AMD Radeon Instinct MI60 vs AMD Radeon Pro W6800X Duo Comparison

AMD
RADEON

AMD Radeon Instinct MI60

CORE STATE Vega 20
VRAM 32 GB
CLOCK SPEED 1800 MHz
TDP 300 W
BUS WIDTH 4096 bit
ARCHITECTURE GCN 5.1
nm
PROCESS 7 nm
LAUNCH DATE 2018
VS
AMD
RADEON

Radeon Pro W6800X Duo

CORE STATE Navi 21
VRAM 32 GB
CLOCK SPEED 1967 MHz
TDP 400 W
BUS WIDTH 256 bit
ARCHITECTURE RDNA 2.0
nm
PROCESS 7 nm
LAUNCH DATE 2021

PERFORMANCE BENCHMARKS

geekbench_opencl
92,488
124,335
geekbench_vulkan
92,444
125,622
geekbench_metal
N/A
157,365

Analysis: AMD Radeon Instinct MI60 vs AMD Radeon Pro W6800X Duo

AMD Radeon Pro W6800X Duo and AMD Radeon Instinct MI60 are both end-of-life AMD workstation accelerators aimed at different corners of the professional market. The W6800X Duo, built on RDNA 2.0, targets high-end Mac Pro graphics and compute, while the MI60, based on GCN 5.1, is a data-center focused compute card from an earlier era. Benchmark data shows a clear performance gap, but the architectural and interface differences matter just as much for real-world selection.

FAQ

Q: Which card has the higher average benchmark score?

A: The AMD Radeon Pro W6800X Duo scores 135774 on average, placing it in the 96th percentile of all GPUs. The AMD Radeon Instinct MI60 averages 92466, sitting in the 93rd percentile.

Q: How big is the performance lead in the shared benchmarks?

A: In Geekbench OpenCL, the W6800X Duo scores 124335 versus 92488 for the MI60, a lead of 34.4%. In Geekbench Vulkan, the W6800X Duo hits 125622 against 92444, which is 35.9% ahead.

Q: Do both cards have the same memory capacity?

A: Yes, both feature 32 GB of memory. However, the W6800X Duo uses GDDR6 on a 256-bit bus providing 512.0 GB/s bandwidth, while the MI60 uses HBM2 on a 4096-bit bus delivering 1.02 TB/s.

Q: What are the physical power requirements?

A: The W6800X Duo has a 400 W TDP and suggests an 800 W power supply. The MI60 has a 300 W TDP and suggests a 700 W power supply.

Q: Which card supports real-time ray tracing hardware?

A: Only the AMD Radeon Pro W6800X Duo includes 60 dedicated ray tracing cores. The AMD Radeon Instinct MI60 has no ray tracing cores listed.

Q: What is the main interface difference?

A: The W6800X Duo uses the Apple MPX bus interface, while the MI60 uses a standard PCIe 4.0 x16 connection. This makes the MI60 compatible with standard PC motherboards, whereas the W6800X Duo is tied to Apple's ecosystem.

Architecture Differences

The two cards represent fundamentally different design philosophies from AMD. The W6800X Duo is built on the Navi 21 chip using the RDNA 2.0 architecture, fabricated on TSMC's 7 nm process. It packs 26,800 million transistors into a 520 mm² die, achieving a transistor density of 51.5M per mm². The MI60 uses the older Vega 20 chip with GCN 5.1 architecture, also on TSMC 7 nm, but with just 13,230 million transistors on a 331 mm² die, for a density of 40.0M per mm².

Core counts differ significantly. The W6800X Duo has 3840 shading units, 240 texture mapping units, and 96 ROPs. The MI60 counters with 4096 shading units, 256 TMUs, but only 64 ROPs. Notably, the W6800X Duo includes 60 ray tracing cores, a feature entirely absent from the MI60. Neither card lists tensor cores.

Clock speeds favor the newer card. The W6800X Duo runs at a base clock of 1800 MHz and boosts to 1967 MHz, with memory at 2000 MHz (16 Gbps effective). The MI60 has a lower base of 1200 MHz but boosts to 1800 MHz, with memory at 1000 MHz (2 Gbps effective). The MI60's memory clock is far lower, but its 4096-bit bus compensates with massive bandwidth.

The W6800X Duo supports DirectX 12 Ultimate (12_2) and Vulkan 1.4, while the MI60 is limited to DirectX 12 (12_1) and Vulkan 1.3. Both support OpenGL 4.6. The W6800X Duo also features a much richer display output configuration with 1x HDMI 2.1 and 4x Thunderbolt, versus the MI60's single mini-DisplayPort 1.4a, reflecting their different intended roles.

The Verdict

The data is unambiguous: the AMD Radeon Pro W6800X Duo is the faster card by a wide margin. It wins both head-to-head benchmarks decisively, with leads of 34.4% in OpenCL and 35.9% in Vulkan. Its average benchmark score of 135774 places it in the 96th percentile, well above the MI60's 92466 in the 93rd percentile.

However, the choice between these two depends entirely on the platform and workload. The W6800X Duo uses the Apple MPX interface, meaning it is designed exclusively for Mac Pro systems. It also requires a quad-slot form factor and 400 W TDP, suggesting a power supply of 800 W. The MI60, conversely, uses a standard PCIe 4.0 x16 slot with a dual-slot design and 300 W TDP, making it far easier to integrate into conventional servers or workstations.

For raw compute performance in a Mac Pro environment, the W6800X Duo is the obvious pick. Its 60 ray tracing cores and newer RDNA 2.0 architecture also make it suitable for graphics-accelerated workloads beyond pure compute. For a standard PC data-center deployment where PCIe compatibility matters more than peak speed, the MI60 remains a viable option, especially given its superior 1.02 TB/s memory bandwidth, which can benefit certain memory-bound compute tasks. The W6800X Duo had a launch MSRP of 4,999 USD; the MI60 had no listed launch MSRP.

Specification Differences

| Specification | AMD Radeon Pro W6800X Duo | AMD Radeon Instinct MI60 |

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

| Architecture | RDNA 2.0 | GCN 5.1 |

| Process Node | 7 nm | 7 nm |

| Transistors | 26,800 million | 13,230 million |

| Die Size | 520 mm² | 331 mm² |

| Base Clock | 1800 MHz | 1200 MHz |

| Boost Clock | 1967 MHz | 1800 MHz |

| Memory Type | GDDR6 | HBM2 |

| Memory Bus Width | 256 bit | 4096 bit |

| Memory Bandwidth | 512.0 GB/s | 1.02 TB/s |

| Shading Units | 3840 | 4096 |

| TMUs | 240 | 256 |

| ROPs | 96 | 64 |

| Ray Tracing Cores | 60 | None |

| Pixel Rate | 188.8 GPixel/s | 115.2 GPixel/s |

| Texture Rate | 472.1 GTexel/s | 460.8 GTexel/s |

| FP32 Performance | 15.11 TFLOPS | 14.75 TFLOPS |

| FP16 Performance | 30.21 TFLOPS (2:1) | 29.49 TFLOPS (2:1) |

| TDP | 400 W | 300 W |

| Slot Width | Quad-slot | Dual-slot |

| Power Connectors | Not listed | 1x 6-pin + 1x 8-pin |

| Suggested PSU | 800 W | 700 W |

| Bus Interface | Apple MPX | PCIe 4.0 x16 |

| Display Outputs | 1x HDMI 2.1, 4x Thunderbolt | 1x mini-DisplayPort 1.4a |

| DirectX Support | 12 Ultimate (12_2) | 12 (12_1) |

| Vulkan Support | 1.4 | 1.3 |

| Release Date | 2021-08-02 | 2018-11-17 |

Head-to-Head Benchmarks

The shared benchmark suite consists of two tests: Geekbench OpenCL and Geekbench Vulkan. The AMD Radeon Pro W6800X Duo dominates both.

In Geekbench OpenCL, the W6800X Duo scores 124335 against the MI60's 92488. That is a delta of 34.4%, meaning the newer card processes compute workloads roughly a third faster. This aligns with its higher clock speeds and newer architecture, despite having fewer shading units (3840 vs 4096). The MI60's higher shading unit count does not overcome the W6800X Duo's efficiency advantage.

The Vulkan test shows an even larger gap. The W6800X Duo scores 125622, while the MI60 manages 92444. This 35.9% difference suggests the RDNA 2.0 architecture handles modern graphics APIs better than the older GCN 5.1 design. The W6800X Duo's support for Vulkan 1.4, compared to the MI60's Vulkan 1.3, likely contributes to this result.

These benchmark results translate into a 2-0 win count for the W6800X Duo. The MI60 records zero wins in the head-to-head comparisons. The average benchmark score difference is also stark: 135774 versus 92466, a gap of roughly 47% in favor of the W6800X Duo.

Where Each One Wins

The AMD Radeon Pro W6800X Duo wins in every measured benchmark category. It is superior in OpenCL compute performance, Vulkan graphics workloads, and overall average score. Its 60 ray tracing cores provide hardware acceleration for ray-traced rendering, which the MI60 lacks entirely. The W6800X Duo also has a higher pixel rate (188.8 GPixel/s vs 115.2 GPixel/s) and texture rate (472.1 GTexel/s vs 460.8 GTexel/s), making it better suited for graphics-intensive tasks. Its 512.0 GB/s memory bandwidth is lower than the MI60's 1.02 TB/s, but the higher clocks and newer architecture compensate in practice.

The AMD Radeon Instinct MI60 has no benchmark wins, but it holds advantages in specific areas that matter for certain use cases. Its 4096-bit memory bus and 1.02 TB/s bandwidth are more than double the W6800X Duo's, which can be critical for large data sets that fit within the 32 GB HBM2 pool. It also has more shading units (4096) and TMUs (256) than the W6800X Duo, though these do not translate into performance wins in the available tests. The MI60's lower 300 W TDP and dual-slot form factor make it easier to deploy in dense server environments. Its standard PCIe 4.0 x16 interface is a significant practical advantage for anyone building a custom compute node outside Apple's ecosystem.

For a Mac Pro user needing maximum compute and graphics acceleration, the W6800X Duo is the clear choice. For a data-center operator with standard PCIe slots who prioritizes memory bandwidth and power efficiency over raw speed, the MI60 still offers a compelling profile, despite its older architecture. The benchmark data favors the W6800X Duo across the board, but the MI60's unique memory subsystem and interface compatibility give it a niche that the performance numbers alone do not capture.

DETAILED SPECIFICATIONS

SPECIFICATION
Instinct MI60
Pro W6800X Duo
Core Specs
Shading Units
4,096
3,840 -6.3%
Shaders
4,096
3,840 -6.3%
TMUs
256
240 -6.3%
ROPs
64
96 +50.0%
Compute Units
64
60 -6.3%
Clocks
Base Clock
1200 MHz
1800 MHz
Boost Clock
1800 MHz
1967 MHz
Memory Clock
1000 MHz 2 Gbps effective
2000 MHz 16 Gbps effective
Memory
Memory Size
32 GB
32 GB
VRAM (MB)
32,768
32,768 0.0%
Memory Type
HBM2
GDDR6
Memory Bus
4096 bit
256 bit
Bandwidth
1.02 TB/s
512.0 GB/s
Cache
L1 Cache
16 KB (per CU)
128 KB per Array
L2 Cache
4 MB
4 MB
L3 Cache
—
128 MB
L0 Cache
—
32 KB per WGP
Performance
Pixel Rate
115.2 GPixel/s
188.8 GPixel/s
Texture Rate
460.8 GTexel/s
472.1 GTexel/s
FP32 (TFLOPS)
14.75 TFLOPS
15.11 TFLOPS
FP64 (TFLOPS)
7.373 TFLOPS (1:2)
944.2 GFLOPS (1:16)
FP16 (TFLOPS)
29.49 TFLOPS (2:1)
30.21 TFLOPS (2:1)
AI/RT
RT Cores
—
60
Power
TDP
300 W
400 W
TDP (W)
300
400 +33.3%
Suggested PSU
700 W
800 W
Power Connectors
1x 6-pin + 1x 8-pin
—
Architecture
Architecture
GCN 5.1
RDNA 2.0
GPU Name
Vega 20
Navi 21
Generation
Radeon Instinct (MIx)
Radeon Pro Mac (Navi II Series)
Process Size
7 nm
7 nm
Transistors
13,230 million
26,800 million
Die Size
331 mm²
520 mm²
Foundry
TSMC
TSMC
Density
40.0M / mm²
51.5M / mm²
API Support
DirectX
12 (12_1)
12 Ultimate (12_2)
OpenGL
4.6
4.6
Vulkan
1.3
1.4
OpenCL
2.1
2.1
Shader Model
6.7
6.8
Physical
Slot Width
Dual-slot
Quad-slot
Length
267 mm 10.5 inches
267 mm 10.5 inches
Height
111 mm 4.4 inches
120 mm 4.7 inches
Outputs
1x mini-DisplayPort 1.4a
1x HDMI 2.14x Thunderbolt
Bus Interface
PCIe 4.0 x16
Apple MPX
Other
Launch Price
—
4,999 USD
Production
End-of-life
End-of-life
Predecessor
FirePro Data Center
—
View Radeon Instinct MI60 Details View Radeon Pro W6800X Duo Details