AMD Radeon R7 240 vs NVIDIA GeForce GT 645M Comparison

The Verdict

The AMD Radeon R7 240 and NVIDIA GeForce GT 645M are both end-of-life mobile and desktop graphics solutions, but the benchmark data tells a clear story. The R7 240 holds the overall advantage, with an average benchmark score of 5063 compared to the GT 645M’s 4411. That is a 14.8% gap in favor of AMD, and it stems from a single decisive OpenCL result.

For users choosing between these two, the R7 240 is the stronger pick for compute workloads. Its Geekbench OpenCL score of 5063 crushes the GT 645M’s 2680, a massive 88.9% lead. The R7 240 also lands in the 30th percentile of all GPUs, while the GT 645M sits at the 26th percentile. If raw computational throughput is the priority, the AMD card wins without question.

The GT 645M, however, is not without merit. It supports additional APIs: Metal and Vulkan, with recorded scores of 5679 and 4875 respectively in those tests. The R7 240 has no recorded Metal or Vulkan benchmarks in the database. So, if the software environment relies on those APIs, the NVIDIA part offers measured capability where AMD has none.

Who should pick which? The R7 240 is for users running OpenCL-heavy applications, where its performance is nearly double that of the competition. The GT 645M is for users who need broader API coverage, particularly in Metal-based environments (common in Apple ecosystems) or Vulkan. Neither card is a modern powerhouse, but the data clearly favors AMD for sheer OpenCL muscle.

Architecture Differences

The two GPUs come from different architectural lineages. The AMD Radeon R7 240 is built on GCN 1.0, using the Oland chip, and belongs to the Volcanic Islands (R7 200) generation. The NVIDIA GeForce GT 645M uses the Kepler architecture, with the GK107 chip, from the GeForce 600M generation.

Both are manufactured on a 28 nm process at TSMC, but the silicon differs significantly. AMD packs 950 million transistors into a 77 mm² die, yielding a transistor density of 12.3 million per mm². NVIDIA’s GK107 is larger, with 1,270 million transistors on a 118 mm² die, but the density is lower at 10.8 million per mm². This means AMD achieves higher packing efficiency, while NVIDIA uses more die area for its design.

The core configurations diverge. The R7 240 has 320 shading units, 20 texture mapping units (TMUs), and 8 raster operation units (ROPs). The GT 645M has 384 shading units, 32 TMUs, and 16 ROPs. On paper, NVIDIA has more of everything. The texture rate reflects this: the GT 645M hits 24.96 GTexel/s, while the R7 240 manages 15.60 GTexel/s. The pixel rate is identical at 6.240 GPixel/s for both, since the ROP count and clock speeds align.

Clock speeds are nearly matched. The R7 240 runs at a base of 730 MHz with a boost of 780 MHz. The GT 645M has a base of 709 MHz and the same 780 MHz boost. Memory is also identical in specification: 2 GB of DDR3, a 128-bit bus, and 28.80 GB/s of bandwidth, with a memory clock of 900 MHz (1800 Mbps effective).

The floating-point performance tells a different story. The R7 240 delivers 499.2 GFLOPS of FP32 compute, while the GT 645M offers 599.0 GFLOPS. Here, NVIDIA is 20% ahead. But the benchmark results do not reflect this theoretical advantage in OpenCL, likely due to driver efficiency and workload characteristics.

Feature support differs. The R7 240 supports DirectX 12 (11_1), OpenGL 4.6, and Vulkan 1.2.170. The GT 645M supports DirectX 12 (11_0), OpenGL 4.6, and Vulkan 1.2.175. NVIDIA’s Vulkan version is slightly newer, and it also has the Metal API (via Geekbench Metal score), which AMD lacks in this comparison.

Physical design varies: the R7 240 is a single-slot, 168 mm long card with no power connectors and a 30 W TDP, suggesting a 200 W PSU. The GT 645M is an IGP (integrated graphics processor) for portable devices, with a 32 W TDP and no power connector. The R7 240 has explicit display outputs (DVI, HDMI 1.4a, VGA), while the GT 645M’s outputs are described as "Portable Device Dependent."

Where Each One Wins

The benchmark data shows a clear split. The R7 240 wins the only head-to-head test: Geekbench OpenCL. Its score of 5063 versus 2680 is a landslide, with a delta of 88.9%. This indicates that for general-purpose GPU compute, OpenCL workloads, the AMD card is far superior. The R7 240 also holds a higher average score and percentile, making it the better all-round performer in the database.

The GT 645M wins in API coverage. It has recorded scores for Geekbench Metal (5679) and Geekbench Vulkan (4875). These are not head-to-head comparisons because the R7 240 lacks such measurements, but they demonstrate that the NVIDIA card can execute Metal and Vulkan tasks. In any application that requires Metal (macOS environments) or Vulkan, the GT 645M is the only option with verified performance data.

For gaming, the picture is more nuanced. The GT 645M has higher texture rate (24.96 GTexel/s vs 15.60 GTexel/s) and more TMUs and ROPs, which typically helps with rasterization and fill-rate-bound scenarios. The R7 240 has higher FP32 compute per pixel, but that rarely translates to gaming wins. However, the OpenCL score suggests the AMD card handles general compute better, which can help in physics or post-processing effects.

Power efficiency is close: the R7 240 has a 30 W TDP, the GT 645M a 32 W TDP. Neither requires external power. The R7 240 is a desktop card with a 200 W PSU recommendation, while the GT 645M is for laptops. For portable use, the GT 645M is the only choice given its IGP form factor.

FAQ

Q: Which GPU has a higher average benchmark score?

A: The AMD Radeon R7 240, with an average score of 5063, versus the NVIDIA GeForce GT 645M’s 4411.

Q: How big is the OpenCL performance gap?

A: The R7 240 scores 5063 in Geekbench OpenCL, while the GT 645M scores 2680. That is an 88.9% lead for AMD.

Q: Does the GT 645M support any APIs that the R7 240 does not?

A: Yes. The GT 645M has recorded Metal and Vulkan scores (5679 and 4875 respectively). The R7 240 has no Metal or Vulkan benchmark data.

Q: Which card has more shading units and TMUs?

A: The GT 645M has 384 shading units and 32 TMUs, versus the R7 240’s 320 shading units and 20 TMUs.

Q: Are the memory specifications the same?

A: Yes, both have 2 GB of DDR3, a 128-bit bus, and 28.80 GB/s bandwidth.

Q: What are the TDPs of each card?

A: The R7 240 has a 30 W TDP, and the GT 645M has a 32 W TDP.

Head-to-Head Benchmarks

The database records only one direct head-to-head test between these two GPUs: Geekbench OpenCL. This is the defining measurement for comparison.

Geekbench OpenCL: AMD Radeon R7 240 wins, 5063 vs 2680 (88.9% delta).

This is a decisive victory. The R7 240 more than doubles the GT 645M’s score. In practical terms, any OpenCL-accelerated application, whether it is video encoding, scientific simulation, or data processing, will run nearly twice as fast on the AMD card. The GT 645M’s higher theoretical FP32 (599 GFLOPS vs 499.2 GFLOPS) does not save it; the R7 240’s architecture and drivers extract far more real-world compute performance.

The GT 645M has no wins in the head-to-head table. Its only counterpoints are its Metal and Vulkan scores, which are not part of the head-to-head comparison because the R7 240 lacks corresponding tests. Those scores (5679 Metal, 4875 Vulkan) are respectable, but they do not affect the direct matchup.

In summary, if you must choose between these two based on the recorded data, the AMD Radeon R7 240 is the clear winner in compute performance. The NVIDIA GeForce GT 645M is the fallback for API compatibility, but it trails significantly in the one test where both are measured.