AMD Radeon HD 8790M vs NVIDIA Quadro 4000M Comparison

Head-to-Head Benchmarks

The recorded data shows a single direct benchmark comparison between these two mobile workstation GPUs. In the Geekbench OpenCL test, the NVIDIA Quadro 4000M scores 5211, while the AMD Radeon HD 8790M scores 5017. This gives the NVIDIA part a 3.7% advantage, which is a modest but clear lead.

Looking at the broader context from the database, the AMD Radeon HD 8790M's average benchmark score of 5691 places it ahead of several familiar rivals. It sits 0.4% behind the Intel Iris Pro Graphics P6300 (5712) and 0.5% behind the NVIDIA GeForce GTX 670MX (5721). More notably, it is 1.6% ahead of the NVIDIA Quadro M500M (5604). The AMD part also achieves a Geekbench Vulkan score of 6365, a result that the Quadro 4000M cannot match because the NVIDIA part has no Vulkan score recorded in the database.

The NVIDIA Quadro 4000M, with its average score of 5211, is positioned 0.5% behind the GeForce GTX 760M (5236) and 1.4% behind the GeForce 940M (5284). However, it leads the AMD Radeon R7 M260X by 1% (5161) and the Quadro K3100M by 1.1% (5154). The Quadro 4000M also holds a percentile ranking of 30 among all GPUs, while the Radeon HD 8790M ranks slightly higher at the 33rd percentile.

The head-to-head result is clear: the Quadro 4000M wins the only shared benchmark. However, the AMD part's Vulkan score and its higher average benchmark total suggest that the Radeon HD 8790M may have strengths in other workloads, particularly those that leverage modern API features. The 3.7% OpenCL gap is significant but not overwhelming, and the two parts are closely matched in raw compute performance as measured by the database.

Architecture Differences

The two GPUs come from different architectural generations and manufacturing processes. The AMD Radeon HD 8790M is built on the GCN 1.0 architecture using the Mars chip, while the NVIDIA Quadro 4000M uses the Fermi architecture with the GF104 chip. This fundamental divergence shapes many of their characteristics.

The manufacturing process differs substantially. AMD uses a 28 nm process at TSMC, while NVIDIA uses a larger 40 nm process, also at TSMC. This translates into significant differences in transistor density and die size. The AMD chip packs 950 million transistors into a 77 mm² die, achieving a density of 12.3 million transistors per square millimeter. The NVIDIA chip, by contrast, has 1,950 million transistors spread across a much larger 332 mm² die, with a density of only 5.9 million per square millimeter. The Radeon HD 8790M is therefore more than twice as dense in terms of transistor packing.

Core configuration also differs. The AMD part features 384 shading units, 24 texture mapping units, and 8 raster output units. The NVIDIA part has fewer shading units at 336, but significantly more texture units at 56 and more raster output units at 32. These differences affect the balance of compute, texturing, and pixel throughput.

Clock speeds tell a partial story. The AMD Radeon HD 8790M has a base clock of 850 MHz and a boost clock of 900 MHz. The Quadro 4000M has no base or boost clocks recorded in the database. Memory clocks also differ: the AMD part runs its GDDR5 memory at 1000 MHz (4 Gbps effective), while the NVIDIA part runs at 625 MHz (2.5 Gbps effective).

The memory subsystems are configured differently. Both have 2 GB of GDDR5, but the bus widths diverge. The AMD part uses a 128-bit interface, yielding 64.00 GB/s of bandwidth. The NVIDIA part uses a 256-bit interface, delivering 80.00 GB/s. Despite the slower memory clock, the wider bus gives the Quadro 4000M a 25% bandwidth advantage.

Theoretical throughput rates reflect these architectural choices. The AMD Radeon HD 8790M achieves 7.200 GPixel/s pixel rate and 21.60 GTexel/s texture rate. The NVIDIA Quadro 4000M posts 6.650 GPixel/s and 26.60 GTexel/s. In FP32 compute, the AMD part delivers 691.2 GFLOPS versus 638.4 GFLOPS for the NVIDIA part, a 8.3% advantage for AMD.

Power consumption is recorded only for the NVIDIA part: the Quadro 4000M has a TDP of 100 W. No TDP is listed for the AMD Radeon HD 8790M. Both use MXM modules, but with different form factor interfaces: the AMD part uses MXM-A (3.0), while the NVIDIA part uses MXM-B (3.0). Neither requires external power connectors, and both have portable-device-dependent display outputs.

API support differs in notable ways. The AMD part supports DirectX 12 (11_1), OpenGL 4.6, and Vulkan 1.2.170. The NVIDIA part supports DirectX 12 (11_0) and OpenGL 4.6, but has no Vulkan support recorded. This explains the AMD part's Vulkan benchmark score of 6365.

FAQ

Q: Which GPU wins the direct benchmark comparison?

A: The NVIDIA Quadro 4000M wins the Geekbench OpenCL test with a score of 5211, beating the AMD Radeon HD 8790M's 5017 by 3.7%.

Q: Does the AMD Radeon HD 8790M support Vulkan?

A: Yes, the AMD part supports Vulkan 1.2.170 and has a recorded Geekbench Vulkan score of 6365. The NVIDIA Quadro 4000M has no Vulkan support listed in the database.

Q: How do the memory bandwidths compare?

A: The NVIDIA Quadro 4000M has 80.00 GB/s of bandwidth from its 256-bit bus, while the AMD Radeon HD 8790M has 64.00 GB/s from a 128-bit bus. The NVIDIA part leads by 25%.

Q: Which GPU has higher FP32 compute performance?

A: The AMD Radeon HD 8790M delivers 691.2 GFLOPS, which is 8.3% higher than the NVIDIA Quadro 4000M's 638.4 GFLOPS.

Q: What are the percentile rankings of these GPUs?

A: The AMD Radeon HD 8790M sits at the 33rd percentile among all GPUs, while the NVIDIA Quadro 4000M sits at the 30th percentile.

Q: How does each GPU compare to its nearest rivals?

A: The AMD part is 1.6% ahead of the Quadro M500M but 0.4% behind the Iris Pro P6300 and 0.5% behind the GTX 670MX. The NVIDIA part is 1% ahead of the R7 M260X and 1.1% ahead of the Quadro K3100M, but 0.5% behind the GTX 760M.

The Verdict

The data supports a split decision. In the only shared benchmark, the NVIDIA Quadro 4000M takes the win with a 3.7% higher OpenCL score. This is a meaningful margin for a single test, and it aligns with the NVIDIA part's higher memory bandwidth of 80.00 GB/s versus 64.00 GB/s. The wider 256-bit bus and larger 32 ROP count suggest the Quadro 4000M is built for scenarios where memory throughput and pixel output matter.

However, the AMD Radeon HD 8790M counters with several advantages. Its FP32 compute of 691.2 GFLOPS exceeds the NVIDIA part by 8.3%. Its Vulkan support, with a strong score of 6365, opens a workload category the Quadro 4000M cannot enter. The AMD part also achieves a higher average benchmark score of 5691 versus 5211, and it ranks at the 33rd percentile versus the 30th for NVIDIA.

The manufacturing differences favor AMD clearly. The 28 nm process versus 40 nm gives the Radeon HD 8790M a much smaller die (77 mm² versus 332 mm²) and higher transistor density (12.3M per mm² versus 5.9M per mm²). This suggests better power efficiency, though the database does not record a TDP for the AMD part to confirm this directly.

For users who prioritize raw compute with modern API support, the AMD Radeon HD 8790M is the better choice based on its FP32 lead and Vulkan capability. For users who need maximum memory bandwidth and higher texture throughput, the NVIDIA Quadro 4000M offers 25% more bandwidth and 23% higher texture rate. The choice depends on workload type.

Specification Differences

The following fields differ between the AMD Radeon HD 8790M and NVIDIA Quadro 4000M:

• Chip: Mars (AMD) versus GF104 (NVIDIA)
• Architecture: GCN 1.0 versus Fermi
• Process Node: 28 nm versus 40 nm
• Transistors: 950 million versus 1,950 million
• Die Size: 77 mm² versus 332 mm²
• Transistor Density: 12.3M / mm² versus 5.9M / mm²
• Base Clock: 850 MHz versus not recorded
• Boost Clock: 900 MHz versus not recorded
• Memory Clock: 1000 MHz (4 Gbps effective) versus 625 MHz (2.5 Gbps effective)
• Memory Bus Width: 128 bit versus 256 bit
• Memory Bandwidth: 64.00 GB/s versus 80.00 GB/s
• Shading Units: 384 versus 336
• Texture Mapping Units: 24 versus 56
• Raster Output Units: 8 versus 32
• Pixel Rate: 7.200 GPixel/s versus 6.650 GPixel/s
• Texture Rate: 21.60 GTexel/s versus 26.60 GTexel/s
• FP32: 691.2 GFLOPS versus 638.4 GFLOPS
• TDP: not recorded versus 100 W
• Bus Interface: MXM-A (3.0) versus MXM-B (3.0)
• DirectX Support: 12 (11_1) versus 12 (11_0)
• Vulkan Support: 1.2.170 versus not recorded
• Release Date: 2013-03-31 versus 2011-02-21
• Predecessor: London versus Quadro FX Mobile
• Successor: Gem System versus Quadro Kepler-M

Where Each One Wins

The AMD Radeon HD 8790M wins in compute-heavy scenarios. Its FP32 output of 691.2 GFLOPS is 8.3% higher than the NVIDIA part, making it the stronger choice for general-purpose compute workloads that are not bandwidth-bound. The Vulkan score of 6365 demonstrates that the AMD part handles modern graphics APIs effectively, while the Quadro 4000M has no Vulkan path at all. The AMD part also wins on manufacturing efficiency, with a 28 nm process yielding a much smaller die and higher transistor density, which typically translates to lower power draw per unit of performance.

The NVIDIA Quadro 4000M wins in memory-intensive and texture-heavy workloads. Its 80.00 GB/s bandwidth is 25% higher than the AMD part's 64.00 GB/s, and its texture rate of 26.60 GTexel/s beats the AMD part's 21.60 GTexel/s by 23%. The 56 texture units and 32 ROPs, versus 24 and 8 for AMD, give the Quadro 4000M a structural advantage in tasks that repeatedly sample textures or write to frame buffers. The OpenCL benchmark result of 5211 versus 5017 confirms that the NVIDIA part translates its memory advantage into a real-world performance lead in at least one compute workload.

In terms of average benchmark scores, the AMD part leads with 5691 versus 5211, a 9.2% difference. The AMD part also holds a higher percentile ranking (33 versus 30). These aggregate figures suggest that the AMD Radeon HD 8790M has broader general performance, while the Quadro 4000M excels in specific memory-bound tasks.

The release dates are worth noting: the AMD part launched in 2013-03-31, while the NVIDIA part launched in 2011-02-21. The two-year gap explains some of the architectural differences, with the AMD part benefiting from a newer process node and API support. Both are now end-of-life products.