Sapphire RX 7700 XT Pure Review: Pretty but Expensive

All of the performance testing on the previous pages tends to be rather unimportant if you're trying to decide if the Sapphire RX 7700 XT Pure is the right card for you. Performance will be nearly the same as any other 7700 XT card, and it's really the other factors like temperature and noise levels that are most noticeable — after aesthetics, perhaps.

Our test PC uses an Nvidia PCAT v2 device to capture in-line power use, as well as GPU clocks and temperatures. The charts below are the geometric mean across all 15 games, though we also have full tables showing the individual results further down the page.

We have separate charts for 1440p ultra and 1080p ultra, since those are the settings we used for testing the Sapphire RX 7700 XT Pure. Besides the power, thermal, and clock speed data, we also check noise levels using an SPL meter. Let's start with power.

Image 1 of 2

It's quite interesting that the Sapphire RX 7700 XT Pure ends up using 12W less power across our test suite compared to the XFX RX 7700 XT QICK 319. Some of that might come from differences in the fans, but it's probably more a case of voltage regulators and tuning. That's not a huge difference, but Sapphire definitely comes out ahead.

And of course we have to put things in perspective. Generally, it's fair to say that the RX 7700 XT outperforms the RTX 4060 Ti by a decent margin, particularly if you're not using ray tracing. Equally true, however, is that Nvidia's GPUs are significantly more efficient. The RTX 4070, which was universally faster than the RX 7700 XT, uses 30W less power. It also costs $100 extra, but that's more a case of Nvidia pricing rather than actual cost of the hardware, we suspect.

It's also interesting that the RX 6800, which offers very similar performance in most cases, lands between the two 7700 XT cards in power use. It gave up a decent amount of compute compared to the 6800 XT, but the drop in compute and clock speeds had a relatively large impact on power use.

Image 1 of 2

Officially, the boost clock on the Sapphire RX 7700 XT Pure is 15 MHz lower than the XFX card's boost clock. In practice, there's almost no difference. Across our test suite, the Sapphire card averaged 2528 MHz at 1440p and 2531 MHz at 1080p, which is 1–2 MHz lower than the XFX card. That also explains why performance ends up basically identical.

You can of course try to overclock and tune your particular GPU to squeeze out some extra performance. Straight overclocking with RDNA 3 GPUs doesn't always yield much of an improvement, but if you can successfully undervolt the GPU (i.e. while maintaining a stable system), that can often result in larger gains than trying to increase the GPU and VRAM clocks alone.

Image 1 of 2

The XFX 7700 XT was already an exceptionally cool running GPU, by modern metrics. Sapphire's 7700 XT Pure doesn't beat the XFX card, but it's within 1 degrees C in our 1440p test suite, and only 0.2C warmer at 1080p. Temperatures are also impacted by the fan speed curve, so looking just at temperatures isn't super useful, especially when cards are sitting below 60C in most cases.

Note also that Nvidia's GPUs tend to land around the 60C mark as well, while AMD's higher tier RX 7800 and 7900 GPUs push temperatures up. The RX 6800 also ran a bit hotter, though all of the cards are still relatively cool compared to most of the previous two generations when looking at the fastest parts.

Image 1 of 2

As we've seen with other recent Sapphire card reviews, the angled fan blades prove very effective at cooling, while keeping noise levels (and RPMs) down. The XFX card had a 23% fan speed in our gaming test, while Sapphire's card had a 24% fan speed. However, the Sapphire card was whisper quiet at just 37 dB(A) — not that the XFX card was loud at 40 dB(A), but it was certainly louder.

The noise level in our room measures 31–32 dB(A) during testing, and that's at the limit of our SPL meter. (We've been in "quieter" environments and it still bottoms out at around 31 dB(A).) We're also using an older PC, a Core i9-9900K, for our noise tests, simply because the CPU cooling on that system is as quiet as we can get. 12900K and 13900K PCs inevitably push higher CPU fan speeds, which makes the overall noise levels higher. We do position the SPL meter just 10cm from the GPU fans as well, and if you're sitting a few feet away the noise from the fans will be quieter.

We also test with a static fan speed of 75%, though mostly that's just to see how loud a card might get in a worst-case scenario. Normally, you should never experience fan speeds anywhere near that level (especially now that Ethereum mining is dead). Sapphire's card bumps to 60 dB(A) in this case, which is still behind the XFX card's 67 dB(A), but you'd need to be in a pretty hot environment to get anywhere near that level.

Image 1 of 12

The above gallery gives the full rundown of all of our testing, including performance per watt and performance per dollar metrics. The prices are based on the best retail price we can find for a new card, at the time of writing. These prices fluctuate a lot, and for the above tables we used a price of $469 for the Sapphire 7700 XT.

Based on the current retail prices, the best GPU value out of the cards tested goes to the RTX 4060, followed by the RTX 4060 Ti, RX 6800, and then the RX 7800 XT. The Sapphire RX 7700 XT (as well as the reference priced cards) land a bit further back, illustrating once again that of the new Navi 32 GPUs, the RX 7800 XT delivers a better overall value.

Nvidia's 40-series GPUs are all more efficient in performance per watt, and the RX 7700 XT (Sapphire or XFX vards) trail the 7800 XT and 7900 XT. At 1440p, the Nvidia 40-series GPUs deliver 0.35–0.39 FPS/W, while AMD's 7000-series GPUs deliver 0.23–0.28 FPS/W. Not that efficiency is the most important metric, but it does show how things haven't really progressed as much for AMD's RDNA3 chips.

• MORE: Best Graphics Cards
• MORE: GPU Benchmarks and Hierarchy
• MORE: All Graphics Content