Any links to online stores should be assumed to be affiliates. The company or PR agency provides all or most review samples. They have no control over my content, and I provide my honest opinion.

Kingston was a bit slower to the market with their PCIe 4.0 NVMe drives, but late last year they launched two new drive models. The Kingston KC3000, which is aimed towards work and professional computing, and the Kingston Fury is geared towards gaming.

The main difference between these two new models is the Kingston Fury has around 5% faster read speeds and a significant 25% increase in the endurance rating. The 2TB Kingston KC3000 I am reviewing today is rated at 1.6PBW, which is already up there with the best, but the Fury is rated at 2.0PBW.

Sony Playstation 5 Comptability

Kingston advertises the Fury as being PS5 compatible, however, both drives meet the specification required for the PS5. The recommended drive read speed is just 5500MB/s, and most/all PCIe 4.0 drives should achieve this.

I assume the high endurance of the Kingston Fury is so that gamers can install and delete large games without worry.

However, neither drive comes with a built-in heat sink, and this is a recommendation for any drive used with the PS5. They are cheap to buy with plenty of compatible heatsinks on Amazon.

Kingston KC3000 vs Fury NVMe vs KC2500 Specification

KingstonFURY RenegadeKC3000
Form factorM.2 2280M.2 2280
InterfacePCIe 4.0 x4 NVMePCIe 4.0 x4 NVMe
Capacities2500GB, 1TB, 2TB, 4TB512GB, 1024GB, 2048GB, 4096GB
ControllerPhison E18Phison E18
NAND3D TLC3D TLC
Sequential read/write1500GB – 7,300/3,900MB/s
1TB – 7,300/6,000MB/s
2TB – 7,300/7,000MB/s
4TB – 7,300/7,000MB/s
512GB – 7,000/3,900MB/s
1024GB – 7,000/6,000MB/s
2048GB – 7,000/7,000MB/s
4096GB – 7,000/7,000MB/s
Random 4K read/write1500GB – up to 450,000/900,000 IOPS
1TB – up to 900,000/1,000,000 IOPS
2TB – up to 1,000,000/1,000,000 IOPS
4TB – up to 1,000,000/1,000,000 IOPS
512GB – up to 450,000/900,000 IOPS
1024GB – up to 900,000/1,000,000 IOPS
2048GB – up to 1,000,000/1,000,000 IOPS
4096GB – up to 1,000,000/1,000,000 IOPS
Total Bytes Written (TBW)3500GB – 500TBW
1TB – 1.0PBW
2TB – 2.0PBW
4TB – 4.0PBW
512GB – 400TBW
1024GB – 800TBW
2048GB – 1.6PBW
4096GB – 3.2PBW
Power consumption500GB – 5mW idle / 0.34W avg / 2.7W (MAX) read / 4.1W (MAX) write
1TB – 5mW idle / 0.33W avg / 2.8W (MAX) read / 6.3W (MAX) write
2TB – 5mW idle / 0.36W avg / 2.8W (MAX) read / 9.9W (MAX) write
4TB – 5mW idle / 0.36W avg / 2.7W (MAX) read / 10.2W (MAX) write
512GB – 5mW idle / 0.34W avg / 2.7W (MAX) read / 4.1W (MAX) write
1024GB – 5mW idle / 0.33W avg / 2.8W (MAX) read / 6.3W (MAX) write
2048GB – 5mW idle / 0.36W avg / 2.8W (MAX) read / 9.9W (MAX) write
4096GB – 5mW idle / 0.36W avg / 2.7W (MAX) read / 10.2W (MAX) write
Storage temperature-40°C~85°C-40°C~85°C
Operating Tempetemperaturerature0°C~70°C0°C~70°C
DimensionsHeat spreader:
80mm x 22mm x 2.21mm (500GB-1TB)
80mm x 22mm x 3.5mm (2TB-4TB)
Heatsink:
80mm x 23.67mm x 10.5mm
80mm x 22mm x 2.21mm (512GB-1024GB)
80mm x 22mm x 3.5mm (2048GB-4096GB)
WeightHeat spreader: 2TB-4TB – 9.7g
Heatsink:
500GB-1TB – 32.1g
2TB-4TB – 34.9g
512GB-1024GB – 7g
2048GB-4096GB – 9.7g
Vibration operating2.17G peak (7-800Hz)2.17G peak (7-800Hz)
Vibration non-operating20G peak (20-1000Hz)20G Peak (20-1000Hz)
MTBF1,800,000 hours1,800,000 hours
Warranty/Support4Limited 5-year warranty with free technical supportLimited 5-year warranty with free technical support

Kingston KC3000 vs KC2500 Performance / Benchmarks

My current system consists of:

  • AMD Ryzen 7 5800X
  • MSI MEG X570 UNIFY
  • NVIDIA GeForce RTX 3080
  • 32GB Kingston Fury Renagade 3600Mhz RAM

The Kingston KC3000 is now being used as my main Windows drive, and I was previously using the excellent PCIe 3.0 based Kingston KC2500.

Thermal Throttling

When I initially installed the drive it was without a heatsink and I didn’t really think anything of it. However, some of the benchmarks I ran generated unusual results. The results were still great but there would be the occasional drop off in performance. ATTO Disk Benchmark was particularly odd, towards the end of the benchmark when it is writing the larger files, speeds dropped off. With me using this as my primary Windows drive, my initial assumption was that it was other applications getting in the way of the read/writes. These issues persisted no matter what I did and I eventually came to the conclusion of thermal throttling.

As soon as I fitted a heatsink, the benchmarks generated the results I had expected.

Heatsinks for NVMe drives have been a thing for years now many motherboards include them as part of the design. In the past, I haven’t noticed that much difference in performance when not using one, but it would seem the increased performance of PCIe 4.0 drives means increased heat.

Benchmarks put the drive under more strain than most real-life scenarios so throttling may not be a significant issue for most workloads, but considering you can pick up a heatsink for, under a tenner, I’d say they are an essential purchase. This applies to all NVMe drives, not just Kingston and it is the reason why the PS5 also requires a heatsink.

Crystal Disk Mark

As you’d expect, there is a significant improvement with all of the benchmark results across the board. Sequential speeds are the big improvements with reading and write speeds more than doubling for Q8T1 and Q32T16.

Even looking at the random 4k Q1T1 results, you are getting a 30% and 39% improvement for read and write, respectively.

ATTO Disk Benchmark

ATTO reflects the Crystal Disk Mark results and also shows significant improvements with the IO/s. In the above screenshots, you can see the drive starts to significantly throttle performance at the end of the test.

The below images shows the input-output performance is roughly double for most of the files sizes.

Anvil Storage Utilities

Again, the KC3000 almost doubles the scores of the KC2500. Here we see significant improvements in IOPs and response time.

Prior to adding the heatsink, I had to rerun the write test twice before I got the expected results

Final Fantasy XIV: Shadowbringers Benchmark

With the Kingston KC3000 achieving such high benchmark scores compared to the KC2500, how does this affect gaming?

The Final Fantasy XIV: Shadowbringers is still a benchmark, but it should help provide some insight into how the Kingston KC3000 can improve game load speeds.

For this test, I run the benchmark in its default settings which is high and 1080p.

In this scenario, there is a 36.8% improvement in load times between each scene.

Kingston KC3000

Loading Times by Scene

  • Scene #1: 1.736 sec
  • Scene #2: 2.063 sec
  • Scene #3: 3.258 sec
  • Scene #4: 1.301 sec
  • Scene #5: 0.612 sec
  • Total Loading Time: 8.97 sec

Kingston KC2500 NVMe

Total benchmark Score: 12829

Loading Times by Scene

  • Scene #1: 2.146 sec
  • Scene #2: 3.509 sec
  • Scene #3: 4.172 sec
  • Scene #4: 2.887 sec
  • Scene #5: 1.509 sec
  • Total Loading Time: 14.223 sec

3DMark Storage Benchmark

3DMark has a new SSD benchmarking utility which simulates gaming-related activities to measure real-world gaming performance. This includes tests such as

  • Loading Battlefield V from launch to the main menu.
  • Loading Call of Duty: Black Ops 4 from launch to the main menu.
  • Loading Overwatch from launch to the main menu.
  • Recording a 1080p gameplay video at 60 FPS with OBS (Open Broadcaster Software) while playing Overwatch.

For comparison, when the benchmark launched, 3DMark posted on Steam the median scores for several 1TB drives:

DeviceMedian score
WD_BLACK SN850 NVMe 1 TB3426
Samsung SSD 980 PRO 1 TB2817
Samsung SSD 980 1 TB2434
Sabrent Rocket 4.0 1 TB2315
Samsung SSD 970 EVO Plus 1 TB2189
Samsung SSD 970 PRO 1 TB2096
Corsair Force Series Gen.4 PCIe MP6002062
WD_BLACK SN750 NVMe 1 TB2056
WD Blue SN550 NVMe 1 TB1992
Samsung SSD 970 EVO 1 TB1908

I should point out that many users using similar PCIe 4.0 drives get similar results to mine and some users getting over 4K. Again, this is quite likely the difference in cooling.

Price and Alternative Options

At the time of writing, the Kingston KC3000 is priced at:

Amazon

  • 512GB: £114.26
  • 1024GB: £152.83
  • 2048GB: £355.44

Insight:

  • 512GB: ££96
  • 1024GB: £156
  • 2048GB: £312

Kingston Fury

  • 512GB: £102
  • 1024GB: £164
  • 2048GB: £307

Samsung 980 PRO

Lower write speeds and 25% lower endurance

  • 512GB: £88
  • 1024GB: £160
  • 2048GB: £335.97

WD_BLACK SN850

Lower write speeds and 25% lower endurance

  • 512GB: £85
  • 1024GB: £140
  • 2048GB: £250 to £350 when available

As always for storage, the price varies a lot and frequently. The Kinston KC3000 isn’t priced that attractively on Amazon, but it has only been on there for a couple of months, while competing models have been available for much longer.

Overall

The Kingston KC3000 PCIe 4.0 is an outstanding drive that provides a significant improvement in performance from the PCIe 3.0 based KC2500.

Compared to other PCIe 4.0 drives, it has one of the best specs on the market. It has a significantly higher write speed and endurance than the Western Digital and Samsung equivalents. It probably won’t make that much difference in the real world, but the Kingston is priced attractively enough that it should be a serious consideration over the other brands.

Kingston KC3000 PCIe 4.0 NVMe Review Rating

Summary

The Kingston KC3000 has incredible performance and is a significant upgrade from the previous KC2500. It has a significantly higher endurance rating than many other competing drives and this makes it ideal for demanding work environments. 

Overall
90%
90%
  • Overall - 90%
    90%

Pros

  • Incredible speeds with double the performance in many benchmarks vs previous generation
  • One of the highest endurance ratings on the market 

Cons

  • Like all PCIe 4.0 drives, it runs hot and you should factor in the cost of a heatsink if you don’t have one included with your motherboard.

Last update on 2024-04-25 / Affiliate links / Images from Amazon Product Advertising API

Similar Posts

Leave a Reply

Your email address will not be published. Required fields are marked *