Tesla V100 PCIe 32GB Specifications
The PCIe V100. Lower clocks than SXM2 and NVLink only via bridge, but it dropped into standard servers — which is how Volta reached most people.
Buy it for
Used FP64 work and older CUDA codebases. Volta introduced tensor cores, so unlike the Pascal Teslas it still runs mixed-precision training meaningfully faster than FP32.
Look elsewhere if
You need FP8, BF16 throughput, or a memory footprint above 32 GB. It also predates the Transformer Engine, so current LLM training recipes lose most of their advantage on it.
Will it fit your system?
Passively cooled — needs server airflow
This card has no fans. It relies on a chassis forcing air through its heatsink. In a normal desktop it will throttle within minutes and can overheat outright. It needs a rack server with high-static-pressure fans, or an aftermarket shroud and blower.
- Form factor
- PCIe dual-slot
- Cooling
- passive (no fans)
- Power connectors
- 1 × 8-pin EPS (CPU-style — not a PCIe 8-pin)
- Board power
- 250 W
- Suggested PSU
- 650 W or better
- Interface
- PCIe Gen 3
PSU figure assumes one card plus a mainstream CPU and drives, with headroom for transient spikes — modern cards can draw well above their rated TDP for milliseconds, which is what trips an undersized supply.
Supported formats & features
What the Tesla V100 PCIe 32GB does and does not support in hardware.
Blackwell-generation feature — B200 has it
Predates hardware FP8; use INT8 or BF16 instead — H100 has it
FP16 only, so training recipes may need loss scaling
Not available on this architecture
Quantised inference supported
Not supported on this generation
Precision must be managed manually
300 GB/s GPU-to-GPU
Compute-only die — no ray tracing hardware
Corrects single-bit errors — expected for long training runs
What fits in memory
Largest model that fits on one Tesla V100 PCIe 32GB, assuming ~20% of the 32 GB is reserved for KV cache and activations.
| Model |
|---|
| Llama 8B |
| Llama 70B |
| Mixtral 8x22B |
| Llama 405B |
Weights only. Long contexts, large batch sizes and optimiser state during training all need substantially more — full fine-tuning typically needs 3–4× the weight memory.
Workload character
Plenty of bandwidth per unit of compute. Memory-bound work — single-stream token decode above all — runs close to its ceiling, and compute is usually what runs out first.
Cost to run
Electricity for one card at 100% duty, including cooling and PSU overhead (3,066 kWh/year).
Where it sits in the line
Availability & market
Plentiful and falling as fleets retire. Sane used buy for FP64 simulation; poor value per watt for LLM work.
Discontinued around 2023. Expect secondary-market stock only, with no manufacturer warranty and limited driver support ahead.
Memory
- Capacity
- 32 GB
- Type
- HBM2
- Bandwidth
- 900 GB/s
- Bus width
- 4,096-bit
- GB per TFLOPMemory relative to compute — higher favours large models
- 0.286
Compute
- FP64
- 7 TFLOPS
- FP32
- 14 TFLOPS
- TF32 dense
- —
- FP16 dense
- 112 TFLOPS
- FP16 sparseRequires structured sparsity; rarely reached in practice
- —
- FP8 dense
- —
- INT8 dense
- —
Silicon
- Codename
- GV100
- Process node
- TSMC 12nm FFN
- Transistors
- 21.1B
- Die size
- 815 mm²
- CUDA cores
- 5,120
- Tensor cores
- 640
- SM count
- 80
- Boost clock
- 1,380 MHz
Platform & power
- TDP
- 250 W
- TFLOPS per wattDense FP16 ÷ TDP — drives your power bill at scale
- 0.45
- Form factor
- PCIe dual-slot
- Cooling
- passive
- Interconnect
- NVLink bridge
- Interconnect speed
- 300 GB/s
- PCIe generation
- Gen 3
- MIG
- No
- Launch
- 2018-03
- MSRP at launch
- —
Tesla V100 PCIe 32GB — common questions
Answered from the specifications above. Questions this card's data cannot answer with confidence are left out rather than guessed.
- How much VRAM does the Tesla V100 PCIe 32GB have?
- The Tesla V100 PCIe 32GB has 32 GB of HBM2 memory with 900 GB/s of bandwidth. In practice plan on roughly 25 GB for weights: KV cache, activations and fragmentation take the rest, and long contexts or large batches take considerably more.
- What size model can the Tesla V100 PCIe 32GB run?
- On a single Tesla V100 PCIe 32GB, about 12B parameters at FP16. Anything larger has to be split across several GPUs. These are weight-only figures — a long context window can consume tens of gigabytes on its own.
- Can you put the Tesla V100 PCIe 32GB in a desktop PC?
- Only with added cooling. The Tesla V100 PCIe 32GB is a passively cooled card — it has no fans at all and depends on a rack chassis forcing air through its heatsink. Dropped into an ordinary desktop it will throttle within minutes and can overheat outright. It needs either a rack server with high-static-pressure fans or an aftermarket shroud and blower.
- How much power does the Tesla V100 PCIe 32GB use?
- The Tesla V100 PCIe 32GB is rated at 250 W board power. It is fed through 1 × 8-pin EPS (CPU-style — not a PCIe 8-pin). For a single-card system, a 650 W supply or better leaves headroom for the CPU, drives and the transient spikes that trip undersized units.
- How much does it cost to run the Tesla V100 PCIe 32GB?
- At the US commercial average of $0.12/kWh, running one Tesla V100 PCIe 32GB continuously costs about $368 a year in electricity. That figure includes cooling and power-supply overhead at a factor of 1.4 over board power.
- Does the Tesla V100 PCIe 32GB support FP8?
- No. The Tesla V100 PCIe 32GB has no hardware FP8 path. FP8-quantised checkpoints will still load, but the values are upcast before they reach the tensor cores, so you get the memory saving without the speedup.
- Does the Tesla V100 PCIe 32GB support FP4?
- No. The Tesla V100 PCIe 32GB predates hardware FP4. You can run 4-bit quantised weights on it and the memory saving is real, but they are dequantised before compute, so there is no arithmetic speedup.
- Does the Tesla V100 PCIe 32GB support INT8?
- Yes. The Tesla V100 PCIe 32GB supports INT8 acceleration. INT8 remains the most widely supported quantised inference format across frameworks.
- Does the Tesla V100 PCIe 32GB support BF16?
- No. The Tesla V100 PCIe 32GB predates bfloat16 support and can only use FP16 for half-precision work. FP16 training on it usually needs loss scaling to stay numerically stable.
- Does the Tesla V100 PCIe 32GB have NVLink?
- Yes. The Tesla V100 PCIe 32GB supports NVLink (NVLink bridge) at 300 GB/s, so GPU-to-GPU traffic bypasses PCIe. This matters for tensor-parallel serving, where every token crosses the link.
- Does the Tesla V100 PCIe 32GB have RT cores?
- No. The Tesla V100 PCIe 32GB is a compute-only part with no ray-tracing hardware. For rendering or virtual workstations, look at the L40S or an RTX workstation card instead.
- Does the Tesla V100 PCIe 32GB have ECC memory?
- Yes. The Tesla V100 PCIe 32GB has ECC, so single-bit memory errors are corrected rather than silently corrupting a long run.
- Is the Tesla V100 PCIe 32GB worth buying?
- Buy it for: Used FP64 work and older CUDA codebases. Volta introduced tensor cores, so unlike the Pascal Teslas it still runs mixed-precision training meaningfully faster than FP32. Look elsewhere if: You need FP8, BF16 throughput, or a memory footprint above 32 GB. It also predates the Transformer Engine, so current LLM training recipes lose most of their advantage on it.