July 23, 2001

Review: Soyo SY-K7ADA motherboard

Author: JT Smith

- By Jeff Field -

With the advent of 266MHz DDR memory, many people are looking to build new systems around this new, faster RAM. Many DDR-capable motherboards are rather expensive, coming in at nearly $200, but there are bargain DDR boards out there. Today, I'm reviewing one of them, the Soyo SY-K7ADA.The board
The K7ADA is an ATX board based on the ALI MAGiK1 DDR Athlon/Duron chipset. ALI has been in the chipset market for several years, focusing mostly on the low end. The board is a fairly compact design, the smallest ATX Athlon board I have tested so far.

Surprisingly, though the board is designed more for cost than expandability, the K7ADA has ample room for most people. It comes with five PCI slots and one AGP slot, and while another PCI slot would have been useful, the number of slots is acceptable. For memory expansion, it has three 184-pin DDR slots capable of holding up to 1.5 gigabytes of DDR memory. The board includes two ATA-100 connectors for up to four ATA devices. For external expansion, it includes two USB ports, with the capability of adding four more, as well as two serial ports, one parallel port, one PS/2 mouse port and one PS/2 keyboard port. The board also includes onboard sound, with a game port, line in, line out and microphone ports. The audio is provided by the on-chip audio.

Layout and design
I found the K7ADA to be generally well designed. I found only two issues with the board, the first that the ATX power connector was too far down the board, making it so you have to extend the power cord further, potentially over or around the heat sink, a minor problem. The other issue comes into play with longer AGP and PCI cards, and this one can be more problematic. The floppy connector is right behind the AGP and first two PCI slots, which may obstruct longer PCI cards from being placed in those slots. This would hardly be an issue, but it also affects the AGP slot, of which there is only one. In most cases, the floppy cable can be pressed down enough for it not to matter, but it is something that users may encounter.

Documentation and configuration
Unlike the Gigabyte 7DXR I reviewed last week, the K7ADA does not come with a complete manual. Rather, it comes with a 23-page booklet that contains all the basic information on installing and configuring the board, such as jumper configurations and how to install a CPU in the socket. The documentation is adequate and well written but does not go into great detail.

Configuration of the board was easy, with most settings available through the BIOS. CPU FSB speed is set with a dip switch and is configurable for either 200 or 266MHz.

System Specifications
AMD Athlon 1.4GHz
256 Megs DDR PC2100 RAM from Crucial.com
Western Digital 7200 RPM 10.2 Gig Hard Drive
3Com 3C905TX-C 10/100 NIC (PCI)
300 Watt AMD-Approved ATX Power Supply
Abit Siluro GeForce 2 MX400 64MB AGP

Kernel compiles
In order to test both the boards stability and speed, I ran three sets of Linux kernel compiles on this board. One is a normal, "uniprocessor" make, or make -j1, which is the default. This uses one process, and does not always maximize system usage. I then did make -j2, which spawns a second process. The last test I ran is with make -j3, spawning two extra processes. I do this for several reasons -- to find the "sweet spot" for the board/CPU, as well as to stress the system as much as possible when trying to rate its stability. Also, the kernel is extremely useful as a measure of integer performance.

Kernel 2.4.6 compile times (Minutes:Seconds)

Kernel Compile
Board -j1 -j2 -j3
Soyo SY-K7ADA 4:37 4:32 4:28
Gigabyte 7DXR 4:29 4:20 4:21

Kernel compile times tend to be similar between motherboards that run the same type of CPU, because they are heavily dependent on the CPU. However, we can also see there is an obvious, if small difference between the K7ADA and the 7DXR -- this can probably be attributed to the known poor memory performance of the ALI MAGiK1 chipset -- not devastatingly slow, but measurable, at least. Overall, however, the K7ADA makes out well in this test.

POVRay benchmarks
POVRay is a multi-platform raytracing program. It is a very floating-point intensive task and serves well to help measure the floating point performance of a CPU. For more information on this benchmark, head to the official POVBENCH homepage. The command to run for this benchmark, once you obtain POVRay, you run povray -i skyvase.pov +v1 +ft -x +mb25 +a0.300 +j1.000 +r3 -q9 -w640 -H480 -S1 -E480 -k0.000 -mv2.0 +b1000 from the command prompt. Results are in seconds.

POVRay (Seconds)

Board Result
Soyo SY-K7ADA 15
Gigabyte 7DXR 17

As you can see from this, the results are extremely close, only varying by two seconds. Clearly, POVRay is more CPU dependent than other tests.

Gears, which is included with Mesa, is, to be blunt, a useless benchmark. It is the most basic of OpenGL benchmarks, but because it takes so little time to run (open a console in X, and type 'gears') I figured I would include it for completeness. I say again, this benchmark is almost as useless as BogoMIPS.

Gears (Frames/Second)

Board Result
Soyo SY-K7ADA 1639.000
Gigabyte 7DXR 1646.000

Quake 3 Arena
Quake 3 Arena is perhaps the most popular 3D game available for Linux, and thus is the best benchmark of 3D game performance under Linux. The benchmark is faily simple, open a console (press the ~ key) and type timedemo 1 followed by demo four.dm_66 in the resolutions you wish to benchmark. Once the test is done, press ~ again and the results will be in the console.

Quake 3 Arena timedemos (Frames Per Second)

Gigabyte 7DXR
Board 640x480 800x600 1024x768 1200x1024 1600x1200
Soyo SY-K7ADA 175.4 129.3 85.1 52.8 36.8
Gigabyte 7DXR 175.4 130.2 85.5 52.9 36.8

As you can see, Quake III is dependent on two things, the graphics card, and the CPU. As such, there is little difference between the boards, even at lower resolutions.

Bonnie++ results
Bonnie++ is a hard drive benchmark that tests the writing and reading
from both a single large file (such as that of a database) and many small files (like a proxy, or mail program). It is
useful for simulating performance under such applications.

Bonnie++ results
Sequential Output
Controller Per-Character Block Rewrite
Gigabyte 7DXR 8608 K/sec, 57% CPU 20787 K/sec, 20% CPU 5168 K/sec, 4% CPU
Soyo SY-K7ADA 15544 K/sec, 99% CPU 26980 K/sec, 20% CPU 8951 K/sec, 5% CPU
Sequential Input Random
Controller Per-Character Block Random
Gigabyte 7DXR 11187 K/sec, 61% CPU 21949 K/sec, 14% CPU 111.2 Seeks/sec, 0% CPU
Soyo SY-K7ADA 12660 K/sec, 70% CPU 22296 K/sec, 8% CPU 134.1 Seeks/sec, 0% CPU
Sequential Create
Controller Create Read Delete
Gigabyte 7DXR 18752 /sec, 99% CPU none 22857/sec, 100% CPU
Soyo SY-K7ADA 18623 /sec, 100% CPU none 22322/sec, 100% CPU
Random Create
Controller Create Read Delete
Gigabyte 7DXR 15559 /sec, 86% CPU none 19043/sec, 99% CPU
Soyo SY-K7ADA 17874 /sec, 100% CPU none 18820/sec, 99% CPU

As you can see, in the Bonnie tests, neither board here scores particularly better than the other, except for the K7ADA in the sequential output test. Bonnie++ results depend more on the hard drive than the controller, and so I run this test only to insure there are no problems with the controller on the boards.

HDparm tests the maximum data transfer rate of a hard drive in with two methods, uncached (but buffered still by the hard drives on board buffer) and cached (buffered by the drive and cached via the operating system cache). While the uncached test should not vary between different controllers that support the drive's ATA-version (such at ATA-66, which is what this drive uses), the cached test varies between boards because it is essentially a test of the CPU, cache and RAM on a system. This is what makes it interesting in this case -- it can help showcase the memory performance of these two boards.

hdparm -t (Uncached)
Soyo SY-K7ADA 23.79 MB/sec
Gigabyte 7DXR 23.88 MB/sec
hdparm -T (Cached)
Soyo SY-K7ADA 216.95 MB/sec
Gigabyte 7DXR 220.69 MB/sec

In this test, the Gigabyte 7DXR pulls slightly ahead, by roughly four megabytes per second. Hardly noticeable, but it again illustrates the slight difference in performance between the ALI MAGiK1 chipset the K7ADA is based on and the AMD760.

Overall, the K7ADA is a solid performer. It's not quite as fast as the 7DXR, but at almost half the price, the K7ADA is quite a bargain. If you are looking for DDR memory but don't want to pay too much for a motherboard, the K7ADA is definitely worth a look. You don't get on-board RAID, but if that is not something you need, then this board might be a good way to save yourself some money. With DDR memory being as low as it is -- in fact, I just picked up 768 megabytes of DDR from Crucial.com for $140 -- and Athlon 1.4GHz CPUs going for as little as $160 on Pricewatch, now is an excellent time to upgrade, and by purchasing DDR memory, you might get longer use out of the memory you buy now, because DDR is rapidly replacing PC133. The Soyo SY-K7ADA is available on Pricewatch for about $103 with shipping.


  • Unix
Click Here!