NVIDIA GeForce GTX 650 Ti Boost vs NVIDIA GeForce GTX 850M Comparison

Head-to-Head Benchmarks

The recorded data shows a clear split between these two NVIDIA parts across the two shared compute workloads. In Geekbench OpenCL, the GeForce GTX 850M posts a score of 9821 against the GTX 650 Ti Boost's 9302, a 5.6% advantage for the mobile Maxwell chip. That is a meaningful margin in a synthetic compute test, and it places the 850M slightly ahead of its own nearest rivals: it sits 0.1% above the GTX 465 and 0.3% above the GTX 960, while trailing the AMD Radeon R7 M380 by just 0.1%. The 850M's average benchmark score of 9302 puts it in the 46th percentile of all GPUs, which is a respectable position for a part designed around a 45 W power envelope.

The Vulkan workload flips the result decisively. Here the GTX 650 Ti Boost scores 10041, which is 12.5% higher than the 850M's 8782. That is a substantial gap, and it is importantly the 650 Ti Boost's Vulkan result is its strongest recorded benchmark. The 850M's Vulkan score is actually lower than its OpenCL score, suggesting the Maxwell architecture's compute strengths do not carry over to this particular API workload. The 650 Ti Boost also benefits from a wider memory bus and higher memory bandwidth in real-world scenarios, though the database only records these two shared tests for the head-to-head comparison.

The wins are split exactly one apiece: the 850M takes OpenCL, the 650 Ti Boost takes Vulkan. The average benchmark scores tell a more complex story, however, because the 650 Ti Boost's average of 8067 is dragged down by a Metal score of 4858, a test that the 850M did not run. That Metal result is far below the card's other scores, and it explains why the 650 Ti Boost's percentile sits at 42nd despite having a higher single-test showing in Vulkan. The 850M's nearest rival list is tighter around its average, with the GTX 650 Ti Boost's 9302 sitting 0.2% above the GRID K2 and 0.2% below the GTX 650 Ti, but the head-to-head data is what matters most for this comparison.

FAQ

Q: Which GPU wins in OpenCL compute performance?

A: The GTX 850M wins decisively in Geekbench OpenCL with a score of 9821, which is 5.6% higher than the GTX 650 Ti Boost's 9302. This is the 850M's strongest recorded result.

Q: Does the GTX 650 Ti Boost have any advantage in modern graphics APIs?

A: Yes, the Vulkan results favor the 650 Ti Boost by a wide margin. It scores 10041 versus the 850M's 8782, a 12.5% difference in favor of the desktop card.

Q: How do these cards compare to their direct rivals in the database?

A: The 850M sits within 0.3% of the GTX 465, GTX 960, and Radeon R7 M380, with a 0.9% lead over the Vega 8. The 650 Ti Boost is effectively tied with the GRID K2, GTX 650 Ti, and GTX 880M, all within 0.3%, and it leads the Quadro P5000 by 0.3%.

Q: Is the 850M a faster card overall despite losing Vulkan?

A: Based on average benchmark scores, the 850M holds a 9302 average versus the 650 Ti Boost's 8067. However, that average includes a weak Metal score for the 650 Ti Boost, and the head-to-head Vulkan test shows the desktop card pulling ahead significantly.

Q: What is the power consumption difference between these two?

A: The 850M is rated at 45 W with no power connectors, while the 650 Ti Boost draws 134 W and requires a single 6-pin power connector. The desktop card also lists a 300 W suggested power supply.

Q: Which GPU has better memory specifications?

A: The 650 Ti Boost uses 2 GB of GDDR5 on a 192-bit bus with 144.2 GB/s bandwidth. The 850M uses 2 GB of DDR3 on a 128-bit bus with 32.03 GB/s bandwidth, a massive difference that likely explains the Vulkan performance gap.

The Verdict

The data supports a straightforward conclusion: the GTX 650 Ti Boost is the stronger GPU for graphics-heavy workloads, while the GTX 850M is the more efficient compute part. The 650 Ti Boost wins the Vulkan test by 12.5%, which is a large enough margin to matter for modern games and applications that use this API. Its memory subsystem, with 144.2 GB/s of bandwidth versus the 850M's 32.03 GB/s, gives it a structural advantage that no amount of architectural efficiency can fully overcome. The 850M's 640 shading units and 40 texture units are also fewer than the 650 Ti Boost's 768 shading units and 64 texture units, further explaining the desktop card's higher pixel rate of 16.51 GPixel/s and texture rate of 66.05 GTexel/s versus the 850M's 14.43 GPixel/s and 36.08 GTexel/s.

For users who prioritize compute throughput in OpenCL, the 850M is the better choice, but the margin is modest at 5.6%. The 850M also has a much lower power draw at 45 W versus 134 W, making it the only viable option for thin laptops or compact systems without dedicated power connectors. The 650 Ti Boost needs a 6-pin connector, a 300 W power supply, and a dual-slot cooler, all of which are reasonable for a desktop but impossible in the 850M's integrated form factor.

The average benchmark scores favor the 850M at 9302 versus 8067, but that gap is misleading because the 650 Ti Boost's Metal score of 4858 is an outlier. In the two tests where both cards are measured directly, the 650 Ti Boost wins the more demanding Vulkan workload by a wide margin. The 850M's 46th percentile ranking versus the 650 Ti Boost's 42nd reflects the 850M's better OpenCL showing, but the desktop card's Vulkan result of 10041 is the highest single score recorded for either product.

Specification Differences

The two cards share a 28 nm process node from TSMC, but almost everything else differs. The GTX 850M uses the GM107 chip with 1,870 million transistors on a 148 mm² die, while the GTX 650 Ti Boost uses the GK106 chip with 2,540 million transistors on a 221 mm² die. The 850M has a transistor density of 12.6M per mm² versus the 650 Ti Boost's 11.5M per mm², showing the Maxwell design packs more transistors into less space.

Clock speeds are only listed for the 650 Ti Boost, which runs at a 980 MHz base and 1032 MHz boost. The 850M has no recorded base or boost clocks. Memory clocks differ dramatically: the 850M runs at 1001 MHz with 2 Gbps effective, while the 650 Ti Boost runs at 1502 MHz with 6 Gbps effective. Memory bandwidth follows suit at 32.03 GB/s versus 144.2 GB/s.

The 650 Ti Boost has 768 shading units, 64 TMUs, and 24 ROPs, producing a 16.9% higher pixel rate and an 83% higher texture rate. The 850M's FP32 output is listed as 1,585 GFLOPS versus the 650 Ti Boost's 1.585 TFLOPS, a 37.2% advantage for the desktop card, though these figures are not directly comparable due to different units. The 650 Ti Boost is a dual-slot card measuring 241 mm in length, while the 850M is an IGP with no dimensions recorded.

Power delivery is a major differentiator: the 850M is rated at 45 W with no connectors, while the 650 Ti Boost is rated at 134 W with a single 6-pin connector and a 300 W suggested power supply. Display outputs also differ, with the 650 Ti Boost offering 2x DVI, 1x HDMI 1.4a, and 1x DisplayPort 1.2, while the 850M uses portable device dependent outputs.

Architecture Differences

The GTX 850M is built on NVIDIA's Maxwell architecture with the GM107 chip, while the GTX 650 Ti Boost uses the older Kepler architecture with the GK106 chip. Both are manufactured on TSMC's 28 nm process, but Maxwell represents a generational leap in efficiency. The 850M achieves higher transistor density at 12.6M per mm² despite having fewer total transistors, and its 45 W power envelope is a fraction of the 650 Ti Boost's 134 W rating.

The 850M uses 2 GB of DDR3 memory, which is unusual for a dedicated GPU and explains its low 32.03 GB/s bandwidth. The 650 Ti Boost uses GDDR5 across a wider 192-bit bus, delivering 144.2 GB/s, more than four times the bandwidth. This architectural difference has a direct impact on the Vulkan benchmark, where memory bandwidth often becomes a bottleneck.

Shader configuration differs as well: the 850M has 640 shading units, 40 TMUs, and 16 ROPs, while the 650 Ti Boost has 768 shading units, 64 TMUs, and 24 ROPs. The Maxwell architecture compensates with higher clock efficiency per watt, but the Kepler chip has more raw resources. Both support DirectX 12 (11_0) and OpenGL 4.6, but their Vulkan support differs: the 850M lists Vulkan 1.4 while the 650 Ti Boost lists Vulkan 1.2.175, which may explain part of the Vulkan score gap.

The 850M belongs to the GeForce 800M generation, succeeding the GeForce 700M and preceding the GeForce 900M. The 650 Ti Boost belongs to the GeForce 600 generation, succeeding GeForce 500 and preceding GeForce 700. The 650 Ti Boost was released in March 2013, while the 850M came a year later in March 2014. Both are end-of-life products, but the architectural differences are significant enough that they should be treated as separate design philosophies rather than direct competitors.

Where Each One Wins

The GTX 850M wins in OpenCL compute workloads, posting a 5.6% advantage over the 650 Ti Boost. It also wins on efficiency, drawing 45 W versus 134 W, and its compact IGP form factor makes it suitable for thin laptops. The 850M's higher transistor density and Maxwell architecture suggest better performance per watt, which shows in the compute results. Its average benchmark score of 9302 also ranks higher than the 650 Ti Boost's 8067, and its 46th percentile places it above the desktop card's 42nd.

The GTX 650 Ti Boost wins decisively in Vulkan rendering, with a 12.5% lead over the 850M. Its memory bandwidth advantage of 144.2 GB/s versus 32.03 GB/s gives it a structural edge in texture-heavy scenes and higher resolutions. The desktop card also has more shading units, TMUs, and ROPs, which helps in traditional rasterization. Its pixel rate of 16.51 GPixel/s and texture rate of 66.05 GTexel/s are both higher than the 850M's figures, meaning it can fill frames faster.

For gaming on a laptop, the 850M is the only realistic option given its power and form factor constraints, but the data shows the 650 Ti Boost is the more capable graphics processor in the Vulkan workload that matters most for modern games. For compute tasks like OpenCL acceleration, the 850M is the better choice. The 650 Ti Boost also carries a launch MSRP of 169 USD, while the 850M has no recorded launch price, but the database records no pricing for either beyond that figure.