GeForce GTX 650 vs. GTX 1650 Performance For Linux Gaming, Performance-Per-Watt

The latest in our benchmarking with the new GeForce GTX 1650 is some "fun" tests seeing how its performance compares to that of the GeForce GTX 650 Kepler. Various OpenGL and Vulkan Linux gaming tests were carried out as well as some compute tests and throughout monitoring the AC power consumption to yield the performance-per-Watt metrics.

The GeForce GTX 650 Kepler graphics card launched nearly seven years ago already with its 40nm GK107 GPU that provided 384 CUDA cores, 1058MHz core clock speed, and 1GB of GDDR5 video memory. The GTX 650 has a 64 Watt TDP but came with a 6-pin PCIe power connector. The GTX 650 / Kepler is the last generation currently supported by the mainline Linux driver and also the oldest NVIDIA hardware with Vulkan driver coverage, making this comparison particularly interesting.

The GeForce GTX 1650 meanwhile with its 12nm TU117 GPU has 896 CUDA cores, 1485MHz base clock speed, 1665MHz boost clock speed, and 4GB of GDDR5 video memory. The GeForce GTX 1650 has a 75 Watt TDP without the need for any external PCIe power connection.

We've already run Linux tests like Radeon RX 560/570/580 vs. GeForce GTX 1060/1650/1660 Linux Gaming Performance while this article is just looking at the GTX 650 vs. GTX 1650 for historical reference. A variety of benchmarks that could run on the GTX 650 were carried out using the Phoronix Test Suite on both graphics cards. The system was running Ubuntu 19.04 with the Linux 5.0 kernel and NVIDIA 430.09 driver. The AC system power consumption was being polled by the Phoronix Test Suite with a WattsUp Pro power meter for generating the overall performance-per-Watt metrics.