Radeon RX 640 OEM vs Tesla T4
Contents
Primary details
GPU architecture, market segment, value for money and other general parameters compared.
| Place in the ranking | 264 | not rated |
| Place by popularity | not in top-100 | not in top-100 |
| Power efficiency | 26.27 | no data |
| Architecture | Turing (2018−2022) | GCN 4.0 (2016−2020) |
| GPU code name | TU104 | Polaris 23 |
| Market segment | Workstation | Desktop |
| Release date | 13 September 2018 (7 years ago) | 9 April 2020 (5 years ago) |
Detailed specifications
General parameters such as number of shaders, GPU core base clock and boost clock speeds, manufacturing process, texturing and calculation speed. Note that power consumption of some graphics cards can well exceed their nominal TDP, especially when overclocked.
| Pipelines / CUDA cores | 2560 | 640 |
| Core clock speed | 585 MHz | 1295 MHz |
| Boost clock speed | 1590 MHz | no data |
| Number of transistors | 13,600 million | 2,200 million |
| Manufacturing process technology | 12 nm | 14 nm |
| Power consumption (TDP) | 70 Watt | 50 Watt |
| Texture fill rate | 254.4 | 51.80 |
| Floating-point processing power | 8.141 TFLOPS | 1.658 TFLOPS |
| ROPs | 64 | 16 |
| TMUs | 160 | 40 |
| Tensor Cores | 320 | no data |
| Ray Tracing Cores | 40 | no data |
| L1 Cache | 2.5 MB | 160 KB |
| L2 Cache | 4 MB | 512 KB |
Form factor & compatibility
Information on compatibility with other computer components. Useful when choosing a future computer configuration or upgrading an existing one. For desktop graphics cards it's interface and bus (motherboard compatibility), additional power connectors (power supply compatibility).
| Interface | PCIe 3.0 x16 | PCIe 3.0 x8 |
| Length | 168 mm | no data |
| Width | 1-slot | 1-slot |
| Supplementary power connectors | None | None |
VRAM capacity and type
Parameters of VRAM installed: its type, size, bus, clock and resulting bandwidth. Integrated GPUs have no dedicated video RAM and use a shared part of system RAM.
| Memory type | GDDR6 | GDDR5 |
| Maximum RAM amount | 16 GB | 4 GB |
| Memory bus width | 256 Bit | 128 Bit |
| Memory clock speed | 1250 MHz | 1500 MHz |
| Memory bandwidth | 320.0 GB/s | 96 GB/s |
Connectivity and outputs
This section shows the types and number of video connectors on each GPU. The data applies specifically to desktop reference models (for example, NVIDIA’s Founders Edition). OEM partners often modify both the number and types of ports. On notebook GPUs, video‐output options are determined by the laptop’s design rather than the graphics chip itself.
| Display Connectors | No outputs | 1x DisplayPort 1.4a, 2x mini-DisplayPort 1.4a |
API and SDK support
List of supported 3D and general-purpose computing APIs, including their specific versions.
| DirectX | 12 Ultimate (12_1) | 12 (12_0) |
| Shader Model | 6.5 | 6.7 (6.4) |
| OpenGL | 4.6 | 4.6 |
| OpenCL | 1.2 | 2.1 |
| Vulkan | 1.2.131 | 1.3 |
| CUDA | 7.5 | - |
| DLSS | + | - |
Pros & cons summary
| Recency | 13 September 2018 | 9 April 2020 |
| Maximum RAM amount | 16 GB | 4 GB |
| Chip lithography | 12 nm | 14 nm |
| Power consumption (TDP) | 70 Watt | 50 Watt |
Tesla T4 has a 300% higher maximum VRAM amount, and a 16.7% more advanced lithography process.
RX 640 OEM, on the other hand, has an age advantage of 1 year, and 40% lower power consumption.
We couldn't decide between Tesla T4 and Radeon RX 640 OEM. We've got no test results to judge.
Be aware that Tesla T4 is a workstation graphics card while Radeon RX 640 OEM is a desktop one.
Other comparisons
We selected several comparisons of graphics cards with performance close to those reviewed, providing you with more options to consider.
