7600 Techspot

[Brian]

Until just days before release, the Radeon RX 7600 was set to cost $300, and we had actually finished our review in which we thoroughly criticized it. Last week, AMD called us to inquire about our impressions of the RX 7600. We told them it was underwhelming and arguably, redundant. We advised them to lower the price to at least $250 or it wouldn't be worth releasing. We were certain that within a few months of release, we would see it priced at $250 or less.

We likely weren't the only reviewers to give them such feedback, prompting AMD to change the price at the last minute. This forced us to go back and make a few changes to our review. What makes matters a little frustrating though is that they only reduced the price by $30, down to $270, which still makes the 7600 a poorly priced product and very underwhelming.

AMD claims that the new $270 Radeon RX 7600 is designed for 1080p gameplay at high to ultra-high settings and that it provides an optimized 1080p gaming experience. We're going to challenge that claim. 8GB of VRAM is already suboptimal for some 1080p games today and certainly won't be sufficient a year or two from now, especially if the information coming out of game developers is anything to go by.

So, this is where we're at today: $270 GPUs still come with just 8GB of VRAM. But at least it's not a $400 GPU. What was Nvidia thinking with their pricing? As poor a value as the $400 RTX 4060 Ti is, the Radeon 7600 isn't significantly better. In terms of cost per frame, there's not a lot in it, but we'll delve into that later in the review.

The only saving grace for AMD here is the price, which isn't great, mind you, but it's better than $400. This is why we won't critique the RX 7600 nearly as harshly as we did the RTX 4060 Ti, which should afford the upcoming RTX 4060 some breathing room as well.

First up, we have Fortnite. Here, the Radeon RX 7600 managed an average of 90 fps at 1080p, which is comparable performance to several pre-existing products such as the RTX 3060, 2060 Super, and 6600 XT. This is a disappointing result. The RTX 3060 not only features more VRAM, but it also didn't cost significantly more when it was released over two years ago.

At 1440p, it managed to match the 6600 XT, but was still 8% slower than the 6650 XT and 17% slower than the RTX 3060. There isn't much to say about that, other than that it's very underwhelming and seems somewhat pointless. Nonetheless, let's move on in the hope that things might improve.

The margins do improve at 1440p, with the 7600 delivering 6700 XT-like performance, which is impressive. It's also approximately 40% faster than the RX 6600, which helps justify the pricing discrepancy, at least for this title.

The 7600 takes the lead over the 6650 XT at 1440p, but these results are somewhat skewed as the 7600 consistently had missing textures in this test. This is a common issue for all 8GB models, leading to inconsistent memory management and unreliable results.

At 1440p, the 8GB graphics cards fail completely, but the 7600 isn't really powerful enough for this resolution anyway. You should be okay with the high-quality preset, but unfortunately, you won't be able to max out textures (which don't impact fps performance) if you have enough VRAM.

So how much should it cost? We believe that the RTX 4060 Ti which was also released this week should have been priced no more than $300, and by that logic, AMD should not ask more than $230 for the Radeon 7600. We advise waiting and watching for inevitable price drops over the next few months.

Here's a quick look at how the AMD reference card performs after an hour of load in an enclosed case playing Hitman 3 at 1440p. The average GPU temperature peaked at 78C with a peak hotspot temperature of 90C. The fan reached speeds of 1600 RPM to achieve these temperatures. The core clock typically ran at 2560 MHz, while the memory operated at the expected 18 Gbps.

It's disappointing to conclude another Radeon review with a thumbs-down, primarily due to its price. Furthermore, it's almost guaranteed that this card will be available for $250 or less within a matter of months.

Regarding the 8GB VRAM discussion, it's our opinion that products priced over $250 should have at least 12GB of VRAM. While an increase in VRAM would need a different GPU configuration, this is something they'd have to account for when planning the lineup.

Texture quality significantly impacts visuals and is much more important than features like ray tracing. As long as there's enough VRAM, textures won't affect FPS performance. This means products like the Radeon RX 6800 might be limited to medium-quality settings in AAA titles released a year or two from now. However, you'll still be able to max out texture options in those games, resulting in stunning, photo-realistic textures.

Since AMD introduced Ryzen 7000 CPUs last September, it's been clear that the series is very impressive in terms of performance, however migrating to the new AM5 platform has proved costly and for that reason fewer people than expected have jumped on board.

The Cinebench R23 results confirm everything we just learned about the operating behavior of these CPUs in this same workload. The 7600 scored 14344 pts and enabling PBO boosted that score by a mere 1.5%, meanwhile out of the box the 7600 was just 4% slower than the 7600X.

The 7900 scored 25062 pts making it quite a bit slower than the 7900X, though that margin was largely closed down simply by enabling PBO, making what is essentially a more efficient version of the 7900X.

Looking at the single core performance we find that the 7600X is 5% faster than the 7600, the 7700X is just 3% faster than the 7700 and the 7900X is 4% faster than the 7900. So the margins are fairly similar, between 3-5% faster for the original X models and not a big difference. In most instances PBO does a good job of closing that small gap.

We'll just show you compression and decompression performance for 7-Zip, but in short the margins are similar to what we saw in Cinebench R23. That is to say the 7600 was typically 3% slower than the 7600X, and the same was also true for the 7700 vs 7700X, while the 7900 was up to 8% slower than the 7900X.

As mentioned before, we're just going to look at the graphs for 3 of the 10 games tested as the results are all very predictable. Starting with Hitman 3 at 1080p, we see that the 7600 is just 3 fps slower than the 7600X, the 7700 4 fps slower than the 7700X and the 7900 10 fps slower than the 7900X. It's a similar story at 1440p, so it's only the 7900X that offers a notable performance gain over its non-X part and even then we're only talking about a 4% increase.

These aren't big margins and for those who care about productivity performance will find these non-X CPUs to make a lot more sense, given they're significantly more efficient and operate at much lower temperatures.

But you can also see why AMD went about things the way they did. These non-X parts don't look nearly as impressive on a graph when compared to Intel 13th-gen range. The 7700X is just a few frames slower than the 13900K and finds itself in second place, the 7700 which is just a few percent slower is a mid-field contender in this graph, trailing the 13700K by a slim margin.

Taking a quick look at cooling performance using the supplied coolers, we'll start with the 7600 using the Wraith Stealth. When placed under an all-core workload in Cinebench for an hour, installed inside an ATX case with a room temperature of 21c, this combo peaked at 97c, which is very hot. That said, we know Zen 4 processors target 95c and are designed to run at that temperature according to AMD, and we saw no evidence of throttling under these conditions, so the 7600 performed as expected. For reference under the same test conditions the 7600 peaked at 78c with a 360mm AIO.

Next we have the 7700 with the Wraith Prism, which is a far more impressive cooler and the results here are far better than what we saw with the 7600/Stealth combo. With the Prism the 7700 peaked at 79c, which is well below the 95c TjMAX. That's a great result given the 360mm AIO saw a peak temperature of 73c.

We've always recommended you buy the non-X versions, and we've done this for a few reasons. First, of course, they're generally cheaper and deliver comparable performance. They're also more efficient and unlocked, so you can overclock them to achieve full performance, and they typically come with a box cooler, which you can sell on eBay if you don't need it. The Wraith Prism sells for around $30-40 which is crazy considering you can get a good quality tower style cooler for that price.

This was already true before the arrival of these non-X 65w parts as AMD had cut pricing aggressively over the past few months, but the new CPUs make it all official. There's a chance the original X-parts head back towards their respective MSRPs (and become more expensive), but at that point we'd simply ignore them and just go for the non-X 65w models instead.

AMD's new Radeon RX 7600 XT is basically an RX 7600 with a 16 GB VRAM buffer, but is that worth paying a premium for? Measuring the benefits of extra VRAM can be complex as we'll explore in this review.

Today, we're taking our first look at AMD's new Radeon RX 7600 XT, which is basically an RX 7600 with a 16 GB VRAM buffer. It probably could have, and perhaps should have, been named the 7600 16GB. Fortunately, it's not an inferior product sharing the same name. The bar is low, we know, but over the years, we've witnessed some dubious practices.

For now, we're eager to discover what the 16GB 7600 has to offer. As before, the 7600 XT is essentially a 7600 with double the VRAM, with the only other alteration being a minor 4% increase in the core clocks, from 2,655 MHz to 2,755 MHz. This adjustment is something that could be manually applied to any 7600, so it's hardly a unique feature of the 7600 XT.

Potential buyers are likely wondering whether the 7600 XT is worth the extra $60 over the standard model. Addressing this question is challenging, as measuring the benefits of 16 GB of VRAM can be complex for several reasons.

c80f0f1006