Ranking of the best PNY graphics cards in 2022

PNY is known around the world as a supplier of a wide range of industrial memory devices, graphics cards and accessories for mobile devices that meet strict standards of quality and reliability. The supplier's portfolio contains a huge list of standard and commercial products. It includes solid state drives, graphics cards, RAM modules, removable drives, and a wide range of mobile accessories.

Despite the rich product range, the company is known in the world as a supplier of NVIDIA brand video card models:

• Tesla;
• Quadro;
• Geforce.

The company does not have its own production plants and is an intermediary between consumers and producers. Why video accelerator manufacturers chose PNY is still not clear. PNY represents the products of manufacturers in the United States, as well as in Russia and the CIS countries.

Briefly about the company and its products

The company entered the US market in 1985 as a supplier of microelectronics. The headquarters is located in the US state of New Jersey. Engaged in the supply of reliable and high-quality OEM equipment for domestic and commercial use. Currently, sales offices are located around the world.

The list of deliveries includes:

• DDRI, DDRII, DDRIII DIMM memory modules;
• SO-DIMM memory modules for DDRI, DDRII, DDRIII laptops;
• Attache branded USB sticks;
• SD drives;
• NVIDIA® graphics cards Tesla®, Quadro®, and GeForce® models.

The production area of ​​the enterprise is several hectares, in the workshops of which assembly, testing, development and printing of packaging for its products are carried out. The headquarters is located on the territory of the enterprise. Product supply logistics starts here.

NVIDIA® Graphics Cards

The first video cards of this brand were produced for game consoles, but as the company grew, production technologies developed. Two years after the founding of the company, the first commercial Quadro video accelerator used on non-workstation devices appears. In 2000, the company signed a contract with Microsoft to supply GeForce cards for game consoles. Further production was established for computers and laptops.

For the first time, CUDA technology was used for accelerators, which allows increasing the computing capabilities based on graphics cards through parallel computing. In 2011, the company practically stops independent production and concludes a franchise contract with other organizations. Since then, video cards have been produced under the logos of those companies that assemble them, including under the PNY brand.

The popularity of models of this brand is due, first of all, to the use of new technologies that allow you to work with 3D images. Cards are in particular demand among gamers, designers, people professionally involved in digital video and photography.

Model GeForce®

A common brand of graphics processors and chipsets for motherboards in portable and stationary computers. It is mainly used for home and office personal computers for games, for working with graphic images that require large resources.

RTX 2080 /2070 Series

Video cards based on the Turing GPU architecture and the RTX platform. Compared to previous versions of cards in this series, the performance of this version has been increased six times. Improved rendering of game textures in real time.

• high-quality rendering of images of video games;
• ready-made drivers for a large number of games;
• virtual reality support;
• built-in NVIDIA® GeForce Experience™ for online game streaming, screenshots, and video recording;
• Ansel technology tool from NVIDIA® allows you to take high-quality photos in games;
• NVIDIA® Highlights system for recording online games;
• compatible with NVIDIA® G-SYNC™ adaptive sync technology to eliminate screen tearing;
• support for DirectX® 12 console interfaces from Microsoft®, Khronos Group Vulkan and OpenGL version 4/5;
• signal interface standard 1.4 DP and 2.0-b HDMI;
• compatible with USB type-C and VirtualLink 4;
• HDCP version 2.2 protection system for 4k video;
• NVIDIA® graphics accelerator;
• NVLink Internal Link (SLI) logic from NVIDIA®.

The RTX 2080 series comes in several flavors.Below are tables of their technical characteristics.

	Titanium 11GB-XLR8 Dual Fan	Titanium 11GB single fan	8GB-XLR8 Dual Fan	8GB-XLR8 for dual fan gaming	8GB -TX 208 single fan	8GB-XLR8 Dual Fan	8GB with one fan
Number of CUDA Cores	4352	4352	2944	2944	2944	2304	2304
Core overclocking	1350 MHz	1350 MHz	1515 MHz	1515 MHz	1515 MHz	1410 MHz	1410 MHz
Maximum core overclock	1635 MHz	1545 MHz	1815 MHz	1815 MHz	1710 MHz	1710 MHz	1620 MHz
Memory size	11GB with GDDR6 memory chip	11GB with GDDR6 memory chip	8GB with GDDR6 memory chip	8GB with GDDR6 memory chip	8GB with GDDR6 memory chip	8 GB with GDDR6 memory chip	8GB with GDDR6 memory chip
Memory interface	352-bit	352-bit	256-bit	256-bit	256-bit	256-bit	256-bit
Reading speed	616 GB/s	616 GB/s	448 GB/s	448 GB/s	448 GB/s	448 GB/s	448 GB/s
heat sink	285 W	285 W	285 W	285 W	285 W	185 W	175 W
Scalable communication interface	NVLink 2-way	NVLink 2-way	NVLink 2-way	NVLink 2-way	NVLink 2-way	Not supported	Not supported
Video output standards	DispPort 1.4, HDMI 2.0-b, USB type C	DispPort 1.4, HDMI 2.0-b, USB type C	DispPort 1.4, HDMI 2.0-b, USB type C	DispPort 1.4, HDMI 2.0-b, USB type C	DispPort 1.4, HDMI 2.0-b, USB type C	DVI-D, DispPort 1.4, HDMI 2.0-b, USB type C	DisPort 1.4, HDMI 2.0-b, DVI-D
Multiwindow	available	available	available	available	available	available	available
Maximum resolution capability	7680x4320 @ 60Hz	7680x4320 @ 60Hz	7680x4320 @ 60Hz	7680x4320 @ 60Hz	7680x4320 @ 60Hz	7680x4320 @ 60Hz	7680x4320 at 60Hz
Tire	2x8-pins	2x8-pins	1x6-pin and 1x8-pin	1x6-pin and 1x 8-pin	1x6-pin and 1x8-pin	1x8-pins	1x8-pins
Tire type	RSIE3.0	RSIE3.0	RSIE3.0	RSIE3.0	RSIE3.0	RSIE3.0	RSIE3.0
Map dimensions	1.73x12.36x5.04	1.65x11.06x5.04	1.57x 11.50x4.41	1.57x 11.50x4.41	1.65x11.06x5.04	235x112	281x128
Width	double slot	double slot	double slot	double slot	double slot	double slot	double slot
Virtual reality	available	available	available	available	available	available	available
Price, $/rub	1 670/103 130	948/104 00	1 203/61 290	799/59 140	948/64 100	629/59 140	948/64 100

GTX 1070/1080 Series

New flagship graphics card based on the NVIDIA Pascal architecture. It is the latest gaming graphics card with improved rendering of game textures, a GDDR5X swap buffer memory array and virtual reality technology.

• simultaneous multi-projection;
• virtual reaction;
• technological tool Ansel;
• support for SLI connection and HB Bridge type;
• compatible with screen tear sync technology;
• GameStream technology;
• accelerator GPU Boost series 3.0;
• support for Microsoft DirectX-12, Vulkan plugins and 3D Open GL;
• PCIe 3.0 bus;
• certified for Windows and Linux;
• maximum resolution: 7680x4320 at 60 Hz;
• video outputs: HDMI 2.0-b, DP 1.4, DLDVI and HDCP type 2.2;
• support for the latest DP type 1.4 interface.

Below is a table of characteristics.

	1080 Titanium single fan	1080 XLR8	1080 XLR8	1080 FOUNDERS issue	1080 single fan	1070 Titanium Dual Fan	1070 XLR8 Dual Fan
Number of CUDA Cores	3584	2560	2560	2560	2560	1607	1920
Basic overclocking	1480 MHz	1708 MHz (versus 1607)	1620 MHz	1607 MHz	1607 MHz	1683 MHz	1518 MHz
Maximum overclocking	1582 MHz	1848 MHz (versus 1734)	1759 MHz	1733 MHz	1734 MHz	1708 MHz	1708 MHz
Standard memory configuration	11 GB with GDDR5X memory chip	8GB with GDDR5X memory chip	8GB with GDDR5X memory chip	8GB with GDDR5X memory chip	8GB with GDDR5X memory chip	8GB with GDDR5 memory modules	8GB with GDDR5 memory modules
Memory interface	352-bit	256-bit	256-bit	256-bit	256-bit	256-bit	256-bit
Memory performance	11 Gb/s	10 Gb/s	10 Gb/s	10 Gb/s	320 Gb/s	256 Gb/s	8 Gb/s
card power	250 watts	190 watts	190 watts	180 watts	180 watts	180W	180 watts
SLI connection	SLI connection/HB bridge	SLI connection/HB bridge	SLI connection/HB bridge	SLI connection/HB bridge	SLI connection / HB-Bridge bridge	SLI connection / HB-Bridge bridge	SLI connection / HB-Bridge bridge
video receiver	DP1.4, HDMI2.0b	DP1.4, HDM2.0b, DLDVI	DP1.4, HDMI2.0b, DLDVI	DVI, DP1-4, HDMI	DP1.4, HDMI2.0b, DLDVI	DP1.4, HDMI2.0b, DLDVI	DP1.4, HDMI2.0b, DLDVI
Permission	7680x4320 @ 60Hz	7680x4320 @ 60Hz	7680x4320 @ 60Hz	7680x4320 @ 60Hz	7680x4320 @ 60Hz	7680x4320 @ 60Hz	7680x4320 @ 60Hz
Tire	6-pin, 8-pin	1x8-pin PCIE	1×8-pin PCIE	1 x PCIE 8-pin	1×8-pin PCIE	one 8-pin	1×8-pin PCIE
Standard Support	PCIE Generation 3.0	RSIE3.0 16x	RSIE3.0 16x	RSIE3.0 16x	RSIE3.0 16x	PCIE3.0x16	PCIE3.0x16
Price, $/rub	699/51 401	499/58 899	924/32 650	699/51 401	699/51 401	499/58 899	924/32 650

Support for PCIe generation 3.0 PCIE 3.0 16x PCIE 3.0 16x PCIE 3.0 16x PCIE 3.0 16x PCIE 3.0×16 PCIE 3.0×16
The series belongs to the middle price category and has less performance than their expensive counterparts.

Model QUADRO®

Performance cards based on the Turing architecture and the RTX platform.They are used at workstations with high computing power. Designed for use by professional graphic designers and artists where performance, color fidelity and large memory capacities are important. This efficiency was achieved through the use of RT and Tensor and Volta cores, as well as through cloud graphical computing technology and a deep analysis system that allows you to work with software products like FEA, CFD, CEM and others that use graphical code for analysis.

Table of characteristics of video cards with RT cores.

	RTX-6000	RTX-5000
Number of CUDA Cores	4608	3072
Number of NVIDIA RT cores	72	48
Number of NVIDIA Tensor Cores	576	384
Memory	24 GB GDDR6 with ECC	16 GB GDDR6 with ECC
RTX-OPS	84T	62T
Raycasting	10 Gigabeams/sec	8 Gigabeams/sec
Peak Performance FP32	16.3TFLOPS	11.2TFLOPS
Peak Performance FP16	32.6TFLOPS	22.3TFLOPS
Peak performance INT8	206.1 TOPS	178.4 TOPS
TFLOPS deep learning technique	130.5 Tensor TFLOPS	89.2 Tensor TFLOPS
Graphics scalability	NVLink	NVLink
NVLink channel	100 GB/s	50 GB/s
memory channel	624 GB/s	870 GB/s
System interface	PCIE3.0x16	PCIE3.0x16
Power consumption	295 W	265 W
Form Factor	4.4x10.5, double slot	4.4x10.5, double slot
video connector	DP1.4, VirtualLink	DP1.4, VirtualLink
Connector DVI-D single-channel	yes, with adapter	yes, with adapter
Support HDMI CONNECTOR	yes, with adapter	yes, with adapter
Number of screens	4	4
Maximum dilution capability DP 1.4	HDR 7680x4320 @ 60Hz	HDR 7680x4320 @ 60Hz
5K video support	HDR 5120x2880 @ 60Hz	HDR 5120x2880 @ 60Hz
4K video support	HDR 4096x2160 or 3840x2160 @ 120Hz	HDR 4096x2160 or 3840x2160 @ 120Hz
Maximum dilution capability DVI-D DL	2560x1600 @ 60Hz	2560x1600 @ 60Hz
Maximum dilution capability DVI-D SL	1920x1200 @ 60Hz	1920x1200 @ 60Hz
HDCP support	Yes	Yes
Compatible with Quadro Sync II	Yes	Yes
Compatible with NVIDIA GPU Direct	Yes	Yes
Graphics APIs	Shader, DirectX, Vulkan, OpenGL	Shader, DirectX, Vulkan, OpenGL
Computing APIs	CUDA; direct computation and 3D OpenCL	CUDA, Direct Computing, and 3D OpenCL
MOSAIC	available	Available
NVIEW	available	available
Price, $/rub	6 299/420 00	2 299/153 000

Graphic card rating

Questions often come up on the forums: how to choose a card by price and quality, which companies are better, what are the criteria for choosing good accelerators, and many others. There is also an ambiguous opinion that even the simplest electronics can outperform the expensive ones. If CUDA-enabled graphics devices are quick to dispel such myths, the benefits of the rest depend on their performance and speed, memory size and operating frequency.

Below is a table of rating of graphic devices by their average cost.

Position	Name	Price, $
1	Quadro RTX-6000	6300
2	Quadro RTX-5000	2300
3	GeForce RTX-2080	1200
4	GeForce RTX-2070	1200
5	GeForce GTX-1080	700
6	GeForce GTX-780	700
7	GeForce GTX-1070	550
8	GeForce GTX-680	400
9	GeForce GTX-1060	330
10	GeForce GTX-970	300

Professional video cards are in the upper price segment, but they have higher performance and support many technologies for working with photo and video materials. This is their advantage over basic digital data visualization devices.
As for the base cards, their differences in frequency and performance are not very different. The range of prices is more likely due to the presence of functionality and memory capacity, as well as the introduction of new technologies.

Overview of video cards

The bandwidth of Quadro processors is higher than gaming GeForce. Therefore, when working with CAD applications, they work much faster, sharpened to use OpenGL. This is noticeable in the speed of drawing three-dimensional graphics. However, for many 3D design plug-ins, there are no optimized drivers that reduce the performance of even powerful devices. For this reason, some software developers are starting to abandon the use of specialized drivers, so ordinary game cards can handle the rendering task quite well. As far as SLI and CrossFire communication systems are concerned, game card users bypass dual-core solutions, preferring that their cards work not in a bundle, but separately from each other. Separate work of video cards allows you to use different models of third-party manufacturers.

The presented novelties work on the Turing architecture, but in all respects it resembles the Volta. As experts suggest, the company simply decided to give a new name, slightly changing the functionality. The number of transistors reaches 18 billion, which corresponds to the number of CUDA cores of 4608. The company announced its products as a ray tracing solution, which is 10 Gigabeams / sec on the RTX 6000 model. The NVLink interface allows you to use video cards in a bundle, which improves overall performance.

The use of the abbreviation RTX and GTX in the product name is primarily due to the use of new RTX and GTX technologies based on the Pascal ray tracing architecture. The innovation takes visualization to a whole new level. This is a completely different technique of reflecting models, darkening, refraction and transmission of light, the appearance of shadows.

Reviews

In their reviews on the markets, buyers note that the cards do not heat up much, they have a good cooling system with several fans. Water-cooled boards do not heat up at all, which keeps their high performance constant. Coolers operate very quietly and do not interfere with work. They are an excellent solution for working with 3D design software products. The price corresponds to the quality, but there are cheaper analogues from other manufacturers. From time to time there are problems with drivers that crash after a while. Consumes a large amount of electricity.

The advantages of the boards include:

• high performance;
• good cooling system;
• good image quality.

Many users note that such cards should be taken for games and programs that require large system resources.

The disadvantages include:

• high price;
• poor quality drivers;
• high power consumption.

Earlier models do not have sufficient cooling, which leads to overheating of the board and reduced performance.

Conclusion

In comparison with models of the same class of early generations, the increase in productivity is obvious. It is increased by six, when cards from other manufacturers show an increase of only 15%.Flagship innovations allow you to play all the games that exist today with a video resolution of 4k.