Platforms
Yandex Compute Cloud provides various types of physical processors. Choosing a platform guarantees the type of physical processor in the data center and determines the supported vCPU and RAM configurations. You can also add a graphics accelerator (GPU) to a VM. You need to choose a platform each time you create a new VM.
Warning
The ru-central1-d zone does not support VMs based on the Intel Broadwell, Intel Broadwell with NVIDIA® Tesla® V100, Intel Cascade Lake with NVIDIA® Tesla® V100, or AMD EPYC™ with NVIDIA® Ampere® A100 platforms. To move such VMs to ru-central1-d, do one of the following:
- Take a disk snapshot and use it to create a new VM in the
ru-central1-dzone on a different platform. - Stop the VM, change the platform, and move the VM by running
relocate.
Standard platforms
| Platform | Processor | Maximum number of cores (vCPUs)per VM | CPU base clock rate (GHz) |
|---|---|---|---|
Intel Broadwell(standard-v1) |
Intel® Xeon® Processor E5-2660 v4 | 32 | 2.00 |
Intel Cascade Lake(standard-v2) |
Intel® Xeon® Gold 6230 | 80 | 2.10 |
Intel Ice Lake(standard-v3) |
Intel® Xeon® Gold 6338 | 96 | 2.00 |
AMD Zen 3(amd-v1)1 |
AMD EPYC™ 7713 | 128 | 2.00 |
AMD Zen 4(standard-v4a) |
AMD EPYC™ 9654 | 288 | 2.40 |
1 amd-v1 is available to a limited number of users for private testing at the Preview stage and is not free of charge. To request access to the platform, contact your account manager.
High-performance platforms
| Platform | Processor | Maximum number of cores (vCPUs) per VM | CPU base clock rate (GHz) |
|---|---|---|---|
Intel Ice Lake Compute-Optimized(highfreq-v3) |
Intel® Xeon® Processor 6354 | 56 | 3.00 |
AMD Zen 4 Compute-Optimized(highfreq-v4a) |
AMD EPYC™ 9374F | 80 | 3.85 |
Platforms with GPUs
| Platform | Graphics accelerator | Processor | Specifications |
|---|---|---|---|
Intel Broadwell withNVIDIA® Tesla® V100(gpu-standard-v1) |
NVIDIA® Tesla® V100 | Intel® Xeon®Processor E5-2660 v4 | Maximum number of GPUs per VM: 4 Number of vCPUs per GPU: 8 RAM per GPU: 96 GB |
Intel Cascade Lakewith NVIDIA® Tesla® V100(gpu-standard-v2) |
NVIDIA® Tesla® V100 | Intel® Xeon® Gold 6230 | Maximum number of GPUs per VM: 8 Number of vCPUs per GPU: 8 RAM per GPU: 48 GB |
AMD EPYC™with NVIDIA® Ampere® A100(gpu-standard-v3) |
NVIDIA® Ampere® A100 | AMD EPYC™ 7702 | Maximum number of GPUs per VM: 8 Number of vCPUs per GPU: 28 RAM per GPU: 119 GB |
Gen2(gpu-standard-v3i) |
Gen2 | Gen2 | Maximum number of GPUs per VM: 8 Number of vCPUs per GPU: 22.5 or 18 RAM per GPU: 180 or 144 GB |
Intel Ice Lake withNVIDIA® Tesla® T4(standard-v3-t4) |
NVIDIA® Tesla® T4 | Intel® Xeon® Gold 6338 | Maximum number of GPUs per VM: 1 Number of vCPUs per GPU: 4, 8, 16, or 32 RAM per GPU: 16, 32, 64, or 128 GB |
Intel Ice Lake with T4i(standard-v3-t4i) |
T4i | Intel® Xeon® Gold 6338 | Maximum number of GPUs per VM: 1 Number of vCPUs per GPU: 4, 8, 16, or 32 RAM per GPU: 16, 32, 64, or 128 GB |