Table of Contents | ||||||||||
---|---|---|---|---|---|---|---|---|---|---|
|
👁️ Overview
The STFC Cloud provides a number of flavors to users but these fall into 4 categories at the moment.
The larger the flavor you choose the more resources you will be holding and the more power is consumed by the instance you have created. This also prevents other users from using these resources.
It is important that you should use the smallest flavor that meets the requirements of your workloads and if possible delete the instance when you are not using it.
Note |
---|
No flavor has resilient storage. If this is required you should boot from Volume at VM creation time |
Note: our flavors are created in
...
such a way that they will efficiently pack onto the hardware we have - due to this we cannot create additional flavors on demand as that may lead to scheduling difficulties
...
under heavy utilisation.
...
🖥️ Compute flavors
l flavors
(e.g.
...
Note |
---|
C flavors are now deprecated and new VMs will not be able to be created after the 12th of April 2023. It is recommend to use l flavors instead. |
These are flavors with dedicated CPU cores meaning that a VCPU core assigned to your VM is dedicated to you.
These flavors are great for computation and are optimised to either be within a single NUMA node providing great performance or balanced equally across the NUMA nodes in the hypervisor.
l*.*
These are l6.c252, l3.nano, l2.xsmall)
These are flavors which utilize local SSD storage on the hypervisors with all of the storage being allocated to the root disk. These instances currently cannot be live migrated so in the event of hardware errors the machine will likely be lost or down for an extended period.
Otherwise the flavors behave as c*.* flavors
...
Larger l flavors may struggle to live migrate due to capacity availability and disk image sizes so where possible be sure that you can recreate your machines using other methods, such as configuration management.
le flavors
(e.g. le3.small, le2.medium)
These are flavors which utilize local SSD storage on the hypervisors with 100GB being allocated to the root disk and the rest to an ephemeral disk. These instances currently cannot be live migrated so in the event of hardware errors the machine will likely be lost or down for an extended period.
Otherwise the flavors behave as c*.* flavors
...
Larger le flavors may struggle to live migrate due to capacity availability and disk image sizes so where possible be sure that you can recreate your machines using other methods, such as configuration management.
Compute flavor list
Name | RAM (MB) | Disk (GB) | Ephemeral Disk (GB) | VCPUs |
---|---|---|---|---|
l3.nano | 8,192 | 50 | 0 | 2 |
l3.micro | 16,384 | 100 | 0 | 4 |
[dep-]l2.tiny | 30,720 | 200 | 0 | 8 |
l3.tiny | 30,720 | 200 | 0 | 8 |
[dep-]l2.xsmall | 61,440 | 400 | 0 | 16 |
l3.xsmall | 61,440 | 400 | 0 | 16 |
[dep-]le2.small | 125,952 | 100 | 700 | 30 |
le3.small | 125,952 | 100 | 700 | 30 |
[dep-]l2.small | 125,952 | 800 | 0 | 30 |
l3.small | 125,952 | 800 | 0 | 30 |
[dep-]le2.medium | 250,880 | 100 | 1500 | 60 |
le3.medium | 250,880 | 100 | 1500 | 60 |
[dep-]l2.medium | 250,880 | 1600 | 0 | 60 |
l3.medium | 250,880 | 1600 | 0 | 60 |
[dep-]le2.large | 501,760 | 100 | 3100 | 124 |
le3.large | 501,760 | 100 | 3100 | 124 |
[dep-]l2.large | 501,760 | 3200 | 0 | 124 |
l3.large | 501,760 | 3200 | 0 | 124 |
l6.c32 | 122,880 | 650 | 0 | 32 |
l6.c64 | 235,520 | 1300 | 0 | 64 |
l6.c124 | 471,859 | 2700 | 0 | 124 |
l6.c252 | 943,718 | 5400 | 0 | 252 |
💪 GPU flavors
Info |
---|
GPU flavors are not available by default and capacity is limited. To request access please contact STFC cloud support. |
g flavors
(e.g. g-a100.x1, g-rtx4000.x4)
These have the same CPU configuration as cl*.* flavors but have a GPU attached:
g-p4000.* have one or more NVidia Quadro P4000 cards attached (the smallest has 1, the largest has 4)g-rtx4000.* have one or more NVidia Quadro RTX 4000 cards attached (the smallest has 1, the largest has 4)
g-rtx4000-ref.* have one or more NVidia Quadro RTX 4000 cards attached (the smallest has 1, the largest has 4)
these These are old older, out of warranty nodes which have been refitted with new GPUs. For this reason these are primarily reserved for short term use such as training courses. If you have a use case you want to use these for then please get in touch with cloud support.
g-a4000.* have one or more NVidia RTX A4000 cards attached (the smallest has 1, the largest has 4)
g-a4000-ref.* have one or more NVidia RTX A4000 cards attached (the smallest has 1, the largest has 4)
these These are old older out of warranty nodes which have been refitted with new GPUs. For this reason these are primarily reserved for short term use such as training courses. If you have a use case you want to use these for then please get in touch with support.
g-p100v100.* have one or more NVidia Tesla P100 V100 32GB cards attached (the smallest has 1, the largest has 2)g-v100.* have one or more NVidia Tesla V100 32GB cards attached (small has 1,medium has 2 and large has 4)
g-a100.* have one or more NVidia Tesla A100 40GB ) cards attached (the smallest has 1, the largest has 4)
g-a100-n100.* have one or more NVidia Tesla A100 40GB cards attached (the smallest has 1, the largest has 4) with higher core count and more network bandwidth
These are currently reserved for a specific projectprojects.
g-a100-80gb.* have one or more NVidia Tesla A100 80GB cards attached (the smallest has 1, the largest has 8)
g-arc-a770.* have one or two Intel Arc A770 16GB cards
These are very limited in number and are primarily in place for prototyping.
g-amd-w6600.* have one or more AMD W6600 8GB cards (the smallest has 1, the largest has 4)
These are very limited in number and are primarily in place for prototyping.
f flavors
(e.g. f-xilinxu200.x1)
These are similar to the g* flavors but have fpga accelerators rather than GPUs
f-xilinxu200.* have one or more Xilinx U200 FPGA cards
These are very limited in number
GPU flavor list
Name | RAM (MB) | Disk (GB) | VCPUs | GPU(s) |
---|---|---|---|---|
g-a4000-ref.x1 | 91,200 | 700 | 12 | 1 x A4000 |
g-a4000.x1 | 122,880 | 400 | 16 | 1 x A4000 |
g-a4000.x2 | 245,760 | 800 | 32 | 2 x A4000 |
g-a4000.x4 | 491,520 | 1600 | 60 | 4 x A4000 |
g-a4000.x8 | 983,040 | 3200 | 124 | 8 x A4000 |
g-rtx4000.x1 | 92,160 | 700 | 12 | 1 x RTX4000 |
g-rtx4000.x2 | 184,320 | 1400 | 28 | 2 x RTX4000 |
g-rtx4000.x4 | 358,400 | 2800 | 60 | 4 x RTX4000 |
g-rtx4000-ref.x1 | 921,600 | 700 | 12 | 1 x RTX4000 |
g-a100.x1 | 102,400 | 1600 | 12 | 1 x A100 40GB |
g-a100.x2 | 204,800 | 3200 | 28 | 2 x A100 40GB |
g-a100.x4 | 460,800 | 6400 | 60 | 4 x A100 40GB |
g-a100.x1hm | 204,800 | 1600 | 12 | 1 x A100 40GB |
g-a100.x2hm | 460,800 | 3200 | 30 | 2 x A100 40GB |
g-a100.x4hm | 870,400 | 6400 | 60 | 4 x A100 40GB |
g-a100-n100.x1 | 102,400 | 800 | 30 | 1 x A100 40GB |
g-a100-n100.x2 | 225,280 | 1600 | 60 | 2 x A100 40GB |
g-a100-n100.x4 | 450,560 | 3200 | 124 | 4 x A100 40GB |
g-a100-80gb.x1 | 102,400 | 800 | 30 | 1 x A100 80GB |
g-a100-80gb.x2 | 225,280 | 1600 | 60 | 2 x A100 80GB |
g-a100-80gb.x4 | 450,560 | 3200 | 124 | 4 x A100 80GB |
g-a100-80gb-2022.x1 | 122,880 | 400 | 16 | 1 x A100 80GB |
g-a100-80gb-2022.x2 | 245,760 | 800 | 32 | 2 x A100 80GB |
g-a100-80gb-2022.x4 | 491,520 | 1600 | 60 | 4 x A100 80GB |
g-a100-80gb-2022.x8 | 983,040 | 3200 | 124 | 8 x A100 80GB |
g-v100.x1 | 76,800 | 700 | 12 | 1 x v100 |
g-v100.x2 | 174,080 | 1400 | 28 | 2 x v100 |
g-v100.x4 | 358,400 | 2800 | 60 | 4 x v100 |
g-arc-a770-ref.x1 | 158,720 | 1400 | 28 | 1 x Intel Arc A770 |
g-amd-w6600-ref.x1 | 92,160 | 700 | 12 | 1 x AMD w6600 |
f-xilinxu200.x1 | 57,344 | 400 | 16 | 1 x Xilinx U200 |
🕚 Flavor Lifetimes
In general flavors are tied to hardware generations. We run these generations for at least the length of their maintenance contracts.
Ideally we try to run these past the end of their maintenance contracts at a reduced SLA. This means that capacity available will decrease over time and all VMs on these flavors should be considered at risk.
There are several things that may shorten the amount of time that we run past the end of the contract.
Hardware that is particularly prone to failure leading to high maintenance overheads
Needing the rack or power space for increases in capacity, functionality or efficiency
Hardware no longer being supported by the software we need to use
Flavor Prefix | Generation Year (FY) | End Of Maintenance | Estimated Draining Date | Estimated Final EOL Date |
---|---|---|---|---|
[dep-]l2/le2 | 2019/20 | 10/04/2023 | 10/04/2024 | 01/09/2024 |
l3/le3 | 2020/21 | 01/04/2026 | 01/04/2027 | 01/09/2027 |
l4 | 2020/21 | 01/04/2026 | 01/04/2027 | 01/09/2027 |
l5 | 2021/22 | 01/04/2025 | 01/04/2027 | 01/09/2027 |
l6 | 2022/23 | 06/11/2025 | 01/04/2028 | 01/09/2028 |
g-rtx4000 | 2020/21 | 01/04/2024 | 01/04/2026 | 01/09/2026 |
g-rtx4000 | 2021/22 | 01/04/2025 | 01/04/2027 | 01/09/2027 |
g-a4000 | 2022/23 | 01/04/2026 | 01/04/2028 | 01/09/2028 |
g-v100 | 2019/20 | 01/04/2023 | 01/04/2025 | 01/09/2025 |
g-v100 | 2020/21 | 01/04/2024 | 01/04/2026 | 01/09/2026 |
g-a100 | 2020/21 | 01/04/2024 | 01/04/2026 | 01/09/2026 |
g-a100-n100 | 2020/21 | 01/04/2024 | 01/04/2026 | 01/09/2026 |
g-a100-80gb | 2021/22 | 01/04/2025 | 01/04/2027 | 01/09/2027 |
g-a100-80gb-2022 | 2022/23 | 01/04/2026 | 01/04/2028 | 01/09/2028 |
The Draining date is the date when access to the flavors will be restricted
The final EOL date is when the hardware turned off
We will re-evaluate these dates periodically and may revise the draining and EOL dates to be closer to the end of maintenance.
🕸️ Deprecated flavors
c flavors
(e.g. c2.xlarge, c3.small)
Note |
---|
C flavors are now deprecated and new VMs will not be able to be created after the 12th of April 2023. It is recommend to use l flavors instead. |
l2 and le2 flavors
(e.g. l2.large, le2.small)
Note |
---|
L2 flavors are now deprecated and new VMs will not be able to be created after the 10th of April 2024. It is recommend to use l3, l5 or l6 flavors instead. |
these flavors have now been renamed to dep-l2 and dep-le2 to reflect that these are deprecated and should be considered at risk. We will be confirming the Final End Of Life date for these shortly. This will be the date when any remaining running VMs will be terminated.
📚 Related articles
Filter by label (Content by label) | ||
---|---|---|
|
Reviewer | Review period |
---|---|
| 6 months |