NVIDIA B100 vs AMD FireStream 9170
We compared two Desktop platform GPUs: 96GB VRAM B100 and 2GB VRAM FireStream 9170 to see which GPU has better performance in key specifications, benchmark tests, power consumption, etc.
Main Differences
NVIDIA B100 's Advantages
Released 17 years late
Boost Clock1837MHz
More VRAM (96GB vs 2GB)
Larger VRAM bandwidth (4.10TB/s vs 51.33GB/s)
16576 additional rendering cores
AMD FireStream 9170 's Advantages
Lower TDP (105W vs 1000W)
Score
Benchmark
FP32 (float)
B100
+12390%
62.08 TFLOPS
FireStream 9170
0.497 TFLOPS
Graphics Card
Nov 2024
Release Date
Nov 2007
Server Blackwell
Generation
FireStream
Desktop
Type
Desktop
PCIe 5.0 x16
Bus Interface
PCIe 2.0 x16
Clock Speeds
1665 MHz
Base Clock
-
1837 MHz
Boost Clock
-
2000 MHz
Memory Clock
802 MHz
Memory
96GB
Memory Size
2GB
HBM3e
Memory Type
GDDR3
4096bit
Memory Bus
256bit
4.10TB/s
Bandwidth
51.33GB/s
Render Config
-
Compute Units
4
132
SM Count
-
16896
Shading Units
320
528
TMUs
16
24
ROPs
16
528
Tensor Cores
-
-
-
-
256 KB (per SM)
L1 Cache
-
50 MB
L2 Cache
256 KB
-
-
-
Theoretical Performance
44.09 GPixel/s
Pixel Rate
12.43 GPixel/s
969.9 GTexel/s
Texture Rate
12.43 GTexel/s
248.3 TFLOPS
FP16 (half)
-
62.08 TFLOPS
FP32 (float)
497.3 GFLOPS
31.04 TFLOPS
FP64 (double)
99.46 GFLOPS
Board Design
1000W
TDP
105W
1400 W
Suggested PSU
300 W
No outputs
Outputs
2x DVI
1x S-Video
-
Power Connectors
1x 6-pin
Graphics Processor
GB102
GPU Name
RV670
-
-
-
Blackwell
Architecture
TeraScale
TSMC
Foundry
TSMC
5 nm
Process Size
55 nm
104 billion
Transistors
0.666 billion
Unknown
Die Size
192 mm²
Graphics Features
N/A
DirectX
10.1 (10_1)
N/A
OpenGL
3.3 (full)
4.0 (partial)
3.0
OpenCL
N/A
N/A
Vulkan
N/A
10.1
CUDA
-
N/A
Shader Model
4.1
