PNY GeForce RTX 5080 Slim OC Review

PNY GeForce RTX 5080 Slim OC in a Compact Format

Now that the GeForce RTX 5080 graphics cards have been on the market for some time and NVIDIA’s board partners have established their own models, there are several refined models tailored to specific needs. This successfully replaces the ** GeForce RTX 4080 (Review), which can now only be found on the market as used merchandise. The RTX 50 Super series has yet to be released, making the GeForce RTX 5080 the high-end model accessible to gamers; thanks to technological improvements in the Blackwell architecture, is likely the best gamers will treat themselves to. While the GeForce RTX 5090 is the flagship model, its price has risen even further and now costs more on its own than a **modern gaming PC should cost in total.

Blackwell GB203 Chip

PNY GeForce RTX 5080 Slim OC GPUz (Image © PCMasters.de)

Naturally, PNY’s GeForce RTX 5080 Slim OC also relies on the Blackwell-generation GB203 chip at its core. With an area of 378 mm², the chip is only about half the size of the full-size ** GeForce RTX 5090 (review), which has 750 mm², but still manages to deliver tremendous performance. Roughly speaking, the chip is almost identical in size to the AD103 chip found in the GeForce RTX 4080 and **GeForce RTX 4080 Super (review), which has a total area of 379 mm². In terms of transistor count, the GeForce RTX 5080 has 45.6 billion, compared to the 45.9 billion transistors in the GeForce RTX 4080 (Super)—a minimal difference. With 10,752 streaming processors (CUDA cores), we still get a higher total compared to the 10,240 CUDA cores in the GeForce RTX 4080 Super and 9,728 cores of the GeForce RTX 4080. The GeForce RTX 5080’s monolithic chip is manufactured by TSMC using the advanced 4-nm process, which is the case for all cards in the RTX 50 series.

Blackwell SM Architecture for More Performance

According to documentation published by NVIDIA, the Blackwell graphics processor features a redesigned Streaming Multiprocessor (SM) architecture that differs from that of the previous generation. As a result of this redesign, each SM now contains a total of 128 FP32 CUDA cores, arranged in 7 Graphics Processing Clusters (GPCs) and organized around a single raster engine.

The Blackwell architecture consolidates the separate 16 "FP32/INT32" units into 32 unified units. Thanks to this design change, the graphics processor can perform either 32 FP32 (floating-point) or 32 INT32 (integer) operations simultaneously, offering greater flexibility and computational power.

The GeForce RTX 5080 also features 84 4th-generation ray tracing cores and as many as 336 5th-generation The RT cores are responsible for light particle calculations and are used for ray tracing in games to achieve greater realism. The Tensor cores are primarily used for DLSS features and running AI models. Another new addition to the Blackwell generation is the 4th generation of DLSS with Multi-Frame Generation. We’ve already reached version 4.5 of DLSS, which demonstrates the technology’s maturity. NVIDIA is currently highlighting this technology because, at first glance, it appears to deliver twice or three times as many frames per second—which, unfortunately, isn’t entirely accurate when examined in detail, since these are by no means real frames, but rather statistically predicted ones. They serve, so to speak, as fillers between two frames, allowing NVIDIA to advertise a tripling of frames—a claim not accepted or tolerated by all users.

The Founders Edition reference design shown by NVIDIA was compact, and PNY’s Slim Edition is set to prove itself against it. In terms of length, they’re almost identical, but there’s a difference in height. PNY uses larger 120-mm fans, while NVIDIA relies on 115-mm variants with massive fan blades.

Measuring 300 x 150 x 40 mm, the GeForce RTX 5080 Slim OC is slightly shorter and 13 mm taller than the Founders Edition’s 304 x 137 x 40 mm dimensions. In the RTX 5080 segment, the PNY GeForce RTX 5080 Slim OC is therefore quite compact.

The current RTX 50 generation features a 5th-generation PCI Express interface. The theoretical maximum bandwidth of 31.5 GB/s is doubled on paper to 63 GB/s. The GeForce RTX 5080 is capable of utilizing 16 PCIe lanes; on our test system, it initially operated with only 8 lanes and, after a few reboots, with 16 lanes. We’ve observed this behavior with other RTX 5080 models as well. However, this isn’t a problem, since 8 PCIe Gen 5 lanes deliver the same performance as 16 Gen 4 lanes.

12V-2x6 Power Connector and Adapter

PNY, too, must rely on the updated 12V-2x6 power connector recommended by NVIDIA for its GeForce RTX 5080 Slim OC graphics card—originally known as the 12VHPWR connector. This revised interface was intended to provide greater safety and reliability when handling high power loads, but the problem remains unresolved to this day. We’ve also already seen a case of fire damage involving a GeForce RTX 5090. So the problem is by no means a thing of the past. The GeForce RTX 5080 can reach a total power draw of up to 360 watts, which isn’t quite as extreme as with an RTX 5090, but it’s still quite substantial. The peak power consumption can even be much higher for short periods.

PNY’s solution also includes the sensor cable, which enables direct communication with the power supply; however, the pins are likely not measured individually, which is the problem with NVIDIA’s design. We didn’t experience any issues with the cable during the testing period. For systems without ATX 3.1 -compliant power supplies, the included adapter converts the single 12V-2x6 connector into three standard 8-pin PCIe power connectors. In test configurations, this approach was successfully used with multiple connectors on standard power supplies.

16 GB GDDR7 Graphics Memory

memory. The PNY GeForce RTX 5080 Slim OC features 16 GB of this memory and also allows for significant overclocking of the memory clock speed. The previous generation used GDDR6X, and AMD still "only" offers the slightly more affordable GDDR6 memory.

The PNY GeForce RTX 5080 Slim OC uses eight Samsung memory modules with the part number K4VAF325ZC-SC32. These modules are organized into cells of 512 M x 32. While NVIDIA specifies data transfer rates of up to 30 Gb/s, the Samsung chips reportedly offer as much as 32 Gb/s according to their specifications. This increase in speed per pin results in a significant boost in total memory bandwidth.

The GeForce TX 5080 achieves a bandwidth of 960 GB/s, representing a measurable performance gain compared to the GeForce RTX 4080 Super’s throughput of 716.8 GB/s.

Advancements in GDDR7 Signal Transmission

The switch to GDDR7 brings a change in the method of data transmission. The architecture has shifted from PAM4 signal transmission—which transfers 2 bits per cycle—to PAM3. Previously, 1.5 bits per cycle were transferred. Although the number of bits per cycle is lower, this transition significantly improves the signal-to-noise ratio. Furthermore, this design allows for a doubling of independent channels without requiring higher I/O density on the board. Together with improvements in I/O training and clock architecture, these changes optimize data flow and reduce electrical overhead.

The architectural refinements in GDDR7 go beyond mere speed. The updated equalization and I/O train mechanisms contribute to improved power management.

PNY GeForce RTX 5080 Slim OC Technical Specifications

Below, we have compiled a table of all key specifications for the GeForce RTX 5080 from KFA2 and the Nvidia reference model, and included the GeForce RTX 4080, GeForce RTX 4080 Super, RTX 5070 Ti, RTX 5070, and GeForce RTX 4090 for comparison.

Connectivity for Four Monitors

The PNY GeForce RTX 5080 Slim OC comes with a standard configuration of output ports. It supports multi-monitor setups, which are enabled by three DisplayPort ports and one HDMI port. The three available DisplayPort 2.1b ports allow the graphics card to handle significantly higher resolutions and faster refresh rates compared to earlier versions, ensuring efficient data throughput.

In addition, there is an HDMI 2.1b port designed for high-quality video output. By utilizing Display Stream Compression (DSC), this port supports extreme display parameters, including 4K resolution at a refresh rate of 480 Hz or 8K resolution at 120 Hz. In addition, the HDMI output supports Variable Refresh Rate (VRR) and High Dynamic Range (HDR).

Design and Construction

In terms of design, we would definitely describe the PNY GeForce RTX 5080 Slim OC as understated. The black finish is well executed overall, and from certain angles, the black coating appears slightly reflective. The design is slightly enlivened by two continuous lines on the front. The stickers on the fan hub are very subtle and highlight the PNY logo in a minimalist way.

The back is covered by a black-painted metal backplate. The white lettering appears a bit too bright because it stands out too much against the black finish. In contrast, all the compliance logos at the end of the card are very subtly painted in gray and are therefore barely noticeable. In gray, the PNY logo certainly have looked good as well, and the GeForce RTX 5080 lettering also appears simply too bright.

We wouldn’t classify this graphics card as an SFF model per se, but we can still understand why PNY and NVIDIA define it that way. For a card of the RTX 5080’s caliber, this graphics card is rather compact and is almost on par with NVIDIA’s Founders Edition. The proportions seem a bit strange at first glance when you hold the card in your hand for the first time. We’re more used to elongated graphics , so the Slim OC model feels very unfamiliar.

Once installed, it definitely fits better, and the graphics card then appears relatively compact.

Another striking feature on the back is that the backplate has a large cutout in the center for the 12V-2x6 power connector. This once again highlights that the graphics card is taller rather than wider. The cutout has the positive effect of giving the cable more room to bend and, so to speak, protecting the graphics card from being bent too sharply.

When viewed from the side, the first thing you notice is a raised section around the fans, which is approximately 3 mm high. The plastic cover extends over the corners on all sides and also helps direct the airflow.

GeForce RTX 5080 Slim OC Overclocking

PNY has not locked the OC variant of the GeForce RTX 5080 Slim, so there is some leeway for overclocking the graphics card. For overclocking, we now use the XTREME Tuner PLUS tool from KFA2 in conjunction with MSI Afterburner for all NVIDIA graphics cards, we now use KFA2’s XTREME Tuner PLUS tool in conjunction with MSI Afterburner, as this allows us to set higher VRAM clock speeds and increase the TGP. We can force the TGP limit to increase from 100% to 110%.

We achieved a clock speed of 203 MHz with GPU Boost and even 2,702 MHz for the graphics memory.

The card operated flawlessly and produced no errors or artifacts. The final GDDR7 clock speed was 2,212.9 MHz instead of the factory setting of 1,875 MHz. This increase in clock speeds and the TGP limit even revealed a clearly noticeable performance boost in the benchmarks.

PNY GeForce RTX 5080 Slim OC Photos

Packaging and Contents

The packaging design of the GeForce RTX 5080 Slim OC is very sleek. The box isn’t as huge as those of the competition and has a modern look. The very light-colored background provides good contrast with the black graphics card, making it stand out better. On the box, we find a note indicating that this is the Slim Edition, which features a dual-slot design. The box has no outer packaging and is designed so that it can be opened via a tab. -

Upon opening the box, the graphics card is immediately visible, wrapped in antistatic film. Overall, in our experience, the foam padding is sufficient for transport and shipping.

Inside the box itself, we find a quick start guide with some instructions, as well as the 12V-2x6 power adapter in plastic packaging. Nothing else is included in the box.

Test Setup

The Ryzen 9 7950X, which runs on faster DDR5-6000 RAM. The 2024 benchmarks omit Battlefield 4 and instead introduce Cyberpunk 2077 and Saints Row. We’ve further reduced 1080p testing and focused more on 4K, enabling ray tracing where appropriate. We’ve also added our first 8K benchmarks.

Another change is that we’re no longer testing inside a case; instead, we’ve returned to an open test setup. Graphics cards are becoming too large and bulky, which is why we opted for the open test bench.

The driver is from Nvidia and was version 610.62.

The Test System

• CPU: AMD Ryzen 7 7800X3D
• Motherboard: ASUS ROG CROSSHAIR X870E HERO

• M.2 SSD: Corsair MP600 Elite, WD_Blue SN580 1 TB

• RAM: G.SKILL Trident Z5 48 GB DDR5-6800 (34-46-46-108-1.35v)
• CPU Cooler: DeepCool LS720

• Power Supply: NZXT C1500 Platinum

• Fans: 0
• Case: CoolerMaster Benchtable (open)

GeForce RTX 5080 Ray Tracing Synthetic Benchmarks

For synthetic benchmarks with ray tracing support, we use only 3DMark Port Royal. The remaining tests are based on games that implement the technology in one way or another.

3DMark PORT ROYAL 1.2

Synthetic Benchmarks

Synthetic benchmarks are useful for an initial comparison of performance, but they do not reflect the cards’ actual gaming performance, as other factors come into play that are not immediately apparent.

3DMark Benchmark

Geekbench 6 OpenCL Benchmark

Unigine Superposition Benchmark

8k Benchmarks

For our first 8k test in the benchmark suite, we use the Unigine Superposition run with the 8k option. This test can also be run on non-8k monitors.

GeForce RTX 5080 Ray Tracing Game Benchmarks

With new game titles and the overhaul of the test station, additional ray tracing titles have been added. This segment now provides a solid data foundation. The results are not commented on individually. Our comments are summarized in the rankings below.

Cyberpunk 2077 with Ray Tracing with/without DLSS/FSR

Fortnite with Ray Tracing

Black Myth Wukong Benchmark with Ray Tracing

Metro Exodus Enhanced Edition Benchmarks

[end of page]