Tesla C1060 Specifications
First Tesla with FP64, at one eighth the FP32 rate. Opened the door to HPC codes that could not tolerate single precision.
Buy it for
Historical interest. Its generation introduced usable double precision to Tesla, which is what first pulled HPC groups onto GPUs.
Look elsewhere if
You need to run anything current. Compute capability 1.3 is far below the floor of any supported CUDA toolkit.
Will it fit your system?
- Form factor
- PCIe dual-slot
- Cooling
- active
- Power connectors
- 1 × 6-pin + 1 × 8-pin PCIe
- Board power
- 187 W
- Suggested PSU
- 550 W or better
- Interface
- PCIe Gen 2
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 C1060 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
Not supported
Not supported on this generation
Precision must be managed manually
Multi-GPU traffic goes over PCIe — a real limit for distributed training
Compute-only die — no ray tracing hardware
No ECC, so silent corruption is possible on multi-day runs
What fits in memory
Largest model that fits on one Tesla C1060, assuming ~20% of the 4 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.
Cost to run
Electricity for one card at 100% duty, including cooling and PSU overhead (2,293 kWh/year).
Where it sits in the line
A newer part existing does not make this one a bad buy — it makes it cheaper. What matters is price per unit of the throughput you actually need.
Availability & market
Effectively worthless for compute; surfaces occasionally in surplus lots.
Discontinued around 2012. Expect secondary-market stock only, with no manufacturer warranty and limited driver support ahead.
Memory
- Capacity
- 4 GB
- Type
- GDDR3
- Bandwidth
- 102 GB/s
- Bus width
- 512-bit
Compute
- FP64
- 0.1 TFLOPS
- FP32
- 0.9 TFLOPS
- TF32 dense
- —
- FP16 dense
- —
- FP16 sparseRequires structured sparsity; rarely reached in practice
- —
- FP8 dense
- —
- INT8 dense
- —
Silicon
- Codename
- GT200
- Process node
- TSMC 65nm
- Transistors
- 1.4B
- Die size
- 576 mm²
- CUDA cores
- 240
- Tensor cores
- —
- SM count
- —
- Boost clock
- 1,300 MHz
Platform & power
- TDP
- 187 W
- Form factor
- PCIe dual-slot
- Cooling
- active
- Interconnect
- None
- Interconnect speed
- —
- PCIe generation
- Gen 2
- MIG
- No
- Launch
- 2008-11
- MSRP at launch
- —
Tesla C1060 — 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 C1060 have?
- The Tesla C1060 has 4 GB of GDDR3 memory with 102 GB/s of bandwidth. In practice plan on roughly 3 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 C1060 run?
- On a single Tesla C1060, about 1B parameters at FP16. Anything larger has to be split across several GPUs, which is slow on this card because it has no NVLink. These are weight-only figures — a long context window can consume tens of gigabytes on its own.
- Can you put the Tesla C1060 in a desktop PC?
- Yes. The Tesla C1060 is a self-cooled PCIe dual-slot card and works in a normal desktop, provided the case has clearance and the power supply is rated around 550 W or better.
- How much power does the Tesla C1060 use?
- The Tesla C1060 is rated at 187 W board power. It is fed through 1 × 6-pin + 1 × 8-pin PCIe. For a single-card system, a 550 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 C1060?
- At the US commercial average of $0.12/kWh, running one Tesla C1060 continuously costs about $275 a year in electricity. That figure includes cooling and power-supply overhead at a factor of 1.4 over board power.
- Does the Tesla C1060 support FP8?
- No. The Tesla C1060 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 C1060 support FP4?
- No. The Tesla C1060 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 C1060 support INT8?
- No. The Tesla C1060 has no INT8 tensor path, so quantised inference gives you a smaller memory footprint but no arithmetic speedup.
- Does the Tesla C1060 support BF16?
- No. The Tesla C1060 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 C1060 have NVLink?
- No. The Tesla C1060 has no NVLink. Multi-GPU work goes over PCIe Gen 2, which is an order of magnitude slower and makes tensor parallelism across cards expensive.
- Does the Tesla C1060 have RT cores?
- No. The Tesla C1060 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 C1060 have ECC memory?
- No. The Tesla C1060 has no ECC. For short interactive work this rarely matters; for multi-week training runs it is a real risk, and it is one reason datacentre parts cost more than consumer ones.
- What replaced the Tesla C1060?
- The Tesla M2090 (2011) is its successor, with 6 GB of GDDR5 at 177 GB/s.. On the used market that gap is what sets a fair price for the Tesla C1060.
- Is the Tesla C1060 worth buying?
- Buy it for: Historical interest. Its generation introduced usable double precision to Tesla, which is what first pulled HPC groups onto GPUs. Look elsewhere if: You need to run anything current. Compute capability 1.3 is far below the floor of any supported CUDA toolkit.