Late last year, NVIDIA released the GeForce GTX 1650 Super graphics card. This is the second GTX 16-series Super model from NVIDIA as the company makes product-stack adjustments across the bottom-end of its "Turing" GPU lineup. These graphics cards are crucial for the company's bottom line as they sell in huge volumes; the sub-$200 market is the gateway to Full HD gaming at 60 FPS. These are the price points at which gamers who choose the PC over consoles spend some money to upgrade their home desktops to gaming-capable machines. AMD's Radeon RX 5500 graphics card threatens NVIDIA's GTX 16-series just in the same way the RX 5700 destabilized the lower end of the RTX 20 lineup, forcing three new product launches to prevent cannibalization.
The RTX 2060 Super was launched as the RTX 2060 lost competitiveness to the RX 5700, and the RTX 2070 Super as the RX 5700 XT beat the RTX 2070. To prevent the RTX 2070 Super from hurting sales of NVIDIA's $700 RTX 2080, NVIDIA also refreshed the RTX 2080 with the faster RTX 2080 Super. With the GTX 16-series, the challenge of keeping these SKUs competitive is higher as NVIDIA can't play its "ray tracing" card here. There is a level playing field between the GTX 16-series and the RX 5500 series—pure raster graphics.
The GeForce GTX 1650 Super was created to restore NVIDIA's competitiveness in the sub-$200 market, as AMD's RX 5500 easily beat the original GTX 1650 in its marketing slides, with a large 30% performance advantage. This invited a two-fold response from NVIDIA. The GTX 1660 Super was launched at $230, offering performance close to the $280 GTX 1660 Ti, and the GTX 1650 Super was launched to restore competition under the $200 mark. The GTX 1650 Super is being launched at an MSRP of $160 with headroom for board partners to price their custom-design cards at up to $200 or thereabouts.
Unlike the original GTX 1650 non-Super, the new GeForce GTX 1650 Super is based on the "TU116" silicon rather than the tiny "TU117" that powers the GTX1650. "TU116" is the same chip on which NVIDIA built its GTX 1660 trio; the "TU116" as the GTX 1650 Super is configured with a 128-bit GDDR6 memory interface holding 4 GB of memory. Even at its memory clock of 12 Gbps, this setup produces 192 GB/s of memory bandwidth, which is a massive 50% increase compared to the 128 GB/s of the original GTX 1650. NVIDIA also endowed the GTX 1650 Super with more muscle than the original—1,280 CUDA cores compared to the original's 896, a 42 (!) percent increase. There are proportionate increases in TMUs: 80 vs. 56. The GPU clock speeds have also been dialed up to 1725 MHz GPU Boost compared to the original's 1665 MHz. These changes also increase the card's typical board power metric to 100 W, up from 75 W on the original. The card needs at least a 6-pin PCIe power connector, while the original GTX 1650 could make do with none.
In this review, we take a close look at the Gigabyte GeForce GTX 1650 Super WindForce OC, a cost-optimized custom-design that is priced just $10 above NVIDIA MSRP, or $170. Included is a dual-fan heatsink, a backplate, and a small factory overclock.
|GTX 1050||$135||640||32||1354 MHz||1455 MHz||1752 MHz||GP107||3300M||2 GB, GDDR5, 128-bit|
|GTX 1050 Ti||$150||768||32||1290 MHz||1392 MHz||1752 MHz||GP107||3300M||4 GB, GDDR5, 128-bit|
|GTX 1650||$150||896||32||1485 MHz||1665 MHz||2000 MHz||TU117||unknown||4 GB, GDDR5, 128-bit|
|RX 570||$120||2048||32||1168 MHz||1244 MHz||1750 MHz||Ellesmere||5700M||4 GB, GDDR5, 256-bit|
|RX 5500 XT||$170||1408||32||1717 MHz||1845 MHz||1750 MHz||Navi 14||6400M||4 GB, GDDR6, 128-bit|
|GTX 1650 Super||$160||1280||32||1530 MHz||1725 MHz||1500 MHz||TU116||6600M||4 GB, GDDR6, 128-bit|
|Gigabyte GTX 1650|
Super WindForce OC
|$170||1280||32||1530 MHz||1755 MHz||1500 MHz||TU116||6600M||4 GB, GDDR6, 128-bit|
|RX 580||$180||2304||32||1257 MHz||1340 MHz||2000 MHz||Ellesmere||5700M||8 GB, GDDR5, 256-bit|
|GTX 1060 3 GB||$170||1152||48||1506 MHz||1708 MHz||2002 MHz||GP106||4400M||3 GB, GDDR5, 192-bit|
|GTX 1060||$210||1280||48||1506 MHz||1708 MHz||2002 MHz||GP106||4400M||6 GB, GDDR5, 192-bit|
|RX 590||$195||2304||32||1469 MHz||1545 MHz||2000 MHz||Polaris 30||5700M||8 GB, GDDR5, 256-bit|
|GTX 1660||$220||1408||48||1530 MHz||1785 MHz||2000 MHz||TU116||6600M||6 GB, GDDR5, 192-bit|
|GTX 1070||$300||1920||64||1506 MHz||1683 MHz||2002 MHz||GP104||7200M||8 GB, GDDR5, 256-bit|
|RX Vega 56||$300||3584||64||1156 MHz||1471 MHz||800 MHz||Vega 10||12500M||8 GB, HBM2, 2048-bit|
|GTX 1660 Super||$230||1408||48||1530 MHz||1785 MHz||1750 MHz||TU116||6600M||6 GB, GDDR6, 192-bit|
|GTX 1660 Ti||$275||1536||48||1500 MHz||1770 MHz||1500 MHz||TU116||6600M||6 GB, GDDR6, 192-bit|
|GTX 1070 Ti||$450||2432||64||1607 MHz||1683 MHz||2000 MHz||GP104||7200M||8 GB, GDDR5, 256-bit|