AMD 3020e vs Intel Celeron N5095

AMD 3020e

The AMD 3020e (Codename Dali) is a mobile APU that integrates two Zen cores (without SMT, therefore 2 threads) clocked at 1.2 to 2.6 GHz. The TDP is specified at 6 Watt and therefore the chip can be passively cooled. The integrated Radeon graphics card offers 3 CUs at up to 1,000 MHz (Radeon RX Vega 3). The dual channel memory controller supports DDR4-2400.

More information on Raven Ridge product family can be found in our launch article.

Performance

The average 3020e in our database sits close to the Celeron N4100 and the Core m3-8100Y, as far as multi-thread benchmark scores are concerned. Neither of the three is a performance monster, their fairly low power consumption figures being their main advantage.

Much like the two Intel chips, the 3020e will let you binge-watch Netflix and type e-mails but anything tougher than that will bring the APU to its limits quickly.

Your mileage may vary depending on how high the CPU power limits are.

Power consumption

This little APU has a default TDP (also known as the long-term power limit) of 6 W. Which is low and thus good enough for tablets, laptops and mini-PCs featuring no active cooling solution.

The AMD 3020e is built with a 14 nm process leading to low, as of early 2023, energy efficiency.

Intel Celeron N5095

The Celeron N5095 is an inexpensive quad-core SoC of the Jasper Lake product family designed for use in affordable SFF desktops and laptops. It features four Tremont CPU cores running at 2 GHz that Boost to up to 2.9 GHz with no thread-doubling Hyper-Threading technology in sight. A pretty basic iGPU is present as well.

The only difference between the N5095 and the N5095A is that the latter comes with support for more proprietary Intel technologies such as the Smart Sound DSP, Wake on Voice and HD Audio.

Architecture and Features

Tremont brings many improvements over Goldmont Plus, the architecture that we know from the N5030 and myriads of other N-class CPUs. An up to 30% boost in single-thread performance is to be expected thanks to smarter prefetchers, branch prediction improvements and other refinements, according to Intel. These new chips are physically larger than their immediate predecessors as a result. Either way, this is still a "small" core rather than a "big" one according to ChipsAndCheese.

The Celeron has 1.5 MB of L2 and 4 MB of L3 cache and is compatible with DDR4-2933 and LPDDR4x-2933 memory or slower. Support for Intel CNVi Wi-Fi 6 modules is baked into the chip, as are 8 PCIe 3.0 lanes for NVMe SSD speeds up to 3.9 GB/s. USB 4 or Thunderbolt aren't supported however.

Please also note that the Celeron gets soldered to the motherboard (BGA1338 socket interface) for good and is thus not user-replaceable.

Performance

The average N5095 in our database is about as fast as the Core i3-10110U, Core i3-1005G1, Celeron N5105 and also the Ryzen 3 3200U in multi-threaded workloads. Which is just enough for the most basic of tasks in late 2024.

Performance will get a significant hit if the power target is set to 10 W or 6 W instead of the Intel-recommended 15 W value.

Graphics

The DirectX 12.1-capable 16 EU UHD Graphics runs at up to 750 MHz and is in many respects similar to what Ice Lake CPUs come equipped with. This graphics adapter is capable of driving up to 3 SUHD displays simultaneously; HEVC, AVC, VP9, MPEG-2 and other popular video codecs can all be hardware-decoded. AV1 and VVC can't.

As far as gaming is concerned, it is reasonable to expect playable framerates in really old games (like Dota 2 Reborn) provided one sticks to lower resolutions such as HD 720p.

Power consumption

While most N-class chips have a 6 W long-term power target, the Celeron N5095 has a 15 W TDP to mimic much faster U-class Core processors. This isn't a great CPU for passively cooled designs.

The N5095 is built with the same 10 nm Intel process as Ice Lake-U processors for pretty unimpressive power efficiency, as of late 2024.