2 live offerings · 1 tracked regions
Availability unconfirmed 1 · Unconfirmed 1
Loading
Loading workspaces…
engineering · GPU0104
Specifications with their source, supported precisions, interconnect, benchmark coverage and which published models fit NVIDIA A100 NVLink 80GB — Hyperstack flavor.
| Specification | Value | Source | Kovara registry |
|---|---|---|---|
| Memory capacity | 80GB | Kovara registry · verified 2026-09-25 | used |
| Memory type | HBM2E | Kovara registry · verified 2026-09-25 | used |
| Memory bandwidth | — | Not published | — |
| Board power (TDP) | — | Not published | — |
| Form factor | — | Not published | — |
| NVLink | Yes | Kovara registry · verified 2026-09-25 | used |
| MIG partitioning | Yes | Kovara registry · verified 2026-09-25 | used |
No public benchmark runs yet
Nobody has published a hardware run on this accelerator. Measured numbers appear here once a run of a compute repository is published.
AcceleratorsGPU0104
NVIDIA
Core hardware and lifecycle, drawn from Kovara master data.
What the recorded specifications mean in practice, and what this accelerator is used for.
Specs, explained
HBM (High Bandwidth Memory) is stacked in layers and sits right next to the processor on the same package. That short, wide path is what gives accelerators with HBM their memory bandwidth, so they are usually chosen for work limited by how fast data reaches the cores, such as large-model inference and HPC simulation.
VRAM is the working space on the card. Model weights, activations and datasets have to fit in it, or be split across several GPUs. Rule of thumb for the weights alone at 16-bit precision (about 2 GB per billion parameters): up to roughly 40 billion parameters, before activations and cache. This is arithmetic on the recorded VRAM, not a benchmark.
Ampere brought third-generation Tensor Cores, TF32 precision and Multi-Instance GPU (MIG) to its data-center parts. (General context about the architecture, not a statement about this specific card.)
Launched in 2020. Kovara's records list it as a previous-generation part.
NVLink is a direct GPU-to-GPU link. Where it is supported, GPUs in the same server can exchange data faster than over PCIe alone, which matters when one training job is spread across several GPUs.
Multi-Instance GPU (MIG) can divide one physical GPU into isolated slices, each with its own memory and compute, so several smaller workloads can share a card.
These are the uses Kovara has recorded for this part. The notes explain the workload in general terms; they are not performance claims.
Figures are Kovara master data, last verified 25 Sept 2026. Explanations describe the technology in general; they are not benchmarks and say nothing about price or availability.
Explore this accelerator on ComputePublic Kovara observations; not a claim of complete or universal market coverage. Incompatible raw price units are excluded.
Observed GPU-hour range
USD / GPU-hour
Spread $0.42 USD / GPU-hour
Providers represented by current Kovara-tracked offerings.
2 live offerings · 1 tracked regions
Availability unconfirmed 1 · Unconfirmed 1
Regions represented by current offerings; this does not describe complete global availability.
2 live offerings · 1 providers
Current configurations, recorded availability, confirmed capacity and the evidence behind each price.
| Provider | Region | Configuration | Availability | Price | Commercial model | Trust / evidence | Action |
|---|---|---|---|---|---|---|---|
| OFR-0705Hyperstack | Unspecified, Multiple / Unspecified | NVIDIA A100 NVLink 80GB — Hyperstack flavor8 GPUs per node | Availability unconfirmedConfirmed capacity: Not confirmed | $0.98 USD / GPU-hourOriginal: USD 0.98 GPU/hour from | ReservedGPU VM · Starting hourly reservation rate | Confidence AChecked 49 days ago · StaleConfirmed 29 Sept 2026Source: SRC000002 | Inspect |
| OFR-0704Hyperstack | Unspecified, Multiple / Unspecified | NVIDIA A100 NVLink 80GB — Hyperstack flavor8 GPUs per node | UnconfirmedConfirmed capacity: Not confirmed | $1.40 USD / GPU-hourOriginal: USD 1.40 GPU/hour | On-DemandGPU VM · Per minute | Confidence AChecked todayConfirmed 29 Sept 2026Source: SRC000002 | Inspect |
Recorded hardware attributes. Not tracked is distinct from No.
Public evidence and confidence signals supporting Kovara-tracked coverage.