Graphcore Bow IPU vs NVIDIA L40S 48GB

The L40S delivers 3% more FP16 throughput than the Bow IPU (362 vs 350 TFLOPS dense).

Bow IPU: Colossus MK2, 2022 L40S: Ada Lovelace, 2023

FP16 dense lead

+3%

L40S: 362 vs 350 TFLOPS

See rental pricing from $0.53/hr Download comparison sheet Printable version →

Graphcore Bow IPU

Full specs →

What is the Graphcore Bow IPU?

The Bow IPU is an ASIC from Graphcore on the Colossus MK2 architecture, an accelerator built for AI workloads. Bow IPU full specifications →

Can I rent the Bow IPU in the cloud?

Flopper is not tracking any live rental listings for the Bow IPU yet. The provider directory lists the clouds that rent comparable accelerators on demand. GPU cloud providers →

How does the Bow IPU compare with other accelerators?

This page sets the Bow IPU against the L40S. The Bow IPU is the older generation of the two. This is a cross-vendor matchup, Graphcore against NVIDIA. Flopper also compares it head to head with the GC200, across every published precision, memory and power. Bow IPU vs GC200 →

Where do the Bow IPU figures come from?

Every Bow IPU figure on this page traces to Graphcore's own datasheet or specification table. How Flopper verifies specifications →

NVIDIA L40S 48GB

Leads on FP16 Full specs →

What is the NVIDIA L40S?

The L40S is a GPU from NVIDIA on the Ada Lovelace architecture, an accelerator for AI training, inference and high-performance computing. It is a PCIe add-in card, so it fits standard servers and workstations. L40S full specifications →

Can I rent the L40S in the cloud?

Yes. Cloud providers tracked by Flopper rent the L40S by the hour, and the pricing page compares their current rates and the price history side by side. L40S rental prices →

How does the L40S compare with other accelerators?

This page sets the L40S against the Bow IPU. The L40S is the newer generation of the two. This is a cross-vendor matchup, NVIDIA against Graphcore. Flopper also compares it head to head with the TPU v5e 16GB and ATOM-Max, across every published precision, memory and power. L40S vs TPU v5e 16GB →

Where do the L40S figures come from?

Every L40S figure on this page traces to NVIDIA's own datasheet or specification table. How Flopper verifies specifications →

Specifications

Bow IPU

L40S

Architecture

Colossus MK2Ada Lovelace

Launch Year

20222023

Form Factor

—PCIe

Memory

—48 GB

Memory Bandwidth

—864 GB/s

TDP

—350 W

Process Node

—4nm

Spec

Bow IPU

L40S

Architecture Colossus MK2Ada Lovelace

Launch Year 20222023

Form Factor —PCIe

Memory —48 GB

Memory Bandwidth —864 GB/s

TDP —350 W

Process Node —4nm

Performance (TFLOPS)

Bow IPU

L40S

1.4 TFLOPS

87 TFLOPS 92 TFLOPS

183 TFLOPS

366 TFLOPS sparse

362 TFLOPS

733 TFLOPS sparse

350 TFLOPS 362 TFLOPS

733 TFLOPS sparse

733 TFLOPS

1,466 TFLOPS sparse

FP6

733 TOPS

1,466 TOPS sparse

Precision

Bow IPU

L40S

1.4 TFLOPS

87 TFLOPS 92 TFLOPS

183 TFLOPS

366 TFLOPS sparse

362 TFLOPS

733 TFLOPS sparse

350 TFLOPS 362 TFLOPS

733 TFLOPS sparse

733 TFLOPS

1,466 TFLOPS sparse

FP6

733 TOPS

1,466 TOPS sparse

Figures are dense. Where a vendor publishes a figure it sits under the dense one and never decides a winner. Hover any precision for its definition, or read the GPU glossary.

FLOPS by Precision

What actually differs

The L40S is the newer part: Ada Lovelace (2023) against the Bow IPU's Colossus MK2 (2022).

Weight the FP8 and FP16 rows for LLM work and the FP64 row for HPC.

New to the terms? The glossary explains FP64, FP32, TF32, BF16, FP16, FP8, INT8, Sparsity and TFLOPS. For how the generations relate, see NVIDIA architectures explained.

Frequently asked questions

Is the Bow IPU faster than the L40S?

At FP16 precision the L40S reaches 362 TFLOPS dense against 350 TFLOPS for the Bow IPU. The performance table on this page lists every published precision for both GPUs.

Which should you choose, the Bow IPU or the L40S?

Choose the L40S for training and high-throughput inference: it has 3% more FP16 throughput than the Bow IPU.

How much do the Bow IPU and the L40S cost to rent?

Live cloud listings tracked by Flopper price no live listings yet for the Bow IPU, and the L40S from $0.53 per GPU-hour across 25 providers. The rental section on this page links to every provider, rate and the price history for each GPU.

Where to Rent

Graphcore Bow IPU

No live rental listings tracked for the Bow IPU yet.

Browse GPU cloud providers

NVIDIA L40S 48GB

Rent from

$0.53 / GPU-hr

Lowest of 44 live listings across 25 providers · median $1.50

See all L40S rental prices →

Price history and every provider on the pricing page. Some links are affiliate links.

Get Comparison Updates

New GPUs added weekly. Be the first to see how they compare.

Open in Full Compare Tool → Bow IPU full specs L40S full specs Estimate cluster cost in the calculator →

Graphcore Bow IPU vs NVIDIA L40S 48GB

Graphcore Bow IPU

What is the Graphcore Bow IPU?

Can I rent the Bow IPU in the cloud?

How does the Bow IPU compare with other accelerators?

Where do the Bow IPU figures come from?

NVIDIA L40S 48GB

What is the NVIDIA L40S?

Can I rent the L40S in the cloud?

How does the L40S compare with other accelerators?

Where do the L40S figures come from?

Specifications

Performance (TFLOPS)

FLOPS by Precision

What actually differs

Frequently asked questions

Is the Bow IPU faster than the L40S?

Which should you choose, the Bow IPU or the L40S?

How much do the Bow IPU and the L40S cost to rent?

Where to Rent

Graphcore Bow IPU

NVIDIA L40S 48GB

Related comparisons

Get Comparison Updates

Stay Updated on GPU Specs