Intel Core i5-5350U vs Intel Celeron J4105 vs Intel Celeron N4000

Intel Core i5-5350U

The Intel Core i5-5350U is an ULV (ultra low voltage) dual-core processor based on the Broadwell architecture, which has been launched in January 2015. In addition to two CPU cores with Hyper-Threading clocked at 1.8 - 2.9 GHz (2 Cores: 2.7 GHz), the chip also integrates an HD Graphics 6000 GPU and a dual-channel LPDDR3-1866/DDR3L-1600 memory controller. The Core i5 is manufactured in a 14 nm process with FinFET transistors.

Compared to the Core i5-5300U, the i5-5350U features a lower CPU base clock, but integrates a somewhat faster GPU.

Architecture

Broadwell represents the "Tick" in Intel's Tick-Tock model, which means a shrink of its predecessor Haswell. The new 14 nm manufacturing process with three-dimensional FinFET transistors allows not only improvements in energy efficiency, but also a significantly smaller die size fit for more compact packages and devices.

Further advantages result from the revised microarchitecture of Broadwell. Due to its improved branch prediction, bigger buffer sizes (1500 instead of 1000 entries in the L2 TLB) and other tweaks, the performance per clock has been increased by more than 5 percent over its predecessor. There are also some new instruction set extensions designed for cryptographic applications.

Performance

Thanks to its improved architecture, the Core i5-5350U ends up 5 - 15 percent faster than the slightly lower clocked Core i5-4350U and matches the Core i7-4510U (both Haswell). Thus, the CPU has sufficient power for office and multimedia purposes as well as more demanding applications and multitasking.

Graphics

The integrated HD Graphics 6000 offers 48 Execution Units (EUs) clocked at 300 - 1000 MHz. Similar to the CPU core, the GPU architecture (Intel Gen 8) has been thoroughly revised for improved performance-per-clock. In addition, the GPU benefits from faster memory speeds up to LPDDR3-1866 (Haswell: DDR3-1600). We expect the HD Graphics 6000 to be significantly faster than the former HD Graphics 5000 (40 EUs), part of the Haswell generation. Nevertheless, demanding games as of 2015 should be playable only in low or medium settings.

Broadwell is the first Intel chip to fully support DirectX 11.2 as well as OpenCL 1.3/2.0 and OpenGL 4.3. Video will output natively via DisplayPort 1.2 or HDMI 1.4a. However, the new HDMI 2.0 standard is not officially supported.

Power Consumption

The i5-5350U is rated at a TDP of 15 W (9.5 W cTDP Down) including the graphics card, memory controller, VRMs and the integrated chipset. Therefore, the CPU is suited even for small ultrabooks 11-inches and greater.

Intel Celeron J4105

The Intel Celeron J4105 is a quad-core desktop SoC primarily for mini PCs and mini ITX mainboards. It runs at 1.5 - 2.5 GHz (Single Core Burst) and is based on the Gemini Lake platform. Similar to the Apollo Lake predecessor, the chip is manufactured in a 14 nm process with FinFETs but offers slightly improved processor cores, double the amount of L2 cache, a smaller package, a new generation of monitor outputs (Gen 10) and a partly integrated WiFi chip. Besides four CPU cores, the chip also includes a DirectX 12 capable GPU as well as a DDR4/LPDDR4 memory controller (dual-channel, up to 2400 MHz). The SoC is not replaceable as it is directly soldered to the mainboard.

Architecture

The processor architecture was slightly reworked and is now called Goldmont Plus. It features an increased level 2 cache (to 4 MB). That means the per-clock-performance should be a bit better, but not near the Core CPUs like Kaby Lake Y.

Performance

While we have not tested a single system built around the J4105 as of August 2023, we have tested a certain number of laptops featuring the Celeron N4120, a chip with very similar specs (4 cores, 4 threads, up to 2.6 GHz). Based on that, expect the J4105 to be very slow in all but the most basic activities.

Intel Celeron N4000

The Intel Celeron N4000 is a dual-core SoC designed for affordable laptops and mini-PCs and was announced late 2017. It runs at 1.1 GHz to 2.6 GHz (Single Core Burst, Multi Core Burst maxes out at 2.5 GHz) and is based on the Gemini Lake platform. Similar to the Apollo Lake predecessor, the chip is manufactured on a 14 nm process with FinFETs but offers slightly improved processor cores, double the amount of L2 cache, all in a smaller package. Partial Wi-Fi 5 support is baked into the chip. Besides two CPU cores, the chip also includes a DirectX 12 capable GPU as well as a DDR4/LPDDR4 memory controller (dual-channel, up to 2400 MHz). The SoC is not replaceable as it is directly soldered to the mainboard.

Architecture

The processor architecture was slightly reworked and is now called Goldmont Plus. It features a larger L2 cache (4 MB). That means the per-clock-performance should be a bit better, but not anywhere near the Core CPUs like Kaby Lake Y.

Performance

The average N4000 in our database only just matches the Core i5-4300Y, a fairly old dual-core processor of nearly the same energy efficiency, in multi-thread performance. In other words, the Celeron is a very slow processor that's good enough for the most basic of tasks only. Expect long load times no matter the app.

Graphics

The UHD Graphics 600 (Gemini Lake) is based on Intel's Generation 9 architecture, which supports DirectX 12 and is also used for the Kaby Lake / Skylake / Apollo Lake graphics adapters (like HD Graphics 520). Equipped with 12 EUs and a clock of up to 650 MHz, the performance should be roughly on par with the older HD Graphics 500 (Apollo Lake).

The chip also includes an advanced video engine with hardware support for the playback of VP9 and H.265 (8-bit color-depth).

Power consumption

Like most other N-class Intel chips, the Celeron N4000 has a 6 W TDP (also known as the long-term power limit). This is not much at all, allowing laptop makers to build passively cooled laptops, tablets, mini-PCs around the chip. The SoC can briefly consume up to 15 W, though, depending on how a system is configured.

Last but not the least, this Celeron is manufactured on one of the old 14 nm Intel processes leading to poor, as of early 2023, energy efficiency.