AMD Instinct MI300X vs NVIDIA RTX PRO 6000D Blackwell Max-Q Comparison

AMD
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
nm
PROCESS 5 nm
LAUNCH DATE 2023
VS
NVIDIA
GEFORCE

RTX PRO 6000D Blackwell Max-Q

CORE STATE GB202
VRAM 96 GB
CLOCK SPEED 2288 MHz
TDP 300 W
BUS WIDTH 512 bit
ARCHITECTURE Blackwell 2.0
nm
PROCESS 5 nm
LAUNCH DATE 2025

PERFORMANCE BENCHMARKS

geekbench_opencl
317,994
N/A
3dmark_3dmark_steel_nomad_dx12
N/A
11,088

Analysis: AMD Instinct MI300X vs NVIDIA RTX PRO 6000D Blackwell Max-Q

Head-to-Head Benchmarks

The recorded data shows a fundamental divergence in how these two accelerators are measured. The AMD Instinct MI300X is evaluated through the Geekbench OpenCL workload, where it posts a score of 317994. That result places the MI300X at the 100th percentile among all GPUs in the database, meaning no other recorded device scores higher. The NVIDIA RTX PRO 6000D Blackwell Max-Q is instead measured using the 3DMark Steel Nomad DX12 test, where it records a score of 11088, landing at the 50th percentile. Because the two products use different benchmark suites, no direct head-to-head comparison can be drawn from those raw numbers alone. The database records zero direct head-to-head benchmark entries between them, and the win tally is zero for each.

Looking at the MI300X against its nearest rivals clarifies its standing. The AMD part sits 5% behind the NVIDIA H200 NVL, which averages 334891. It is also 8% behind the NVIDIA B200, which averages 345482. Conversely, the MI300X leads the NVIDIA L40S by 7.5%, with that rival averaging 295763, and it beats the NVIDIA RTX 6000 Ada Generation by 10.7%, where that rival averages 287237. These deltas indicate that the MI300X occupies a competitive position in the high-end compute segment, trailing only the most recent NVIDIA data center parts while clearly outpacing the prior generation workstation offerings.

The RTX PRO 6000D Blackwell Max-Q tells a different story through its nearest rivals. It matches the NVIDIA RTX PRO 6000 Blackwell Max-Q exactly, with both scoring 11088, a 0% delta. The AMD Radeon RX 550 is effectively tied at 11075, a 0.1% gap, and the NVIDIA GeForce GTX 1650 SUPER trails by 0.4% with a score of 11047. The AMD FirePro W4300 posts 11225, putting it 1.2% ahead of the RTX PRO 6000D. These results suggest that in the DX12 Steel Nomad workload, the RTX PRO 6000D performs in line with entry-level consumer and older professional cards, which is a surprising outcome for a part with such a high launch MSRP. The percentile rank of 50 confirms this middling position within the broader database.

The MI300X, by contrast, delivers a benchmark score that places it at the absolute top of the recorded distribution. Its nearest rivals are all substantial data center accelerators, and the deltas show a tight cluster of performance among the top tier. The RTX PRO 6000D, on the other hand, finds itself grouped with much older and less powerful hardware in its specific workload, which points to either a workload mismatch or a very different optimization target.

FAQ

Q: How does the AMD Instinct MI300X compare to its closest rival, the NVIDIA H200 NVL?

A: The MI300X records a Geekbench OpenCL score of 317994, which is 5% lower than the H200 NVL's average score of 334891.

Q: What benchmark score does the NVIDIA RTX PRO 6000D Blackwell Max-Q achieve, and what does its percentile mean?

A: It scores 11088 in 3DMark Steel Nomad DX12, placing it at the 50th percentile among all GPUs in the database, meaning half of recorded devices perform better and half perform worse.

Q: Are the two benchmark scores directly comparable?

A: No. The MI300X is measured with Geekbench OpenCL, while the RTX PRO 6000D is measured with 3DMark Steel Nomad DX12. The database records no head-to-head benchmark entries between the two products, so direct score comparison is not possible.

Q: Which GPUs are nearest rivals to the RTX PRO 6000D in the database?

A: The nearest rivals are the NVIDIA RTX PRO 6000 Blackwell Max-Q with the same score (11088), the AMD Radeon RX 550 at 11075, the NVIDIA GeForce GTX 1650 SUPER at 11047, and the AMD FirePro W4300 at 11225.

Q: How does the MI300X's score compare to the NVIDIA B200?

A: The MI300X trails the B200 by 8%, as the B200 averages 345482 compared to the MI300X's 317994.

Q: What does the RTX PRO 6000D's 0% delta with the RTX PRO 6000 Blackwell Max-Q indicate?

A: It indicates that the two NVIDIA parts produce identical scores in the 3DMark Steel Nomad DX12 workload, suggesting equivalent measured performance in that specific test.

Architecture Differences

The two accelerators share a 5 nm process node and both use TSMC as the foundry, but nearly every other architectural detail diverges. The AMD Instinct MI300X is built on the CDNA 3.0 architecture with the Aqua Vanjaram chip, while the NVIDIA RTX PRO 6000D Blackwell Max-Q uses the Blackwell 2.0 architecture with the GB202 chip. The MI300X belongs to the Instinct (MIx) generation, whereas the RTX PRO 6000D is part of the Blackwell PRO W (x000) generation.

Transistor counts differ substantially. The MI300X packs 153,000 million transistors on a 1017 mm² die, yielding a transistor density of 150.4 million per square millimeter. The RTX PRO 6000D contains 92,200 million transistors on a 750 mm² die, with a density of 122.9 million per square millimeter. The MI300X therefore has roughly 66% more transistors and a larger die area, while the RTX PRO 6000D achieves a smaller physical footprint.

Memory architecture presents another stark contrast. The MI300X uses 192 GB of HBM3 across an 8192-bit bus, delivering 5.32 TB/s of bandwidth. The RTX PRO 6000D uses 96 GB of GDDR7 on a 512-bit bus, providing 1.79 TB/s. The MI300X offers double the capacity and nearly three times the bandwidth. Memory clocks also differ: the MI300X runs at 1300 MHz with 5.2 Gbps effective, while the RTX PRO 6000D runs at 1750 MHz with 28 Gbps effective.

Compute resources are arranged differently. The MI300X has 19456 shading units and 1216 texture mapping units, but no ROPs, no RT cores, and no tensor cores recorded. The RTX PRO 6000D has 24064 shading units, 752 TMUs, and 192 ROPs, plus 188 RT cores and 752 tensor cores. The MI300X achieves a texture rate of 2,553.6 GTexel/s, while the RTX PRO 6000D records 1,720.6 GTexel/s. Pixel rates diverge sharply: the MI300X has no pixel output capability (0 MPixel/s), whereas the RTX PRO 6000D reaches 439.3 GPixel/s.

Floating-point performance favors the NVIDIA part. The RTX PRO 6000D delivers 110.1 TFLOPS for both FP32 and FP16, while the MI300X delivers 81.72 TFLOPS for both. Clock speeds also differ, with the MI300X running at 1000 MHz base and 2100 MHz boost, while the RTX PRO 6000D runs at 1590 MHz base and 2288 MHz boost. Power consumption shows a major gap: the MI300X has a TDP of 750 W with a suggested PSU of 1150 W, while the RTX PRO 6000D has a TDP of 300 W and a suggested PSU of 700 W.

Form factors and connectivity differ as well. The MI300X is an OAM module with no power connectors and no display outputs, while the RTX PRO 6000D is a dual-slot card with a single 16-pin power connector and four DisplayPort 2.1b outputs. Both use PCIe 5.0 x16 interfaces. API support distinguishes them: the RTX PRO 6000D supports DirectX 12 Ultimate (12_2), OpenGL 4.6, and Vulkan 1.4, while the MI300X records no API support at all. The RTX PRO 6000D has physical dimensions of 267 mm length, 111 mm height, and 40 mm width.

The Verdict

The data indicates that the AMD Instinct MI300X targets the top tier of data center compute, as evidenced by its 100th percentile ranking and a Geekbench OpenCL score of 317994 that places it ahead of the NVIDIA RTX 6000 Ada Generation by 10.7% and the L40S by 7.5%. Its nearest rivals are exclusively high-end data center accelerators like the H200 NVL and B200, which edge it out by 5% and 8% respectively. The MI300X offers 192 GB of HBM3 memory with 5.32 TB/s bandwidth, 153,000 million transistors, and a 750 W TDP, making it a heavyweight compute solution without display output or graphics APIs.

The NVIDIA RTX PRO 6000D Blackwell Max-Q, by contrast, records a 50th percentile score in its DX12 workload and sits alongside the Radeon RX 550 and GTX 1650 SUPER in its nearest rival group. Its 11088 Steel Nomad score is essentially tied with those older consumer parts, which suggests that its strengths lie outside the measured workload. The card does deliver 110.1 TFLOPS of FP32 performance, 96 GB of GDDR7 memory, 188 RT cores, and 752 tensor cores, all within a 300 W TDP and a dual-slot form factor with four DisplayPort outputs. It supports DirectX 12 Ultimate, OpenGL 4.6, and Vulkan 1.4, positioning it as a workstation graphics solution rather than a pure compute accelerator.

The choice between the two depends on workload requirements. The MI300X is the stronger option for memory-bandwidth-intensive compute tasks, given its 5.32 TB/s bandwidth and 192 GB capacity, and its benchmark performance places it among the fastest accelerators recorded. The RTX PRO 6000D is the option for graphics-oriented work, with its display outputs, RT cores, tensor cores, and full API support, though its measured DX12 performance does not distinguish it from far less expensive hardware. The MI300X demands a 1150 W suggested PSU and an OAM slot, while the RTX PRO 6000D fits a standard dual-slot PCIe layout with a 700 W PSU recommendation. The production status of the MI300X is not recorded, while the RTX PRO 6000D is listed as active.

Specification Differences

| Specification | AMD Instinct MI300X | NVIDIA RTX PRO 6000D Blackwell Max-Q |

|---|---|---|

| Architecture | CDNA 3.0 | Blackwell 2.0 |

| Process Node | 5 nm | 5 nm |

| Transistors | 153,000 million | 92,200 million |

| Die Size | 1017 mm² | 750 mm² |

| Base Clock | 1000 MHz | 1590 MHz |

| Boost Clock | 2100 MHz | 2288 MHz |

| Memory Size | 192 GB | 96 GB |

| Memory Type | HBM3 | GDDR7 |

| Memory Bus | 8192 bit | 512 bit |

| Memory Bandwidth | 5.32 TB/s | 1.79 TB/s |

| Shading Units | 19456 | 24064 |

| TMUs | 1216 | 752 |

| ROPs | 0 | 192 |

| RT Cores | N/A | 188 |

| Tensor Cores | N/A | 752 |

| FP32 Performance | 81.72 TFLOPS | 110.1 TFLOPS |

| FP16 Performance | 81.72 TFLOPS (1:1) | 110.1 TFLOPS (1:1) |

| TDP | 750 W | 300 W |

| Slot Width | OAM Module | Dual-slot |

| Power Connectors | None | 1x 16-pin |

| Suggested PSU | 1150 W | 700 W |

| Display Outputs | No outputs | 4x DisplayPort 2.1b |

| DirectX Support | N/A | 12 Ultimate (12_2) |

| OpenGL Support | N/A | 4.6 |

| Vulkan Support | N/A | 1.4 |

| Release Date | 2023-12-05 | 2025-03-17 |

| Launch MSRP | N/A | 8,565 USD |

The specification table highlights the divergent design philosophies. The MI300X emphasizes raw memory capacity and bandwidth, while the RTX PRO 6000D emphasizes clock speeds, shading units, and graphics features. The MI300X uses a larger die with more transistors but lower clocks, while the RTX PRO 6000D uses a smaller die with higher clocks and better power efficiency. The RTX PRO 6000D also carries the only recorded production status (Active) and the only launch MSRP (8,565 USD, stated once here as recorded).

DETAILED SPECIFICATIONS

SPECIFICATION
Instinct MI300X
RTX PRO 6000D Blackwell Max-Q
Core Specs
Shading Units
19,456
24,064 +23.7%
Shaders
19,456
24,064 +23.7%
TMUs
1,216
752 -38.2%
ROPs
0
192 +∞%
Compute Units
304
—
SM Count
—
188
Clocks
Base Clock
1000 MHz
1590 MHz
Boost Clock
2100 MHz
2288 MHz
Memory Clock
1300 MHz 5.2 Gbps effective
1750 MHz 28 Gbps effective
Memory
Memory Size
192 GB
96 GB
VRAM (MB)
196,608
98,304 -50.0%
Memory Type
HBM3
GDDR7
Memory Bus
8192 bit
512 bit
Bandwidth
5.32 TB/s
1.79 TB/s
Cache
L1 Cache
16 KB (per CU)
128 KB (per SM)
L2 Cache
16 MB
128 MB
L3 Cache
256 MB
—
Performance
Pixel Rate
0 MPixel/s
439.3 GPixel/s
Texture Rate
2,553.6 GTexel/s
1,720.6 GTexel/s
FP32 (TFLOPS)
81.72 TFLOPS
110.1 TFLOPS
FP64 (TFLOPS)
40.86 TFLOPS (1:2)
1.721 TFLOPS (1:64)
FP16 (TFLOPS)
81.72 TFLOPS (1:1)
110.1 TFLOPS (1:1)
AI/RT
RT Cores
—
188
Tensor Cores
—
752
Matrix Cores
1,216
—
Power
TDP
750 W
300 W
TDP (W)
750
300 -60.0%
Suggested PSU
1150 W
700 W
Power Connectors
None
1x 16-pin
Architecture
Architecture
CDNA 3.0
Blackwell 2.0
GPU Name
Aqua Vanjaram
GB202
Generation
Instinct (MIx)
Blackwell PRO W (x000)
Process Size
5 nm
5 nm
Transistors
153,000 million
92,200 million
Die Size
1017 mm²
750 mm²
Foundry
TSMC
TSMC
Density
150.4M / mm²
122.9M / mm²
AMD MCM
MCM
2
—
API Support
DirectX
—
12 Ultimate (12_2)
OpenGL
—
4.6
Vulkan
—
1.4
OpenCL
3.0
3.0
CUDA
—
12.0
Shader Model
—
6.9
Physical
Slot Width
OAM Module
Dual-slot
Length
—
267 mm 10.5 inches
Height
—
111 mm 4.4 inches
Outputs
No outputs
4x DisplayPort 2.1b
Bus Interface
PCIe 5.0 x16
PCIe 5.0 x16
Other
Launch Price
—
8,565 USD
Production
—
Active
Predecessor
Radeon Instinct
Workstation Ada
View Instinct MI300X Details View RTX PRO 6000D Blackwell Max-Q Details