AMD Radeon Pro V520 vs Nvidia A40

Two data center GPUs, one RDNA and one Ampere: 8 GB HBM2 against 48 GB GDDR6 and 36% more memory bandwidth on the A40.

At a glance

Updated 9 minutes ago

AMD Radeon Pro V520

Nvidia A40

Cheapest listed price per GPU per hour by billing type, and the median across providers.

Billing	Radeon Pro V520	A40	Difference
On-demand	$0.38	$1.05	2.8x more expensive on A40
Reserved	$0.19	$0.61	3.2x more expensive on A40
Median	$0.87	$1.05	17% cheaper on Radeon Pro V520

Median: Median across providers, one vote each, so a large catalog does not outweigh a small one.

Weekly median on-demand price per GPU per hour

Last 12 months

Aggregating historical prices...

Spec	Radeon Pro V520	A40	Difference
Memory
Capacity	8 GB HBM2	48 GB GDDR6	6x more on A40
Bandwidth	512 GB/s	696 GB/s	1.4x higher on A40
Bus width	2,048-bit	384-bit	5.3x wider on Radeon Pro V520
Compute · dense
FP16 / BF16	14.8 TFLOPS	149.7 TFLOPS	10x higher on A40
INT8		299.3 TOPS
FP32	7.4 TFLOPS	37.4 TFLOPS	5.1x higher on A40
Platform
Interconnect	PCIe only	PCIe only
Board power	225 W	300 W	1.3x higher on A40
Released	Q2 2020	Q4 2020
Precision
Supported formats	FP16FP32INT4INT8	FP16BF16TF32FP32INT4INT8

Compute: Vendor peak figures, dense and per GPU. Not measured throughput.

Providers listing these GPUs, those with both first, so you can switch without moving clouds.