NVIDIA GeForce 920M vs NVIDIA GeForce GT 745M Comparison

Head-to-Head Benchmarks

The recorded data shows two distinct benchmark tests where these mobile GPUs face off: Geekbench OpenCL and Geekbench Vulkan. The results are split exactly one win each, but the margins tell a very different story about where each part excels.

In Geekbench OpenCL, the NVIDIA GeForce 920M takes the win with a score of 3725 against the GT 745M's 3580. That is a delta of 3.9 percent in favor of the 920M. This is a modest but real advantage for compute workloads that rely on OpenCL. The 920M's higher clock speeds (954 MHz base and boost) compared to the GT 745M's unspecified clocks likely explain this edge. The 920M also delivers significantly higher raw throughput numbers: 732.7 GFLOPS of FP32 performance versus 421.6 GFLOPS on the GT 745M. In texture fill rate, the 920M reaches 30.53 GTexel/s versus 17.57 GTexel/s. These are substantial computational advantages that show up most clearly in GPU-accelerated compute tasks.

The Geekbench Vulkan result flips the script entirely. The GT 745M scores 5502, while the 920M manages only 2849. That is a 93.1 percent advantage for the older GT 745M, nearly doubling the 920M's output. This is a massive gap and suggests the GT 745M has a far more robust Vulkan implementation or driver support in this test environment. For any workload that relies on Vulkan, the GT 745M is in a completely different class. The GT 745M's memory configuration also helps in bandwidth-hungry scenarios: 64.00 GB/s of bandwidth from GDDR5 on a 128-bit bus versus the 920M's 14.40 GB/s from DDR3 on a 64-bit bus. That fourfold-plus bandwidth advantage is decisive in many real-world rendering tasks.

Looking at the average benchmark scores, the GT 745M holds a clear overall lead. Its average across all recorded tests is 3953, while the 920M averages 3287. That is a 20.3 percent gap in favor of the GT 745M. The GT 745M also places in the 23rd percentile of all GPUs in the database, versus the 920M's 20th percentile. Neither chip is a high performer by modern standards, but the GT 745M is measurably stronger overall.

FAQ

Q: Which GPU has the higher average benchmark score?

A: The GT 745M averages 3953 across all recorded tests, while the 920M averages 3287, a difference of roughly 20 percent.

Q: Why does the Geekbench Vulkan result differ so dramatically between the two?

A: The GT 745M scores 5502 in Vulkan, versus 2849 for the 920M, a 93.1 percent difference. The GT 745M's GDDR5 memory and wider 128-bit bus, giving 64.00 GB/s of bandwidth, are likely major contributors to this advantage.

Q: Which GPU has the higher clock speed?

A: The 920M has a base and boost clock of 954 MHz. The GT 745M's base and boost clocks are not listed in the database, but its memory runs at 1000 MHz (4 Gbps effective) versus the 920M's 900 MHz (1800 Mbps effective).

Q: What is the transistor density difference between the two chips?

A: The 920M's GK208B chip has a higher transistor density at 11.7M per mm², packing 1,020 million transistors into an 87 mm² die. The GT 745M's GK107 packs 1,270 million transistors into a 118 mm² die, for a density of 10.8M per mm².

Q: Which GPU is more power efficient?

A: The 920M has a TDP of 33 W, while the GT 745M is rated at 45 W. The 920M draws less power while delivering higher FP32 compute (732.7 GFLOPS versus 421.6 GFLOPS), but the GT 745M wins on memory bandwidth and Vulkan performance.

Q: Are these GPUs still in production?

A: Both are end-of-life products. The GT 745M was released in 2013, and the 920M followed in 2015.

The Verdict

From the data, the GT 745M is the stronger overall GPU. Its average benchmark score of 3953 versus 3287 for the 920M represents a meaningful lead. The Vulkan result is the standout: 5502 versus 2849, a 93.1 percent advantage. Anyone running Vulkan-based games or applications should favor the GT 745M without hesitation.

The 920M has its own niche. It leads in OpenCL by 3.9 percent, and it does so at a lower TDP of 33 W versus 45 W. For users who prioritize compute workloads under OpenCL and want lower power draw, the 920M is the better fit. It also offers a higher clock speed at 954 MHz, which contributes to its raw FP32 and texture rate numbers.

For a general-purpose laptop GPU, the GT 745M's memory subsystem is the deciding factor. Its GDDR5 memory on a 128-bit interface provides 64.00 GB/s, a figure the 920M's DDR3 on a 64-bit bus cannot approach (14.40 GB/s). The database shows that bandwidth advantage translates into a massive Vulkan lead, and it will likely help in texture-heavy workloads.

The GT 745M's percentile ranking is slightly higher (23rd versus 20th), and its nearest competitors (AMD Radeon R5 M420 at 3956, NVIDIA GeForce 830M at 3957, NVIDIA Quadro K2000 at 3964) all score within 0.3 percent of it. The 920M's nearest rivals are lower-scoring parts like the NVIDIA GeForce GT 730M (3316), Intel HD Graphics 530 (3332), and GT 640 (3210), which reinforces its position as a lower-tier mobile chip.

Specification Differences

The two GPUs differ in several key specifications. The GT 745M uses a 128-bit memory bus, while the 920M uses a 64-bit bus. Memory type differs: GDDR5 on the GT 745M versus DDR3 on the 920M. This produces a large bandwidth gap: 64.00 GB/s versus 14.40 GB/s. Memory clock speeds are 1000 MHz (4 Gbps effective) for the GT 745M and 900 MHz (1800 Mbps effective) for the 920M.

The GT 745M has 16 ROPs, while the 920M has only 8, a 2x difference. The 920M has higher pixel rate (7.632 GPixel/s versus 4.392 GPixel/s) and texture rate (30.53 GTexel/s versus 17.57 GTexel/s), despite having fewer ROPs, because of its higher clock speed. FP32 performance goes to the 920M at 732.7 GFLOPS versus 421.6 GFLOPS.

The 920M has an official base and boost clock of 954 MHz; the GT 745M's clocks are not listed in the database. The 920M also has a lower TDP (33 W versus 45 W) and uses a PCIe 3.0 x8 interface, while the GT 745M uses PCIe 3.0 x16. The 920M has no power connector listed, while the GT 745M's is not specified. Both share the same 2 GB memory size, 384 shading units, and 32 TMUs.

Architecture Differences

The GT 745M is based on the GK107 chip under the Kepler architecture, while the 920M uses the GK208B chip under Kepler 2.0. Both are fabricated on a 28 nm process at TSMC, but the dies differ: the GT 745M has 1,270 million transistors on a 118 mm² die, while the 920M has 1,020 million transistors on an 87 mm² die.

The transistor density is higher on the 920M (11.7M per mm² versus 10.8M per mm²), showing a more efficient packing of transistors, likely due to the later design. The 920M's GK208B is also a newer chip, released in 2015 versus 2013 for the GT 745M.

Both GPUs support DirectX 12 (11_0), OpenGL 4.6, and Vulkan 1.2.175, according to the database. Neither has ray tracing cores or tensor cores. The GT 745M's generation is GeForce 700M, while the 920M is GeForce 900M, but both are end-of-life products. The GT 745M's predecessor is GeForce 600M and successor is GeForce 800M; the 920M's predecessor is GeForce 800M and successor is GeForce 10 Mobile.

The architecture differences boil down to the newer GK208B chip in the 920M, which brings higher clocks and compute throughput, but the GT 745M's larger die and memory bus give it an advantage in memory-bound tasks. The 920M's higher transistor density suggests a more refined design, but the GT 745M's better memory system is the more significant factor in the benchmark results.

Where Each One Wins

The GeForce GT 745M wins clearly in Vulkan graphics workloads. The 93.1 percent advantage in Vulkan benchmark is a dominant force, and it is backed by the much higher memory bandwidth of 64.00 GB/s. For users running Vulkan games, compute shaders, or modern graphics APIs, the GT 745M is the obvious pick. It also wins on average benchmark score and percentile ranking, so for general mixed use it is faster overall.

The GeForce 920M wins in OpenCL compute workloads. Its 3725 versus 3580 score shows a 3.9 percent advantage, and its FP32 throughput of 732.7 GFLOPS is 73.7 percent higher than the GT 745M's 421.6 GFLOPS. Its texture rate of 30.53 GTexel/s is also much higher. The 920M also has a lower TDP at 33 W versus 45 W, making it a better fit for thinner, cooler laptops where sustained load matters.

If you prioritize Vulkan gaming or any workload that leans on memory bandwidth, choose the GT 745M. If you need raw compute speed in OpenCL and want to reduce power consumption, the 920M is the better pick. The data gives each GPU one win, but the margin of victory matters. The GT 745M's Vulkan lead is huge (93.1 percent), while the 920M's OpenCL lead is a smaller 3.9 percent. The GT 745M also wins in the overall average and percentile rankings, so for a general-purpose laptop GPU, it is the stronger recommendation based on the recorded measurements.