GPU Comparison
RADEON
AMD Radeon RX 640 OEM
CORE STATE
Polaris 23
VRAM
4 GB
CLOCK SPEED
—
TDP
50 W
BUS WIDTH
128 bit
ARCHITECTURE
GCN 4.0
PROCESS
14 nm
LAUNCH DATE
2020
VS
GEFORCE
A100 PCIe 40 GB
CORE STATE
GA100
VRAM
40 GB
CLOCK SPEED
1410 MHz
TDP
250 W
BUS WIDTH
5120 bit
ARCHITECTURE
Ampere
PROCESS
7 nm
LAUNCH DATE
2020
DETAILED SPECIFICATIONS
SPECIFICATION
RX 640 OEM
A100 PCIe 40 GB
Core Specs
Shading Units
640
6,912
+980.0%
Shaders
640
6,912
+980.0%
TMUs
40
432
+980.0%
ROPs
16
160
+900.0%
Compute Units
10
—
SM Count
—
108
Clocks
Base Clock
—
765 MHz
Boost Clock
—
1410 MHz
GPU Clock
1295 MHz
—
Memory Clock
1500 MHz
6 Gbps effective
1215 MHz
2.4 Gbps effective
Memory
Memory Size
4 GB
40 GB
VRAM (MB)
4,096
40,960
+900.0%
Memory Type
GDDR5
HBM2e
Memory Bus
128 bit
5120 bit
Bandwidth
96.00 GB/s
1.56 TB/s
Cache
L1 Cache
16 KB (per CU)
192 KB (per SM)
L2 Cache
512 KB
40 MB
Performance
Pixel Rate
20.72 GPixel/s
225.6 GPixel/s
Texture Rate
51.80 GTexel/s
609.1 GTexel/s
FP32 (TFLOPS)
1.658 TFLOPS
19.49 TFLOPS
FP64 (TFLOPS)
103.6 GFLOPS (1:16)
9.746 TFLOPS (1:2)
FP16 (TFLOPS)
1.658 TFLOPS (1:1)
77.97 TFLOPS (4:1)
AI/RT
Tensor Cores
—
432
BF16
—
311.84 TFLOPS (16:1)
TF32
—
155.92 TFLOPs (8:1)
Power
TDP
50 W
250 W
TDP (W)
50
250
+400.0%
Suggested PSU
250 W
600 W
Power Connectors
None
8-pin EPS
Architecture
Architecture
GCN 4.0
Ampere
GPU Name
Polaris 23
GA100
Generation
Polaris
(RX 600)
Server Ampere
(Axx)
Process Size
14 nm
7 nm
Transistors
2,200 million
54,200 million
Die Size
103 mm²
826 mm²
Foundry
GlobalFoundries
TSMC
Density
21.4M / mm²
65.6M / mm²
API Support
DirectX
12 (12_0)
—
OpenGL
4.6
—
Vulkan
1.3
—
OpenCL
2.1
3.0
CUDA
—
8.0
Shader Model
6.7
—
Physical
Slot Width
Single-slot
Dual-slot
Length
—
267 mm
10.5 inches
Height
—
111 mm
4.4 inches
Outputs
1x DisplayPort 1.4a2x mini-DisplayPort 1.4a
No outputs
Bus Interface
PCIe 3.0 x8
PCIe 4.0 x16
Other
Production
End-of-life
End-of-life
Predecessor
Polaris
Tesla Turing
Successor
Vega
Server Ada