Kingston FURY Renegade RGB DDR5 96GB Kit Review

Two high-performance DIMMs, 96GB. Any questions?

Late last year we were introduced to the idea of so-called “non-binary” memory (24Gb rather than 16Gb DRAM density), with 24GB and 48GB modules eventually making their way to the market in 2023. For their part, Intel offered initial platform support (BIOS updates began in February), and AMD followed a few months later (BIOS updates began in May).

If you are the type to keep up-to-date on BIOS revisions, or have simply purchased a DDR5 motherboard recently enough to have support already baked in, you too can outfit your system with memory capacities previously reserved for those who rather questionably mixed their modules.

While it is certainly possible to achieve a capacity of 96GB by simply installing three 32GB DIMMs, thus breaking the symmetry of memory channels and robbing oneself of bandwidth, there is a better way. We will explore this dual-DIMM 96GB kit from Kingston today – the maximum capacity available in the FURY Renegade RGB lineup – first seen as part of a high-end configuration from Falcon Northwest.

Features (via Kingston):

• Dynamic customizable RGB lighting effects
• Patented Kingston FURY Infrared Sync Technology
• Engineered to maximize performance
• Tap into extreme overclocking potential
• Intel XMP 3.0 Certified
• Qualified by the world’s leading motherboard manufacturers
• Aggressive aluminum heat spreader design

The Fury RENEGADE 96GB Kit

These are, as one may have noticed, DIMMs with aluminum heat spreaders and RGB lighting along the top edge. Nothing revolutionary in the design, but nice-looking (subjective, of course).

Beyond the hefty capacity is the speed of these DIMMs, which are rated for 6400 MT/s at a CAS latency of 32 (part number KF564C32RSAK2-96). This speed and latency is achieved by running the sticks at 1.4 volts, making it unsuitable for AMD systems – but have no fear, the more AMD-appropriate 6000 MT/s version is offered at 1.35V (KF560C32RSAK2-96).

Performance Testing

Memory benchmarking is about as useful as any number of other purely academic pursuits, but what sort of “review” would this be if I just claimed that it worked as advertised and showed some photos of the product? So, test results must be provided.

It may not surprise the astute reader to discover  that we will be using an Intel platform for our testing of this kit. Why? Well, running above 6000 MT/s on AMD, regardless of voltage, offers little to no advantage, depending on application. On Intel, on the other hand, the sky’s the limit.

The specs of the test system appear on the charts, and here I will clarify that our Intel Core i7-14700K was manually limited to the officially specified 253W Turbo and ICCMax 307A limits. No 4095W motherboard defaults here. (Read more about 14th Gen power in Gavin’s excellent review for AT.) Oh, and I forgot to indicate the GPU used, which was an NVIDIA GeForce RT4090 FE card.

First, a synthetic memory benchmark to show potential memory scaling:

As expected, running the AIDA64 v6.92.6600 Cache & Memory Benchmark, DDR5 clearly scales with speed in this test (latency, too, but we didn’t get that far into testing). But what does this mean for a more real-world workload? Thanks to PCMark 10 and its full system benchmark, we can test a variety of scenarios and compare results:

The differences are slight, but overall this 96GB DDR5-6200 kit had the highest scores in every category. Shaving cumulative seconds off of various workloads over time may not be compelling to everyone, but at least I can report that I had zero stability issues running this RAM over several days of testing and evaluating both the Kingston kit, and the Talon PC.

Conclusion

One area in which these kits really shine is rather obvious – they only contain two DIMMs. You may be rolling your eyes, but consider that until recently one had to run four DIMMs (or three, if you are uncivilized) to exceed 64GB of system memory, it’s kind of a big deal. Plus, when we are talking about overclocking, it is an unfortunate fact that not every platform can support the same memory speeds when populating the board with more than two DIMMs.

Rather than falling back to some slower speed, or relaxing memory timings, isn’t it easier to simply pop in two DIMMs, set the XMP profile, and be on your way? I prefer this lifestyle, myself. Life is too short to mess around with down-clocking gaming RAM due to platform limitations.

This 96GB kit may very well be a solid option for high-performance memory at this price level, but the fact is I can’t actually find this kit for sale in the usual places online, and Kingston’s online storefront doesn’t offer it yet, either. I did find a listing at PROVANTAGE, but with a $394 USD price tag (down from $447 list) from this seller, this particular speed/capacity mix will have limited appeal.

Anyhow, there may be a certain premium attached to a dual-DIMM kit of this particular speed and capacity, but I eagerly anticipate the availability of such kits considering how smooth and uneventful my experience with this 96GB SKU was. The 6000 MT/s SKU (KF560C32RSAK2-96) is probably more than sufficient, and should be more affordable than the 6400 MT/s SKU reviewed today.