Graphics servers with Tesla A100
All graphics servers with Tesla A100 are based on two Intel Xeon Gold 3rd generation 6336Y CPUs with a base clock frequency of 2.4 GHz and a maximum clock frequency with Turbo Boost technology of 3.6 GHz.
Each processor contains two Intel® AVX-512 units and supports Intel® AVX-512 Deep Learning Boost functions. This set of instructions speeds up multiplication and addition operations with reduced accuracy, which are used in many internal cycles of the deep learning algorithm.
Each server has up to 4096 GB of DDR4 ECC Reg 3200 MHz RAM. Local storage with a total capacity of 1920 GB is organized on Samsung solid-state drives, designed specifically for data centers.
GPU Tesla A100
Tesla A100 GPU provides unsurpassed acceleration for AI tasks, data analysis and for solving the most complex computing tasks. The A100 is the most productive integrated platform for AI and HPC, allowing you to get real-time results and deploy scalable solutions.
When training Deep Learning algorithms, tensor cores with Tensor Float (TF32) support increase performance by 20 times without requiring changes in the code and speed up the automatic function of working with different accuracies, including FP16, by 2 times.
Double-precision tensor cores provide the greatest performance for high-performance computing since the advent of the GPU. HPC applications can also use TF32 to achieve up to 11 times the throughput for precision operations.
Data science specialists need to analyze and visualize large datasets, extracting valuable information from them. Servers equipped with A100 graphics accelerators provide the necessary computing power, thanks to large amounts of high-speed memory with high bandwidth, to cope with these workloads.
| Video memory capacity |
80 GB |
| Type of video memory |
HBM2e |
| Memory bandwidth |
1935 Gb/s |
| Encode/decode |
1 encoder, 2 decoder (+AV1 decode) |
GPU performance benchmarks
Performance benchmarks results in a virtual environment for 1 Tesla A100 graphics card.
-
OctaneBench 2020
pts
-
Matrix multiply example
GFlop/s
-
Hashcat bcrypt
H/s
Configurations with Tesla A100 80 GB with NVLink
Deliver strong performance for deep learning, scientific computing, and large-scale model inference. Unlike single A100 setups, these configurations link two (or more) GPUs via NVLink — a high-speed interconnect providing up to 600 GB/s of bandwidth per GPU pair, vastly outperforming PCIe 4.0 and enabling near-zero-latency data exchange between GPUs.
The key advantage is unified memory address space. Two A100 GPUs with NVLink provide up to 160 GB of shared HBM2e memory, which is critical for running large LLMs, MoE models, and complex HPC applications without node-level sharding. This is especially important when working with long contexts, training foundation models, or running multi-GPU simulations where gradient and activation exchange must not become a bottleneck.
The Ampere architecture, third-generation Tensor Cores, support for TF32, FP64, FP16, and INT8, plus MIG (Multi-Instance GPU) technology make the A100 a versatile platform: from high-precision scientific computations to energy-efficient production inference. In NVLink configurations, these capabilities are fully realized — without losses from inter-node communication.
If you need powerful GPUs for real AI workloads, you can host a server in immers.cloud. This is the ideal solution for those seeking a server for neural network training with guaranteed performance, full resource isolation, and enterprise-grade infrastructure support without the need to purchase and maintain your own hardware.
↔
Custom configuration
100% Performance
Each physical core or GPU adapter is dedicated to a single client.
This means:
- 100% vCPU time is available;
- Physical pass-through of GPUs inside virtual servers;
- Reduced storage and network load on hypervisors, delivering more storage and network performance to clients.
Up to 75,000 IOPS1 for RANDOM READ and up to 20,000 IOPS for RANDOM WRITE on Virtual Servers with local SSDs
Up to 70 000 IOPS1 for RANDOM READ and up to 60 000 IOPS for RANDOM WRITE on Virtual Servers with block storage volumes
You can be confident that Virtual Servers do not share vCPU or GPU resources with one another.
Ready-made OS images with the required software
Create virtual servers by utilizing our pre-configured OS images with either Windows or Linux, along with specialized pre-installed software.
View all the pre-installed images
in the Marketplace.
Pure OpenStack API
Developers and system administrators can manage the cloud using the full OpenStack API.
Authenticate ninja_user example:
$ curl -g -i -X POST https://api.immers.cloud:5000/v3/auth/tokens \
-H "Accept: application/json" \
-H "Content-Type: application/json" \
-H "User-Agent: YOUR-USER-AGENT" \
-d '{"auth": {"identity": {"methods": ["password"], "password": {"user": { "name": "ninja_user", "password": "ninja_password", "domain": {"id": "default"}}}}, "scope": {"project": {"name": "ninja_user", "domain": {"id": "default"}}}}}'
Create ninja_vm example:
$ curl -g -i -X POST https://api.immers.cloud:8774/v2.1/servers \
-H "Accept: application/json" \
-H "Content-Type: application/json" \
-H "User-Agent: YOUR-USER-AGENT" \
-H "X-Auth-Token: YOUR-API-TOKEN" \
-d '{"server": {"name": "ninja_vm", "imageRef": "8b85e210-d2c8-490a-a0ba-dc17183c0223", "key_name": "mykey01", "flavorRef": "8f9a148d-b258-42f7-bcc2-32581d86e1f1", "max_count": 1, "min_count": 1, "networks": [{"uuid": "cc5f6f4a-2c44-44a4-af9a-f8534e34d2b7"}]}}'
STOP ninja_vm example:
$ curl -g -i -X POST https://api.immers.cloud:8774/v2.1/servers/{server_id}/action \
-H "Accept: application/json" \
-H "Content-Type: application/json" \
-H "User-Agent: YOUR-USER-AGENT" \
-H "X-Auth-Token: YOUR-API-TOKEN" \
-d '{"os-stop" : null}'
START ninja_vm example:
$ curl -g -i -X POST https://api.immers.cloud:8774/v2.1/servers/{server_id}/action \
-H "Accept: application/json" \
-H "Content-Type: application/json" \
-H "User-Agent: YOUR-USER-AGENT" \
-H "X-Auth-Token: YOUR-API-TOKEN" \
-d '{"os-start" : null}'
SHELVE ninja_vm example:
$ curl -g -i -X POST https://api.immers.cloud:8774/v2.1/servers/{server_id}/action \
-H "Accept: application/json" \
-H "Content-Type: application/json" \
-H "User-Agent: YOUR-USER-AGENT" \
-H "X-Auth-Token: YOUR-API-TOKEN" \
-d '{"shelve" : null}'
DELETE ninja_vm example:
$ curl -g -i -X DELETE https://api.immers.cloud:8774/v2.1/servers/{server_id} \
-H "User-Agent: YOUR-USER-AGENT" \
-H "X-Auth-Token: YOUR-API-TOKEN"
Documentation
Subscribe to our newsletter
Get notifications about new promotions and special offers by email.
