GPU Comparison
GEFORCE
NVIDIA A2
CORE STATE
GA107
VRAM
16 GB
CLOCK SPEED
1770 MHz
TDP
60 W
BUS WIDTH
128 bit
ARCHITECTURE
Ampere
PROCESS
8 nm
LAUNCH DATE
2021
VS
GEFORCE
Tesla K20Xm
CORE STATE
GK110
VRAM
6 GB
CLOCK SPEED
—
TDP
235 W
BUS WIDTH
384 bit
ARCHITECTURE
Kepler
PROCESS
28 nm
LAUNCH DATE
2012
PERFORMANCE BENCHMARKS
geekbench_opencl
geekbench_metal
DETAILED SPECIFICATIONS
SPECIFICATION
A2
Tesla K20Xm
Core Specs
Shading Units
1,280
2,688
+110.0%
Shaders
1,280
2,688
+110.0%
TMUs
40
224
+460.0%
ROPs
32
48
+50.0%
SM Count
10
—
Clocks
Base Clock
1440 MHz
—
Boost Clock
1770 MHz
—
GPU Clock
—
732 MHz
Memory Clock
1563 MHz
12.5 Gbps effective
1300 MHz
5.2 Gbps effective
Memory
Memory Size
16 GB
6 GB
VRAM (MB)
16,384
6,144
-62.5%
Memory Type
GDDR6
GDDR5
Memory Bus
128 bit
384 bit
Bandwidth
200.1 GB/s
249.6 GB/s
Cache
L1 Cache
128 KB (per SM)
16 KB (per SMX)
L2 Cache
2 MB
1536 KB
Performance
Pixel Rate
56.64 GPixel/s
40.99 GPixel/s
Texture Rate
70.80 GTexel/s
164.0 GTexel/s
FP32 (TFLOPS)
4.531 TFLOPS
3.935 TFLOPS
FP64 (TFLOPS)
70.80 GFLOPS (1:64)
1,311.7 GFLOPS (1:3)
FP16 (TFLOPS)
4.531 TFLOPS (1:1)
—
AI/RT
RT Cores
10
—
Tensor Cores
40
—
Power
TDP
60 W
235 W
TDP (W)
60
235
+291.7%
Suggested PSU
250 W
550 W
Power Connectors
None
—
Architecture
Architecture
Ampere
Kepler
GPU Name
GA107
GK110
Generation
Workstation Ampere
(Ax000)
Tesla Kepler
(Kxx)
Process Size
8 nm
28 nm
Transistors
8,700 million
7,080 million
Die Size
200 mm²
561 mm²
Foundry
Samsung
TSMC
Density
43.5M / mm²
12.6M / mm²
API Support
DirectX
12 Ultimate (12_2)
12 (11_0)
OpenGL
4.6
4.6
Vulkan
1.4
1.2.175
OpenCL
3.0
3.0
CUDA
8.6
3.5
Shader Model
6.8
6.5 (5.1)
Physical
Slot Width
Single-slot
Dual-slot
Length
—
267 mm
10.5 inches
Outputs
No outputs
No outputs
Bus Interface
PCIe 4.0 x8
PCIe 3.0 x16
Other
Launch Price
—
7,699 USD
Production
End-of-life
End-of-life
Predecessor
Quadro Turing
Tesla Fermi
Successor
Workstation Ada
Tesla Maxwell