News
NVIDIA Announces New RTX A5000 and RTX A4000 GPUs for Desktop and A30 and A10 GPUs for Data Center
April 13, 2021
7 min read
Earlier this week during GTC 2021, amidst the action packed keynote NVIDIA also announced the arrival of several new GPUs to their product line. Among the releases, this blog will focus on the NVIDIA® RTX™ A5000, NVIDIA® RTX™ A4000, NVIDIA A30, and NVIDIA A10.
For desktop GPUs, to further bolster their professional graphics line the NVIDIA RTX A5000 and NVIDIA RTX A4000 GPUs were released.
Next-Generation RTX Technology
All the NVIDIA RTX GPUs feature the latest technologies in the NVIDIA Ampere architecture:
- Second-Generation RT Cores: Up to 2x the throughput of the previous generation, with the ability to run concurrent ray tracing, shading and denoising tasks.
- Third-generation Tensor Cores: Up to 2x the throughput of the previous generation, up to 10x with sparsity, with support for new TF32 and BFloat16 data formats.
- CUDA Cores: Up to 2.5x the FP32 throughput of the previous generation for significant increases in graphics and compute workloads.
The new NVIDIA RTX A5000
The NVIDIA® RTX™ A5000 combines the latest-generation RT Cores, Tensor Cores, and CUDA cores with 24GB of graphics memory to deliver a robust desktop experience for designers, engineers, and artists to create the future, today.
The new NVIDIA RTX A4000
The NVIDIA RTX A4000 is the most powerful single-slot GPU for professionals that combines the latest-generation RT Cores, Tensor Cores, and CUDA cores with 16 GB of graphics memory in a compact form factor that fits a wide range of systems.
Looking for a new NVIDIA Powered Workstation or Server?
Shop NVIDIA Systems
NVIDIA RTX A5000 & NVIDIA RTX A4000 Specs Compared
| NVIDIA RTX A5000 | NVIDIA RTX A4000 | |
| Architecture | NVIDIA Ampere Architecture | NVIDIA Ampere Architecture |
| Foundry | Samsung | Samsung |
| Process Size | 8nm | 8nm |
| Transistors | 28.3 billion | 17.4billion |
| Die Size | 628.4 mm2 | 392.5 mm2 |
| CUDA Parallel Processing cores | 8,192 | 6,144 |
| NVIDIA Tensor Cores | 256 | 192 |
| NVIDIA RT Cores | 64 | 48 |
| Single-Precision Performance1 | 27.8 TFLOPS | 19.2 TFLOPS |
| RT Core Performance1 | 54.2 TFLOPS | 37.4 TFLOPS |
| Tensor Performance1 | 222.2 TFLOPS | 153.4 TFLOPS |
| GPU Memory | 24 GB GDDR6 with ECC | 16 GB GDDR6 with ECC |
| Memory Interface | 384-bit | 256-bit |
| Memory Bandwidth | 768 GB/s | 448 GB/s |
| Max Power Consumption | 230W | 140W |
| Graphics Bus | PCI Express 4.0 x16 | PCI Express 4.0 x16 |
| Display Connectors | DP 1.4 (4) | DP 1.4 (4) |
| Form Factor | 4.4” H x 10.5” L Dual Slot | 4.4” H x 9.5” L Single Slot |
| Product Weight | 1.025 kg | 500 g |
| Thermal Solution | Active | Active |
| vGPU Software Support4 | NVIDIA ® Virtual PC/Virtual Applications (vPC/vApps), NVIDIA RTX® Virtual Workstation (vWS), NVIDIA Virtual Compute Server (vCS)3 | n/a |
| vGPU Profiles Supported | See vGPU Pricing & Licensing Guide | n/a |
| NVIDIA® 3D Vision® and 3D Vision Pro | Support via 3 pin mini DIN | Support via 3 pin mini DIN |
| Frame lock | Compatible (with Quadro Sync II) | Compatible (with Quadro Sync II) |
| NVLink | 2-way low profile (2-slot and 3-slot bridges) | n/a |
| NVLink Interconnect | 112.5 GB/s (bidirectional) | n/a |
| Power Connector | 1x 8-pin PCIe | 1x 6-pin PCIe |
| NVENC | NVDEC | 1x | 2x (+AV1 decode) | 1x | 1x (+AV1 decode) |
NVIDIA Adds New Data Center GPUs
For the data center, and server workloads we’ll examine the new NVIDIA A30 GPU and NVIDIA10 GPU. The NVIDIA A30 Tensor Core GPU is being hailed as the most versatile mainstream compute GPU for AI inference and mainstream enterprise workloads.
NVIDIA A30 Tensor Core GPU
The A30 features NVIDIA Ampere architecture Tensor Core technology, and supports a broad range of math precisions, providing a single accelerator to speed up every workload.
NVIDIA A10 Tensor Core GPU
the A10 combines second generation RT Cores, third-generation Tensor Cores, and new streaming microprocessors with 24 gigabytes (GB) of GDDR6 memory—all in a 150W power envelope—for versatile graphics, rendering, AI, and compute performance. From virtual workstations, accessible anywhere in the world, to render nodes to the data centers running a variety of workloads, A10 is built to deliver optimal performance in a single-wide, full-height, full-length PCIe form factor.
| NVIDIA A10 | NVIDIA A30 | |
| FP64 | n/a | 5.2 TF |
| FP64 Tensor Core | n/a | 10.3 TF |
| FP32 | 31.2 TF | 82 TF | 165 TF* |
| TF32 Tensor Core | 62.5 TF | 125 TF* | 165 TF | 330 TF* |
| BFLOAT16 Tensor Core | 125 TF | 250 TF* | 165 TF | 330 TF* |
| FP16 Tensor Core | 125 TF | 250 TF* | 330 TOPS | 661 TOPS* |
| INT8 Tensor Core | 250 TOPS | 500 TOPS* | 661 TOPS | 1321 TOPS* |
| INT4 Tensor Core | 500 TOPS | 1000 TOPS* | 661 TOPS | 1321 TOPS* |
| RT Cores | 72 | n/a |
| Encode / Decode | 1 encoder, 1 decoder, (+AV1 decode) | n/a |
| Media engines | n/a | 1 optical flow accelerator (OFA), 1 JPEG decoder (NVJPEG) 4 Video decoders (NVDEC) |
| GPU Memory | 24 GB GDDR6 | 24GB HBM2 |
| GPU Memory Bandwidth | 600 GB/s | 933GB/s |
| Interconnect | PCIe Gen4: 64 GB/s | PCIe Gen4: 64GB/s Third-gen NVIDIA® NVLINK® 200GB/s** |
| Form Factor | 1-slot FHFL | 2-slot FHFL |
| Max TDP Power | 150W | 165W |
| vGPU Software Support | NVIDIA vPC/vApps, NVIDIA RTX™ vWS, NVIDIA Virtual Compute Server (vCS) | NVIDIA AI Enterprise for Vmware, NVIDIA Virtual Compute Server |
| Multi-Instance GPU (MIG) | n/a | 4 MIGs @ 6GB each, 2 MIGs @ 12GB each, 1 MIGs @ 24GB |
| Secure and Measured Boot with Hardware Root of Trust | Yes | Yes |
| NEBS Ready | Level 3 | Level 3 |
| Power Connector | PEX 8-pin | PEX 8-pin |
Have any questions?
Contact Exxact Today
News
NVIDIA Announces New RTX A5000 and RTX A4000 GPUs for Desktop and A30 and A10 GPUs for Data Center
April 13, 20217 min read
Earlier this week during GTC 2021, amidst the action packed keynote NVIDIA also announced the arrival of several new GPUs to their product line. Among the releases, this blog will focus on the NVIDIA® RTX™ A5000, NVIDIA® RTX™ A4000, NVIDIA A30, and NVIDIA A10.
For desktop GPUs, to further bolster their professional graphics line the NVIDIA RTX A5000 and NVIDIA RTX A4000 GPUs were released.
Next-Generation RTX Technology
All the NVIDIA RTX GPUs feature the latest technologies in the NVIDIA Ampere architecture:
- Second-Generation RT Cores: Up to 2x the throughput of the previous generation, with the ability to run concurrent ray tracing, shading and denoising tasks.
- Third-generation Tensor Cores: Up to 2x the throughput of the previous generation, up to 10x with sparsity, with support for new TF32 and BFloat16 data formats.
- CUDA Cores: Up to 2.5x the FP32 throughput of the previous generation for significant increases in graphics and compute workloads.
The new NVIDIA RTX A5000
The NVIDIA® RTX™ A5000 combines the latest-generation RT Cores, Tensor Cores, and CUDA cores with 24GB of graphics memory to deliver a robust desktop experience for designers, engineers, and artists to create the future, today.
The new NVIDIA RTX A4000
The NVIDIA RTX A4000 is the most powerful single-slot GPU for professionals that combines the latest-generation RT Cores, Tensor Cores, and CUDA cores with 16 GB of graphics memory in a compact form factor that fits a wide range of systems.
Looking for a new NVIDIA Powered Workstation or Server?
Shop NVIDIA Systems
NVIDIA RTX A5000 & NVIDIA RTX A4000 Specs Compared
| NVIDIA RTX A5000 | NVIDIA RTX A4000 | |
| Architecture | NVIDIA Ampere Architecture | NVIDIA Ampere Architecture |
| Foundry | Samsung | Samsung |
| Process Size | 8nm | 8nm |
| Transistors | 28.3 billion | 17.4billion |
| Die Size | 628.4 mm2 | 392.5 mm2 |
| CUDA Parallel Processing cores | 8,192 | 6,144 |
| NVIDIA Tensor Cores | 256 | 192 |
| NVIDIA RT Cores | 64 | 48 |
| Single-Precision Performance1 | 27.8 TFLOPS | 19.2 TFLOPS |
| RT Core Performance1 | 54.2 TFLOPS | 37.4 TFLOPS |
| Tensor Performance1 | 222.2 TFLOPS | 153.4 TFLOPS |
| GPU Memory | 24 GB GDDR6 with ECC | 16 GB GDDR6 with ECC |
| Memory Interface | 384-bit | 256-bit |
| Memory Bandwidth | 768 GB/s | 448 GB/s |
| Max Power Consumption | 230W | 140W |
| Graphics Bus | PCI Express 4.0 x16 | PCI Express 4.0 x16 |
| Display Connectors | DP 1.4 (4) | DP 1.4 (4) |
| Form Factor | 4.4” H x 10.5” L Dual Slot | 4.4” H x 9.5” L Single Slot |
| Product Weight | 1.025 kg | 500 g |
| Thermal Solution | Active | Active |
| vGPU Software Support4 | NVIDIA ® Virtual PC/Virtual Applications (vPC/vApps), NVIDIA RTX® Virtual Workstation (vWS), NVIDIA Virtual Compute Server (vCS)3 | n/a |
| vGPU Profiles Supported | See vGPU Pricing & Licensing Guide | n/a |
| NVIDIA® 3D Vision® and 3D Vision Pro | Support via 3 pin mini DIN | Support via 3 pin mini DIN |
| Frame lock | Compatible (with Quadro Sync II) | Compatible (with Quadro Sync II) |
| NVLink | 2-way low profile (2-slot and 3-slot bridges) | n/a |
| NVLink Interconnect | 112.5 GB/s (bidirectional) | n/a |
| Power Connector | 1x 8-pin PCIe | 1x 6-pin PCIe |
| NVENC | NVDEC | 1x | 2x (+AV1 decode) | 1x | 1x (+AV1 decode) |
NVIDIA Adds New Data Center GPUs
For the data center, and server workloads we’ll examine the new NVIDIA A30 GPU and NVIDIA10 GPU. The NVIDIA A30 Tensor Core GPU is being hailed as the most versatile mainstream compute GPU for AI inference and mainstream enterprise workloads.
NVIDIA A30 Tensor Core GPU
The A30 features NVIDIA Ampere architecture Tensor Core technology, and supports a broad range of math precisions, providing a single accelerator to speed up every workload.
NVIDIA A10 Tensor Core GPU
the A10 combines second generation RT Cores, third-generation Tensor Cores, and new streaming microprocessors with 24 gigabytes (GB) of GDDR6 memory—all in a 150W power envelope—for versatile graphics, rendering, AI, and compute performance. From virtual workstations, accessible anywhere in the world, to render nodes to the data centers running a variety of workloads, A10 is built to deliver optimal performance in a single-wide, full-height, full-length PCIe form factor.
| NVIDIA A10 | NVIDIA A30 | |
| FP64 | n/a | 5.2 TF |
| FP64 Tensor Core | n/a | 10.3 TF |
| FP32 | 31.2 TF | 82 TF | 165 TF* |
| TF32 Tensor Core | 62.5 TF | 125 TF* | 165 TF | 330 TF* |
| BFLOAT16 Tensor Core | 125 TF | 250 TF* | 165 TF | 330 TF* |
| FP16 Tensor Core | 125 TF | 250 TF* | 330 TOPS | 661 TOPS* |
| INT8 Tensor Core | 250 TOPS | 500 TOPS* | 661 TOPS | 1321 TOPS* |
| INT4 Tensor Core | 500 TOPS | 1000 TOPS* | 661 TOPS | 1321 TOPS* |
| RT Cores | 72 | n/a |
| Encode / Decode | 1 encoder, 1 decoder, (+AV1 decode) | n/a |
| Media engines | n/a | 1 optical flow accelerator (OFA), 1 JPEG decoder (NVJPEG) 4 Video decoders (NVDEC) |
| GPU Memory | 24 GB GDDR6 | 24GB HBM2 |
| GPU Memory Bandwidth | 600 GB/s | 933GB/s |
| Interconnect | PCIe Gen4: 64 GB/s | PCIe Gen4: 64GB/s Third-gen NVIDIA® NVLINK® 200GB/s** |
| Form Factor | 1-slot FHFL | 2-slot FHFL |
| Max TDP Power | 150W | 165W |
| vGPU Software Support | NVIDIA vPC/vApps, NVIDIA RTX™ vWS, NVIDIA Virtual Compute Server (vCS) | NVIDIA AI Enterprise for Vmware, NVIDIA Virtual Compute Server |
| Multi-Instance GPU (MIG) | n/a | 4 MIGs @ 6GB each, 2 MIGs @ 12GB each, 1 MIGs @ 24GB |
| Secure and Measured Boot with Hardware Root of Trust | Yes | Yes |
| NEBS Ready | Level 3 | Level 3 |
| Power Connector | PEX 8-pin | PEX 8-pin |
Have any questions?
Contact Exxact Today
