To answer this question, you have to do your research, or you can read this GTX 970 vs RX 570 comparison instead. I’ve done all of the research, so you won’t have to. You’ll find all the information you need in this article to make an informed decision before purchasing either of these. And to answer the question, yeah, you should buy one in 2023. Keep reading our RX 570 vs GTX 970 comparison to find out which one.
GTX 970 vs RX 570 – Quick Comparison
|GeForce GTX 970||Specs||Radeon RX 570|
|GM204-200-A1||GPU||Polaris 20 XL|
|PCIe 3.0 x16||Interface||PCIe 3.0 x16|
|1,664 (CUDA Cores)||Cores||2,048 (Stream Processors)|
|1,050 MHz||Base Clock|
|1,178 MHz||Boost Clock|
|3.5 GB + 512 MB GDDR5||Memory||8 GB GDDR5|
4 GB GDDR5
(7 Gbps effective)
|Memory Speed||1,750 MHz|
(7 Gbps effective)
|196 GB/s + 28 GB/s||Bandwidth||224.0 GB/s|
|224-bit + 32-bit||Memory Bus||256-bit|
|300 W||Required PSU|
|64℃ (147.2℉)||Max Recorded Temp|
|47.5dB||Max Fan Noise|
1x HDMI 2.0
3x DisplayPort 1.4a
1x HDMI 2.0b
3x DisplayPort 1.4a
Nvidia GeForce GTX 970
Nvidia did an excellent job with this high-end graphics card. Even though it was released in 2014, its performance surpasses some of Nivida’s later-released cards from the GTX 10 and 16 series. However, it was the center of a U.S. class action lawsuit for false advertising. Its advertised 4 GB GDDR5 VRAM was really 3.5 GB. The remaining 0.5 GB was a much slower “spillover” section.
Despite these issues, the GTX 970 is a popular card that offers great performance, as many Nvidia cards do. Nvidia has produced many excellent GPUs since it developed the first-ever GPU in 1999. Nvidia GPUs now power AI and machine learning projects all over the world.
- Lower prices (currently)
- Better 1080p performance
- Better 1440p performance
- Better when overclocked
- Fewer cores
- Smaller VRAM
AMD Radeon RX 570
This mid-range card from AMD was released in 2017 as part of the RX 500 series based on the second-generation Polaris architecture. It was offered in two versions, with a 4 or 8 GB VRAM. Even though it outdoes the GTX 970 in specifications, it lacks where it’s most important — performance!
AMD was founded in Silicon Valley over 50 years ago. It continuously offers great CPUs, while its GPUs always seem a bit behind the competition, especially since it didn’t offer high-end cards for some time.
- More cores
- Higher clock speeds
- Larger VRAM
- More TMUs
- Requires a stronger PSU
- Under-average overclocking
RX 570 vs GTX 970 – Key Specifications
The Second-generation Maxwell introduced new technologies such as VR Direct, MFAA, and support for HDMI 2.0. It had its fair share of problems as well. Async computing was reported not to perform well on some Maxwell cards because they relied on a driver to forward asynchronous tasks to the correct units. This bred controversy but did not affect its DirectX 12 compliance.
Polaris, or GCN 4.0, had no trouble with async computing because it had the built-in hardware to support the technology. Even that was not an upgrade from GCN 3.0. Instead, AMD relied on advancements in manufacturing to achieve better results on the same hardware. This allowed for a boost in clock speeds, which, together with HBM2 (High Bandwidth Memory 2), offered better performance than its predecessor.
Winner: RX 570
Also Read: Differences between GTX and RTX Explained
Clock Speeds & Overclocking
Clock speeds are the operating frequency of a component. GPUs, CPUs, RAM, and VRAM all have them. Over a decade or so ago, most clock speeds were static. We now live in the age of dynamic clock speeds. Dynamic clock speeds fluctuate between “base” and “boost.”
The GTX 970 FE has a 1,050 MHz base clock speed and a 1,178 MHz boost clock speed. The RX 570 FE has a 1,168 MHz base clock and a 1,244 MHz boost clock. Take these speeds with a grain of salt because they will drop to 300 MHz when idle and jump above the official boost clock.
Back when all cards had static clock speeds, we all resorted to overclocking (OC) to achieve higher speeds and better performance. Both AMD and Nvidia have their own dedicated software for overclocking, but a lot of people (me included) choose MSI Afterburner.
The GTX 970 offers (18%) better results when overclocked, according to users. Third-party cards almost always come factory-overclocked. The GTX 970 variants offer higher clock speeds than the RX 570 variants, as seen in the list below, and offer even better performance than the FE version.
|Best GTX 970 Variants||Boost Clock||Best RX 570 Variants||Boost Clock|
|GALAX GTX 970 HOF||1,380 MHz||Sapphire NITRO+ RX 570||1,340 MHz|
|ZOTAC GTX 970 AMP! Extreme Core Edition||1,380 MHz||Sapphire NITRO+ RX 570 8 GB||1,340 MHz|
|EVGA GTX 970 FTW ACX 2.0||1,367 MHz||Dataland RX 570 X-Serial||1,320 MHz|
|EVGA GTX 970 FTW+ ACX 2.0+||1,367 MHz||PowerColor Red Devil RX 570 OC||1,320 MHz|
|ASUS DRAGON GTX 970 DirectCU II TOP||1,355 MHz||AREZ STRIX RX 570 GAMING OC||1,300 MHz|
Winner: GTX 970
Also Read: Undervolting your CPU and GPU How-To Guide
A GPU’s core configuration encompasses its main cores (CUDA or Stream), TMUs (Texture Mapping Units), ROPs (render output units), and RT and Tensor cores (when available). They used to be equal to one another. This changed in 2004 when these areas were decoupled for better transistor allocation, resulting in more “main” cores and fewer of the rest.
The GTX 970 has 1,664 CUDA cores, and the RX 570 has 2,048 Stream processors. The RX 570 uses the same GPU variant as the RX 580, but AMD locked 256 cores to achieve the targeted 2,048, still more than the GTX 970 has. So, how is it that more cores result in poorer performance? It’s down to how the cores are used.
Shaders are generic programming languages that use GPU cores to create many of the aspects critical to 3D rendering. DirectX is a prime example of a shading language. The table below lists the various shaders and their version in each card. The GTX 970 uses slightly newer versions of some shaders, resulting in better performance by the GTX 970.
|GTX 970||Shaders||RX 570|
|12 (12_1)||DirectX||12 (12_0)|
TMUs handle textures, another critical part of 3D rendering. Virtually every virtual surface is covered with some sort of texture. Generally, AMD cards offer more TMUs and have the upper hand when it comes to texture rendering. The GTX 970 has 104, while the RX 570 has 128. Just enough to give the latter a 58% advantage in texture rendering.
Winner: RX 570
VRAM & Memory Specs
The GTX 970 was advertised as a 4 GB graphics card. This turned out to be not entirely true, as users reported that only 3.5 GB was accessible. Nvidia later admitted that 3.5 GB was paired with a 224-bit BUS offering a 196 GB/s bandwidth. The remaining 512 MB was paired to a 32-bit BUS offering a much slower bandwidth — 28 GB/s. This led to a $30 refund for those who purchased the GTX 970.
Needless to say that this is not very comparable to the RX 570’s 4, let alone 8 GB VRAM, paired with a 256-bit BUS offering a bandwidth of 224 GB/s. Nvidia’s segmented VRAM was not usable synchronously. Both cards’ VRAMs operate at up to 1,750 MHz (7 Gbps effective).
Winner: RX 570
Also Read: RAM vs VRAM: What are the key differences?
Both cards are capable of running a multitude of titles with maxed-out settings, but the GTX 970 does it better by an average of 8%, according to both benchmark and user tests. 4K gaming is not much of an option using these cards, except in some less performance-heavy titles such as CS:GO, Minecraft, or LoL.
The GTX 970 has 8% better performance at 1080p.
The GTX 970 has 8% better performance at 1440p.
Certain titles work better on one card than the other, but the majority is in the GTX 970’s favor.
Winner: GTX 970
The FE versions of both cards support up to five screens with one DVI, one HDMI, and three DisplayPorts. Both cards also require additional PSU connections to draw power. The GTX 970 requires two 6-pin connectors. The RX 570 uses one 6-pin connector.
TDP refers to two things important to this comparison, Thermal Design Point and Thermal Design Power. The GTX 970 requires a 300 W PSU to draw the 148 W it needs. The RX 570 requires a 450 W to draw the similar 150 W it uses. That’s Thermal Design Power.
Thermal Design Point is how much temperature a subsystem is allowed to produce. Generally, AMD cards run hotter than Nvidia cards. It’s no surprise that the RX 570 maxed out at 74℃ (165.2°F) during testing, 10℃ higher than the GTX 970.
Winner: GTX 970
Pricing & Availability
The GTX 970 launched with a $329 price tag. This dropped lower after the lawsuit against Nvidia. Prices are now at an all-time low, and you can grab a few third-party variants on Amazon right now. The most expensive option I could find is this EVGA SSC variant, priced 39% below MSRP. The lowest-priced is an EVGA SC variant priced 65% below MSRP. This MSI variant is the middle ground priced 45% below.
The RX 570 is an entirely different ballpark. It launched at a much lower MSRP — $169. It now costs much closer to the GTX’s MSRP (on average). The cheapest option I found was this Sapphire Nitro+ variant (one of the top five), priced 17% above MSRP. The next best option is an MSI Gaming variant priced 136% above MSRP. The most audacious offer is for an XFX Black Edition priced 402% above MSRP!
*Prices and availability may differ when you read this article.
So, should you buy one of these in 2023? I recently purchased a three-year-old model, and I’m more than impressed. The GTX 970 rivals that card in many ways, let alone the older RX 570. So, as I said, my answer is, “yes, you should.” That is if you’re not looking for the best of the best for 4K gaming in the newest titles.
The only thing I see going for the AMD in this RX 570 vs GTX 970 comparison is the larger VRAM in the RX 570. Games eat up VRAM as you push settings toward the max, and having enough VRAM is crucial. Nevertheless, the RX 570 falls short in the only aspect that really matters here – performance. The GTX 970 beats it in almost every title, regardless of the RX 570’s larger VRAM.