NVIDIA GeForce RTX 3060 12GB
Overview
The GeForce RTX 3060 is the budget AI card of its generation, and the reason is its frame buffer: 12 GB on a 192-bit bus, more memory than the 10 GB RTX 3080 that sat three tiers above it. That made it the cheapest new card able to hold a 7B model in half precision, and it is still the entry point people are pointed at for local inference and small fine-tunes. Throughput is modest, roughly 12.7 TFLOPS of FP32 and 25.5 TFLOPS of dense BF16 tensor, and the GeForce tensor restriction means TF32 runs at the plain shader rate rather than double it. INT8 and INT4 are untouched by that restriction, so quantised inference is where this card earns its keep. It draws 170 W and needs no exotic power supply.
Performance
Peak theoretical throughput by precision type
| Precision | Dense | 2:4 Sparse |
|---|---|---|
FP64 64-bit floating point | 0.2TFLOPS | Structured sparsity is a tensor-core feature; this vector precision has no sparse form |
FP32 32-bit floating point | 13TFLOPS | Structured sparsity is a tensor-core feature; this vector precision has no sparse form |
TF32 TensorFloat-32 | 13TFLOPS | 25TFLOPS |
BF16 Brain Float 16 | 25TFLOPS | 51TFLOPS |
FP16 16-bit floating point | 25TFLOPS | 51TFLOPS |
FP8 | No verified data available | The part supports 2:4 sparsity, but the vendor publishes no sparse figure for this precision |
FP6 | No verified data available | The part supports 2:4 sparsity, but the vendor publishes no sparse figure for this precision |
FP4 | No verified data available | The part supports 2:4 sparsity, but the vendor publishes no sparse figure for this precision |
INT8 8-bit integer | 102TOPS | 204TOPS |
INT4 4-bit integer | 204TOPS | 408TOPS |
Per watt divides the dense figure by the 170 W board power. Where only a sparse figure exists, the cell says so.
The GeForce RTX 3060 12GB in the GPU landscape
Peak FP16 TFLOPS (dense) against TDP, single-GPU parts tracked by Flopper
Higher and further left is better: more half-precision throughput for less power.
Specifications
Architecture
Ampere
Form Factor
PCIe
Launch Year
2021
Process Node
8nm
Memory
12 GB GDDR6
Bandwidth
No verified data available
TDP
170 W
Max Power
170 W
CUDA Cores
3,584
Spec Confidence
Official
Full Specifications
| Compute Engine | |
|---|---|
| CUDA Cores | 3,584 |
| Tensor Cores | 112 (3rd Gen) |
| Streaming Multiprocessors | 28 |
| Base Clock | 1.32 GHz |
| Boost Clock | 1.78 GHz |
| Memory | |
| VRAM | 12 GB |
| Memory Type | GDDR6 |
| Bandwidth | No verified data available |
| Interface Width | 192-bit |
| Interconnect & I/O | |
| PCIe | 4.0 x16 |
| Power & Thermal | |
| TDP | 170 W |
| Max Board Power | 170 W |
| Cooling | Active |
| Enterprise Features | |
| ECC Memory | No |
| Sparsity | Yes |
| Compute APIs | CUDA, DirectX 12, Vulkan, OpenGL |
| General | |
| Form Factor | PCIe |
| Architecture | Ampere |
| Process Node | 8nm |
| Launch Year | 2021 |
Datasheet & Resources
Flopper Datasheet
NVIDIA GeForce RTX 3060 12GB specifications, generated from our database. Printable.
Vendor product page
NVIDIA documentation for the GeForce RTX 3060 12GB
NVIDIA Introduces GeForce RTX 3060, Next Generation of the World's Most Popular GPU
NVIDIA · 2021-01-12
NVIDIA Developer Documentation
Technical specs, programming guides
Data Provenance
Every figure traced to a source
Primary Source
- Publisher
- NVIDIA
- Published
- 2021-01-12
Data Quality
- Spec confidence
- Official
- Clock basis
- Boost
- Core precisions with figures
- 6 of 9
- Normalization
- All values in TFLOPS
Rent this GPU
$0.04 / GPU-hr
Lowest of 1 live listing across 1 provider
Price history and every provider on the pricing page. Some links are affiliate links.
Similar GPUs
Frequently Asked Questions
How many TFLOPS does the NVIDIA GeForce RTX 3060 have?
The NVIDIA GeForce RTX 3060 delivers 13 TFLOPS FP32 and 25 TFLOPS FP16 at peak. Flopper does not currently have a verified FP8 throughput figure for it.
What is the power consumption of the NVIDIA GeForce RTX 3060?
The NVIDIA GeForce RTX 3060 has a TDP (Thermal Design Power) rating of 170 watts.
How much memory does the NVIDIA GeForce RTX 3060 have?
The NVIDIA GeForce RTX 3060 is equipped with 12 GB of VRAM. Flopper does not currently have a verified memory bandwidth figure for it.
What architecture is the NVIDIA GeForce RTX 3060 based on?
The NVIDIA GeForce RTX 3060 is based on the Ampere architecture, launched in 2021.
Stay Updated on GPU Releases
Get notified when new GPUs are added or specifications are updated.
No spam, unsubscribe anytime.