GPU Comparison
RADEON
AMD Instinct MI300X
CORE STATE
Aqua Vanjaram
VRAM
192 GB
CLOCK SPEED
2100 MHz
TDP
750 W
BUS WIDTH
8192 bit
ARCHITECTURE
CDNA 3.0
PROCESS
5 nm
LAUNCH DATE
2023
VS
GEFORCE
B200
CORE STATE
GB100
VRAM
90 GB
CLOCK SPEED
1965 MHz
TDP
1000 W
BUS WIDTH
4096 bit
ARCHITECTURE
Blackwell
PROCESS
5 nm
LAUNCH DATE
—
PERFORMANCE BENCHMARKS
geekbench_opencl
DETAILED SPECIFICATIONS
SPECIFICATION
Instinct MI300X
B200
Core Specs
Shading Units
19,456
18,944
-2.6%
Shaders
19,456
18,944
-2.6%
TMUs
1,216
592
-51.3%
ROPs
0
24
+∞%
Compute Units
304
—
SM Count
—
148
Clocks
Base Clock
1000 MHz
700 MHz
Boost Clock
2100 MHz
1965 MHz
Memory Clock
1300 MHz 5.2 Gbps effective
2000 MHz
8 Gbps effective
Memory
Memory Size
192 GB
90 GB
VRAM (MB)
196,608
92,160
-53.1%
Memory Type
HBM3
HBM3e
Memory Bus
8192 bit
4096 bit
Bandwidth
5.32 TB/s
4.10 TB/s
Cache
L1 Cache
16 KB (per CU)
256 KB (per SM)
L2 Cache
16 MB
50 MB
L3 Cache
256 MB
—
Performance
Pixel Rate
0 MPixel/s
47.16 GPixel/s
Texture Rate
2,553.6 GTexel/s
1,163.3 GTexel/s
FP32 (TFLOPS)
81.72 TFLOPS
74.45 TFLOPS
FP64 (TFLOPS)
40.86 TFLOPS (1:2)
37.22 TFLOPS (1:2)
FP16 (TFLOPS)
81.72 TFLOPS (1:1)
1,191.2 TFLOPS (16:1)
AI/RT
Tensor Cores
—
592
Matrix Cores
1,216
—
Power
TDP
750 W
1000 W
TDP (W)
750
1,000
+33.3%
Suggested PSU
1150 W
1400 W
Power Connectors
None
—
Architecture
Architecture
CDNA 3.0
Blackwell
GPU Name
Aqua Vanjaram
GB100
Generation
Instinct (MIx)
Server Blackwell
(Bxx)
Process Size
5 nm
5 nm
Transistors
153,000 million
104,000 million
Die Size
1017 mm²
—
Foundry
TSMC
TSMC
Density
150.4M / mm²
—
AMD MCM
MCM
2
—
API Support
OpenCL
3.0
3.0
CUDA
—
10.0
Physical
Slot Width
OAM Module
SXM Module
Outputs
No outputs
No outputs
Bus Interface
PCIe 5.0 x16
PCIe 5.0 x16
Other
Production
—
Active
Predecessor
Radeon Instinct
Server Hopper
Successor
—
Server Rubin