments/systemd/config-default.yaml index a7e8f279..1ddf3049 100644 --- a/deployments/systemd/config-default.yaml +++ b/deployments/systemd/config-default.yaml @@ -4,6 +4,11 @@ mig-configs: - devices: all mig-enabled: false + all-enabled: + - devices: all + mig-enabled: true + mig-devices: {} + # A100-40GB, A800-40GB all-1g.5gb: - devices: all @@ -44,7 +49,7 @@ mig-configs: # H100-80GB, H800-80GB, A100-80GB, A800-80GB, A100-40GB, A800-40GB all-1g.10gb: # H100-80GB, H800-80GB, A100-80GB, A800-80GB - - device-filter: ["0x233110DE", "0x233010DE", "0x232210DE", "0x20B210DE", "0x20B510DE", "0x20F310DE", "0x20F510DE", "0x232410DE"] + - device-filter: ["0x233010DE", "0x233110DE", "0x232210DE", "0x20B210DE", "0x20B510DE", "0x20F310DE", "0x20F510DE", "0x232410DE"] devices: all mig-enabled: true mig-devices: @@ -126,75 +131,91 @@ mig-configs: mig-devices: "4g.24gb": 1 - # PG506-96GB + # H100 NVL, H800 NVL, GH200 all-1g.12gb: - devices: all mig-enabled: true mig-devices: "1g.12gb": 7 - all-2g.24gb: + all-1g.12gb.me: - devices: all mig-enabled: true mig-devices: - "2g.24gb": 3 + "1g.12gb+me": 1 - all-3g.48gb: + all-1g.24gb: - devices: all mig-enabled: true mig-devices: - "3g.48gb": 2 + "1g.24gb": 4 - all-7g.96gb: + all-2g.24gb: - devices: all mig-enabled: true mig-devices: - "7g.96gb": 1 + "2g.24gb": 3 - # H100 NVL 94GB - all-1g.12gb.me: + # H100 NVL, H800 NVL + all-3g.47gb: - devices: all mig-enabled: true mig-devices: - "1g.12gb+me": 1 + "3g.47gb": 2 - all-1g.24gb: + all-4g.47gb: - devices: all mig-enabled: true mig-devices: - "1g.24gb": 4 + "4g.47gb": 1 - all-3g.47gb: + all-7g.94gb: - devices: all mig-enabled: true mig-devices: - "3g.47gb": 2 + "7g.94gb": 1 - all-4g.47gb: + # H100-96GB, PG506-96GB, GH200 + all-3g.48gb: - devices: all mig-enabled: true mig-devices: - "4g.47gb": 1 + "3g.48gb": 2 - all-7g.94gb: + all-4g.48gb: - devices: all mig-enabled: true mig-devices: - "7g.94gb": 1 + "4g.48gb": 1 + + all-7g.96gb: + - devices: all + mig-enabled: true + mig-devices: + "7g.96gb": 1 - # H100-94GB, H100-80GB, H800-80GB, A100-40GB, A100-80GB, A800-40GB, A800-80GB, A30-24GB, PG506-96GB + # H200-141GB, H100-96GB, GH200, H100 NVL, H800 NVL, H100-80GB, H800-80GB, A800-40GB, A800-80GB, A100-40GB, A100-80GB, A30-24GB, PG506-96GB all-balanced: - # A100-40GB, A800-40GB - - device-filter: ["0x20B010DE", "0x20B110DE", "0x20F110DE", "0x20F610DE"] + # H200 141GB + - device-filter: ["0x233510DE"] devices: all mig-enabled: true mig-devices: - "1g.5gb": 2 - "2g.10gb": 1 - "3g.20gb": 1 + "1g.18gb": 2 + "2g.35gb": 1 + "3g.71gb": 1 + + # H100 NVL, H800 NVL + - device-filter: ["0x232110DE", "0x233A10DE"] + devices: all + mig-enabled: true + mig-devices: + "1g.12gb": 2 + "2g.24gb": 1 + "3g.47gb": 1 # H100-80GB, H800-80GB, A100-80GB, A800-80GB - - device-filter: ["0x233110DE", "0x233010DE", "0x232210DE", "0x20B210DE", "0x20B510DE", "0x20F310DE", "0x20F510DE", "0x232410DE"] + - device-filter: ["0x233010DE", "0x233110DE", "0x232210DE", "0x20B210DE", "0x20B510DE", "0x20F310DE", "0x20F510DE", "0x232410DE"] devices: all mig-enabled: true mig-devices: @@ -202,6 +223,15 @@ mig-configs: "2g.20gb": 1 "3g.40gb": 1 + # A100-40GB, A800-40GB + - device-filter: ["0x20B010DE", "0x20B110DE", "0x20F110DE", "0x20F610DE"] + devices: all + mig-enabled: true + mig-devices: + "1g.5gb": 2 + "2g.10gb": 1 + "3g.20gb": 1 + # A30-24GB - device-filter: "0x20B710DE" devices: all @@ -210,8 +240,8 @@ mig-configs: "1g.6gb": 2 "2g.12gb": 1 - # PG506-96GB - - device-filter: "0x20B610DE" + # H100-96GB, PG506-96GB, GH200, H20 + - device-filter: ["0x234210DE", "0x233D10DE", "0x20B610DE", "0x232910DE"] devices: all mig-enabled: true mig-devices: @@ -219,11 +249,45 @@ mig-configs: "2g.24gb": 1 "3g.48gb": 1 - # H100 NVL 94GB - - device-filter: "0x232110DE" - devices: all + # H200-141GB + all-1g.18gb: + - devices: all mig-enabled: true mig-devices: - "1g.12gb": 2 - "2g.24gb": 1 - "3g.47gb": 1 + "1g.18gb": 7 + + all-1g.18gb.me: + - devices: all + mig-enabled: true + mig-devices: + "1g.18gb+me": 1 + + all-1g.35gb: + - devices: all + mig-enabled: true + mig-devices: + "1g.35gb": 4 + + all-2g.35gb: + - devices: all + mig-enabled: true + mig-devices: + "2g.35gb": 3 + + all-3g.71gb: + - devices: all + mig-enabled: true + mig-devices: + "3g.71gb": 2 + + all-4g.71gb: + - devices: all + mig-enabled: true + mig-devices: + "4g.71gb": 1 + + all-7g.141gb: + - devices: all + mig-enabled: true + mig-devices: + "7g.141gb": 1 \ No newline at end of file