FireStream 9250 vs H200 NVL
We compared two Desktop platform GPUs: 1024MB VRAM FireStream 9250 and 141GB VRAM H200 NVL to see which GPU has better performance in key specifications, benchmark tests, power consumption, etc.
Main Differences
FireStream 9250 's Advantages
Lower TDP (150W vs 600W)
H200 NVL 's Advantages
Released 16 years and 5 months late
Boost Clock1785MHz
More VRAM (141GB vs 1024GB)
Larger VRAM bandwidth (4.89TB/s vs 63.55GB/s)
16096 additional rendering cores
Score
Benchmark
FP32 (float)
FireStream 9250
1000
H200 NVL
+5932%
60320
Graphics Card
Jun 2008
Release Date
Nov 2024
FireStream
Generation
Server Hopper
Desktop
Type
Desktop
PCIe 2.0 x16
Bus Interface
PCIe 5.0 x16
Clock Speeds
-
Base Clock
1365 MHz
-
Boost Clock
1785 MHz
993 MHz
Memory Clock
1593 MHz
Memory
1024MB
Memory Size
141GB
GDDR3
Memory Type
HBM3e
256bit
Memory Bus
6144bit
63.55GB/s
Bandwidth
4.89TB/s
Render Config
10
Compute Units
-
-
SM Count
132
800
Shading Units
16896
40
TMUs
528
16
ROPs
24
-
Tensor Cores
528
16 KB (per CU)
L1 Cache
256 KB (per SM)
256 KB
L2 Cache
50 MB
Theoretical Performance
10.00 GPixel/s
Pixel Rate
42.84 GPixel/s
25.00 GTexel/s
Texture Rate
942.5 GTexel/s
-
FP16 (half)
241.3 TFLOPS
1000 GFLOPS
FP32 (float)
60.32 TFLOPS
200.0 GFLOPS
FP64 (double)
30.16 TFLOPS
Board Design
150W
TDP
600W
450 W
Suggested PSU
1000 W
1x DVI
Outputs
No outputs
1x 6-pin
Power Connectors
8-pin EPS
Graphics Processor
RV770
GPU Name
GH100
RV770 PRO
GPU Variant
-
TeraScale
Architecture
Hopper
TSMC
Foundry
TSMC
55 nm
Process Size
5 nm
0.956 billion
Transistors
80 billion
256 mm²
Die Size
814 mm²
Graphics Features
10.1 (10_1)
DirectX
N/A
3.3
OpenGL
N/A
1.1
OpenCL
3.0
N/A
Vulkan
N/A
-
CUDA
9.0
4.1
Shader Model
N/A