We are a participant in the Amazon Services LLC Associates Program, an affiliate advertising program designed to provide a means for us to earn fees by linking to Amazon.com and affiliated sites.
[nextpage title=”Introduction”]
Both AMD and Intel recently released new families of low cost, low TDP desktop CPUs. AMD launched the AM1 platform with Sempron and Athlon “Kabini” processors, while Intel released the “Bay Trail-D” Celeron and Pentium CPUs, recognizable by the use of the letter “J” on the model naming. Among the lowest-end models of each family are, respectively, the AMD Sempron 2650, and the Intel Celeron J1800. Let’s compare the performance of those CPUs and discover which one is the best buy in the low-end market segment.
Both CPUs are dual-core models, while most processors of each family are quad-core ones. Because of this, they are the most inexpensive desktop processors available on the market nowadays.
The AMD “Kabini” and Intel “Bay Trail-D” processors are SoC (System On a Chip) CPUs, which means they include in the chip the processing cores, the memory controller, the video controller, and also the chipset. This helps both platforms to keep their low cost and low TDP.
The Sempron 2650 is the simplest (and cheapest) desktop processor from AMD. It is a socketed CPU and uses the FS1b socket (also informally called AM1). It looks like other processors from AMD, but it is smaller and uses a different (and smaller) cooler. We will test it using the ASUS AM1M-A motherboard, which we analyzed recently.
Figures 1 and 2 unveil the Sempron 2650 CPU and its stock cooler.
Figure 1: the Sempron 2650 processor
Figure 2: the Sempron 2650 and its stock cooler
The Intel Celeron J1800 is one of the simplest desktop processors on the market. Differently from its contender from AMD, it comes soldered to the motherboard. The model we are testing comes on the ASRock D1800M motherboard, which we also analyzed recently.
Figure 3: the Celeron J1800 processor
Let’s compare the main specs of the reviewed CPUs in the next page.
[nextpage title=”The Reviewed CPUs”]
In the tables below, we compare the main features of the CPUs included in our review.
| CPU | Cores | HT | IGP | Internal Clock | Turbo Clock | Base Clock | Core | Tech. | TDP | Socket | Price |
| Sempron 2650 | 2 | No | Yes | 1.45 GHz | No | 100 MHz | Kabini | 28 nm | 25 W | AM1 | USD 40* |
| Celeron J1800 | 2 | No | Yes | 2.41 GHz | 2.58 GHz | 100 MHz | Bay Trail-D | 22 nm | 10 W | FCBGA1170 | USD 60** |
* Motherboards can be found starting at USD 31
** The price includes the motherboard, since this CPU is soldered to the motherboard
Prices were researched at Newegg.com on the day we published this review. TDP stands for Thermal Design Power and states the maximum amount of heat the CPU can dissipate.
Below you can see the memory configuration for each CPU.
| CPU | L2 Cache | L3 Cache | Memory Support | Memory Channels |
| Sempron 2650 | 1 MiB | No | Up to DDR3-1333 | One |
| Celeron J1800 | 1 MiB | No | Up to DDR3-1333 | Two |
Below we have a quick comparison of the video engine of the CPUs.
| CPU | Video Engine | DirectX | Clock | Cores |
| Sempron 2650 | Radeon R3 | 11.2 | 400 MHz | 128 |
| Celeron J1800 | Intel HD | 11 | 688/792 MHz | 4 |
[nextpage title=”How We Tested”]
During our benchmarking sessions, we used the configuration listed below. Between our benchmarking sessions, the only variable device was the CPU being tested and the motherboard, which had to be replaced to match the different CPU sockets.
Hardware Configuration
- Motherboard (Sempron 2650): ASUS AM1M-A
- Motherboard (Celeron J1800): ASRock D1800M
- CPU Cooler: Intel/AMD stock
- Memory: 8 GiB DDR3-2133, two G.Skill Ripjaws F3-17000CL9Q-16GBZH 4 GiB memory modules configured at 1333 MHz
- Boot drive: Kingston HyperX Fury 240 GB
- Video Card: integrated
- Video Monitor: LG Flatron W1942S
- Power Supply: Seventeam ST-550P-AM
Operating System Configuration
- Windows 7 Home Premium 64-bit
- NTFS
- Video resolution: 1440×900 60 Hz
Driver Versions
- AMD driver version: 14.9
- Intel video driver version: 15.33.13.3408
- Intel Inf chipset driver version: 9.4.5.1006
Software Used
Error Margin
We adopted a 3% error margin. Thus, differences below 3% cannot be considered relevant. In other words, products with a performance difference below 3% should be considered as having similar performance.
[nextpage title=”PCMark 8″]
PCMark 8 is a benchmarking software that uses real-world applications to measure the computer performance. We ran three tests: Home, which includes web browsing, writing, light gaming, photo editing, and video chat tests; Creative, which includes web browsing, photo editing, video editing, group video chat, media transcoding, and gaming; and Work, which runs tasks such as writing documents, web browsing, spreadsheets, editing, and video chatting. Let’s analyze the results.
The Celeron J1800 achieved a score 12% higher than the Sempron 2650 in the Home test.
On the Creative benchmark, the Celeron J1800 achieved a score 21% higher than Sempron 2650.
On the Work benchmark, the Celeron J1800 was 24% faster than the Sempron 2650.
[nextpage title=”DivX”]
We used the DivX converter, a tool included in the DivX package, in order to measure the encoding performance using this codec. The DivX codec is capable of recognizing and using not only more than one CPU (i.e., more than one core) but also the SSE4 instruction set.
We converted a Full HD, six-minute long .mov video file into a .avi file, using the “HD 1080p” output profile.
The results below are given in seconds, so the lower the better.
On DivX encoding, the Celeron J1800 was 98% faster than the Sempron 2650.
[nextpage title=”Media Espresso 6.7″]
Media Espresso is a video conversion program that uses the graphics processing unit of the video engine to speed up the conversion process. We converted a 1 GiB, 1920x1080i, 23,738 kbps, .mov video file to a smaller 320×200, H.264, .MP4 file for viewing on a smartphone. The results below are given in seconds, so the lower the better.
Here the Celeron J1800 was 44% faster than the Sempron 2650.
[nextpage title=”DVDShrink”]
DVDShrink is an old but still very useful software to “shrink” video DVDs that have more than 4.7 GiB of data to fit single-layer DVD media. We used it to compress the DVD of “The Lord of the Rings: The Fellowship of the Ring” DVD to 4.7 GiB. The results below are given in seconds, so the lower the better.
In this test, the Celeron J1800 was 35% faster than the Sempron 2650.
[nextpage title=”Far Cry 3″]
Far Cry 3 is based on the Dunia 2 engine, which is DirectX 11. In order to measure performance using this game, we played the same mission three times, measuring the number of frames per second using FRAPS. We ran this game at the lowest possible settings, using 800 x 600 resolution, and overall image quality at “low.”
The results below are expressed in frames per second and are an arithmetic average of the three results collected.
The performance of the Celeron J1800 was 36% higher than the Sempron 2650. However, this does’t really matter, since both CPUs were not powerful enough for running this game at a playable framerate, even at the lowest possible resolution and quality settings.
[nextpage title=”3DMark”]
3DMark is a program with a set of three benchmarks: Ice Storm, Cloud Gate, and Fire Strike.
The 3DMark Ice Storm benchmark measures DirectX 9 performance and it is aimed at low-end computers. The 3DMark Cloud Gate benchmark measures DirectX 10 performance, running at 1280×720 resolution. The 3DMark Fire Strike benchmark measures DirectX 11 performance, and is aimed at high-end gamer PCs, barely running in both tested systems, so we did not include it in this review.
On the Ice Storm Extreme benchmark, the Sempron 2650 was 7% faster than the Celeron J1800.
On the Cloud Gate benchmark, the Sempron 2650 was 11% faster than the Celeron J1800.
[nextpage title=”Conclusions”]
Nowadays, the desktop computer market has specific products for specific needs. And a very large share of the market requires computers where value is more important than performance. May be a grandson buying a computer for his grandma to browse the web, or a company buying dozens of desktops for simple tasks such as data typing, or a store installing price-consulting terminals. Those (among several other) applications need a cheap computer, and if it has a low power consuption, small footprint, and low noise level, it is even better.
As expected, the two biggest desktop CPU manufacturers have products for this market segment. AMD offers the AM1 platform, with Athlon and Sempron processors, while Intel presents the Bay Trail-D family with Celeron J and Pentium J models. Most of the chips from both families are quad-core CPUs, but we chose two of the most inexpensive models, with only two cores each.
The AMD platform has the advantage of being socketed: you can buy the motherboard from a vendor, and the processor from another one. If you want more performance in the future, you just need to replace the chip. The Intel solution, on the other hand, is soldered to the motherboard, which means if you want to upgrade your system, you have to replace the motherboard.
Our tests showed that, between the Sempron 2650 and the Celeron J1800, the Intel solution is faster, and it is probably due to its higher clock rate and dual-channel memory access. Additionally, the model from Intel has a lower TDP (10 W against 25 W of the Sempron 2650), and goes well with a passive refrigeration solution (meaning less noise), which is a good thing to have on this market segment. The Celeron J1800 was faster in all the tests we performed, except in the 3DMark tests.
In addition, a solution based on the Celeron J1800 is a little bit cheaper than a Sempron 2650 solution, since a motherboard with a soldered J1800 costs about USD 60, while if you buy a Sempron 2650 CPU (which costs about USD 40), you will also have to buy a motherboard (which costs today USD 31 or more).
Another important conclusion from our benchmarks is that the integrated video engine of both CPUs are not powerful enough to run a recent game, even at the lowest possible settings. Using a discrete video card with those platforms does not make sense, since the CPU will be a bottleneck. So, it is better to choose a CPU from another family if you are planning to build an inexpensive computer for casual gaming.