Intel Celeron N5095 vs Intel Core i5-8365U vs Intel Pentium Silver N6005

Intel Celeron N5095

The Intel Celeron N5095 is a quad-core SoC of the Jasper Lake product family designed for use in affordable SFF desktops and laptops. The Celeron was announced in early 2021. It features four Tremont CPU cores running at 2 GHz (base clock speed) Boosting up to 2.9 GHz (single-core Boost) with 1.5 MB of L2 and 4 MB of L3 cache but no thread-doubling Hyper-Threading technology. This processor is manufactured on Intel's first-gen 10 nm process to be not unlike the Ice Lake-U Core-series processors. The faster Celeron N5105 is notable for being more energy efficient (10 W TDP vs N5095's 15 W) yet featuring a faster iGPU model (24 EU UHD Graphics versus 16 EU UHD Graphics, with a higher clock rate to boot).

Architecture

Tremont architecture brings many improvements over the outgoing Goldmont architecture we know from Pentium N5030 and so many other N-class CPUs. According to Intel, the new architecture brings a 30% boost in single thread performance (+10% to +80% in SPECint and SPECfp depending on the test).

While the older N-class Celerons and Pentiums had to be content with up to 8 GB of DDR4-2400 RAM, Celeron N5095 is officially compatible with up to 16 GB of dual-channel DDR4-2933 or quad-channel LPDDR4x-2933 memory. Wi-Fi 6 (Gig+) support is partially baked into the chip. The Celeron supports up to 8 PCI-Express 3.0 lanes, up to 14 USB 3.2 ports and up to two SATA III ports. Four PCI-Express 3.0 lanes allow for read/write rates of up to 3.9 GB/s, provided a suitably fast NVMe SSD is used.

Jasper Lake processor package is larger compared to what Gemini Lake models had (35 x 24 mm vs. 25 x 24 mm). Please note that the Celeron gets soldered on to the motherboard (BGA1338 socket interface) and is thus not user-replaceable.

Performance

The average N5095 in our database matches the Core i3-10110U, Core i3-1005G1, Celeron N5105 and also the Ryzen 3 3200U, as far as multi-thread benchmark scores are concerned.

Even if the cooling system is great and the Power Limits are high as the sky, some slight architectural improvements together with the updated manufacturing node and higher-than-usual wattage of the Celeron are not nearly enough to let it compete with modern Core and Ryzen-series processors. It's a good option for basic day-to-day tasks, but anything beyond that will be a struggle.

Graphics

The CPU features the DirectX 12-capable 16 EU UHD Graphics iGPU running at 450 MHz to 750 MHz. The graphics adapter is capable of driving up to 3 displays with resolutions up to 4096x2160@60; it will happily decode HEVC, AVC, VP9, MPEG-2 and other popular video codecs. The latest AV1 codec is not supported, though. (You can still play such a video but it will be software-decoded rather than hardware-decoded which puts a hard limit on the watchable resolution as the CPU cores are not very fast here.)

This is not a gaming GPU by any stretch of imagination. It is just as fast as an HD Graphics 515; it will let you play the least demanding titles (like Dota 2 Reborn) provided you are OK with 720p resolution and lowest quality preset possible.

Power consumption

Most N-class Celerons and Pentiums have a 6 W TDP (also known as the long-term Power Limit). The Celeron N5095 on the other hand has a 15 Watt TDP to mimic much faster U-class Core i3/i5/i7 processors. This basically means that passively cooled designs are out of the question with this chip.

This Intel processor is manufactured on Intel's first-gen or second-gen 10 nm process [no precise data available] for OK, as of late 2022, energy efficiency.

Intel Core i5-8365U

The Intel Core i5-8365U is a power efficient quad-core SoC for notebooks and Ultrabooks based on the Whiskey Lake generation and was announced in April 2019 as part of two vPro capable CPUs of the WHL-U series. Compared to the similar named Kaby Lake-R processors (e.g. Core i5-8250U), the Whiskey Lake CPUs are now produced in a further improved 14nm process (14nm++) and offer higher clock speeds. The architecture and features are the same. The i5-8365U offers e.g. high Turbo clock speeds of 4.1 GHz (versus 3.4 GHz of the i5-8250U) for a single core. The integrated GPU is still named Intel UHD Graphics 620 and the dual-channel memory controller still supports the same RAM speeds as Kaby-Lake-R (DDR4-2400 / LPDDR3-2133). Thermal Velocity Boost is not supported (only in the Core i7-8565U).

Compared to the predecessor, the Core i5-8265U, the i5-8365U offers a 200 MHz higher Turbo Boost. As the i5-8365U first appeared in a Thinkpad, the CPU most likely supports vPRO for additional hardware management features.

The Whiskey Lake SoCs are used with a new PCH produced in 14nm that supports USB 3.1 Gen 2 (10 Gbps) and CNVi WiFi/BT parts.

Architecture

Intel basically uses the same micro architecture compared to Skylake and Kaby Lake, so the per-MHz performance does not differ. That means Whiskey Lake is a Kaby Lake chip manufactured in the improved 14nm++ process.

Performance

The performance of the i5-8365U depends on the cooling solution of the laptop and the defined TDP limits for short and long term performance. We already saw big differences for Kaby Lake-R (e.g., i7-8550U benchmarks), especially for long term (sustained) performance. Therefore, it will be interesting to see how the additionalTurbo  clock speed can be made use of. Compared to the Core i5-8265U (see for benchmarks), the 200 MHz higher Turbo Boost should result in some gains in single core benchmarks.

Graphics

The integrated Intel UHD Graphics 620 (Intel Gen 9.5) is untouched from the 8th Gen Kaby Lake chips. With fast dual-channel memory it can reach the performance of a dedicated GeForce 920M.

Contrary to Skylake, Kaby Lake and Whiskey Lake now also supports H.265/HEVC Main 10 with a 10-bit color depth as well as Google's VP9 codec. The dual-core Kaby Lake processors announced in January should also support HDCP 2.2.

Power Consumption

The chip is manufactured in a further improved 14nm process with FinFET transistors (14nm++), the same as the 8th Gen Coffee Lake processors. Intel still specifies the TDP with 15 Watts, which is typical for ULV chips. Depending on the usage scenario, the TDP can vary between 7.5 (cTDP Down) and 25 Watts.

Intel Pentium Silver N6005

The Intel Pentium Silver N6005 is a quad-core SoC of the Jasper Lake series that is primarily intended for inexpensive and small desktop and was announced in early 2021. The four Tremont CPU cores clock between 2 and 3.3 GHz (single core Burst) and offer no HyperThreading (SMT). The N6005 uses 1.5 MB L2 and 4 MB L3 cache. The chip is manufactured in 10nm at Intel (most likely in the same process as Ice Lake).

CPU Architecture

The processor architecture is called Tremont and a complete redesign compared to the old Golmont Plus cores in the predecessor. According to Intel, the single thread performance of a core could be improved by 30% on average (10 - 80% in all tests of SPECint and SPECfp).

Features

In addition to the quad-core CPU block, the SoC integrates a 32 EU Intel UHD Graphics GPU clocked from 450 - 900 MHz and a LPDDR4(x) dual channel memory controller (up to 16 GB and 2933 MHz). The chip now also partly integrates Wi-Fi 6 (Gig+), 8 PCIe 3.0 lanes, 14 USB 2.0 / 3.2 ports and two SATA 6.0 ports. The package got bigger and measures 35 x 24 mm (compared to 25 x 24 mm for the N5030 e.g.). The SoC is directly soldered to the mainboard (BGA) and can't be easily replaced.

Performance

The average N6005 in our database is in the same league as the Core i3-1110G4, as far as multi-thread benchmark scores are concerned. This is enough to enjoy Web browsing and watch videos online without noticeable slowdowns, as of mid 2023 - which is what most consumers want.

Thanks to its decent cooling solution and a long-term CPU power limit of 15 W, the Intel NUC11ATKPE is among the fastest systems built around the N6005 that we know of. It can be about 50% faster in CPU-bound workloads than the slowest system featuring the same chip in our database, as of August 2023.

Power Consumption

Similar to the predecessor, Intel specifies the TDP with 10 Watts (versus the 6 W of the mobile N6000). The chip can therefore be cooled passively in theory, but SKUs with fans are possible as well.