NVIDIA GRID A100B vs NVIDIA Tesla K40m
Graphics Processor
GPU model | GA100 | GK110B |
---|---|---|
Architecture | Ampere | Kepler |
Foundry | TSMC | TSMC |
Manufacturing process | 7 nm | 28 nm |
Number of transistors | 54,200 million | 7,080 million |
Die size | 826 mm² | 561 mm² |
Graphics Card
Release date | May 14th, 2020 | Nov 22nd, 2013 |
---|---|---|
Generation | GRID | Tesla |
Production | Active | End-of-life |
Interface | PCIe 4.0 x16 | PCIe 3.0 x16 |
Launch price | — | 7,699 USD |
Clocks
Base clock | 900 MHz | 745 MHz |
---|---|---|
Boost clock | 1005 MHz | 876 MHz |
Memory clock | 1215 MHz 2.4 Gbps effective | 1502 MHz 6 Gbps effective |
Memory Configuration
Memory size | 48 GB | 12 GB |
---|---|---|
Memory type | HBM2E | GDDR5 |
Memory bus width | 6144 bit | 384 bit |
Bandwidth | 1,866 GB/s | 288.4 GB/s |
Render Configuration
Shading units | 6912 | 2880 |
---|---|---|
TMUs | 432 | 240 |
ROPs | 192 | 48 |
SM count | 108 | — |
Tensor cores | 432 | — |
Cache L1 | 192 KB (per SM) | 16 KB (per SMX) |
Cache L2 | 48 MB | 1536 KB |
SMX count | — | 15 |
Performance
Pixel rate | 193.0 GPixel/s | 52.56 GPixel/s |
---|---|---|
Texture rate | 434.2 GTexel/s | 210.2 GTexel/s |
FP16 (half) performance | 55.57 TFLOPS (4:1) | — |
FP32 (float) performance | 13.89 TFLOPS | 5.046 TFLOPS |
FP64 (double) performance | 6.947 TFLOPS (1:2) | 1.682 TFLOPS (1:3) |
Dimensions & Outputs
Slot width | IGP | Dual-slot |
---|---|---|
TDP | 400 W | 245 W |
Suggested PSU | 800 W | 550 W |
Outputs | No outputs | No outputs |
Power connectors | None | — |
Length | — | 267 mm 10.5 inches |
API Support & Features
DirectX | — | 12 (11_1) |
---|---|---|
OpenGL | — | 4.6 |
OpenCL | 3.0 | 3.0 |
Vulkan | — | 1.1 |
CUDA | 8.0 | 3.5 |
Shader model | — | 5.1 |