General Impressions
After running a number of tests on the R581A with Cyrix, Intel, AMD and IDT chips it appears that M Tech has outdone themselves with this product. All benchmarks ran flawlessly. SDRAM timings were set to X-1-1-1 for 66MHz and 75MHz bus speeds and never failed. The layout is very nice and the features are well implemented. All-in-all, we would have to say that this appears to be M Tech’s best offering to date.
Features
In a nutshell, M Tech included some of the most desirable features from the SiS 5591 chipset on this board. Documented jumper settings are available for bus speeds up to 83MHz, and two undocumented settings. One of those appears to be about 90MHz, while the other simply doesn’t work. Linear Burst mode for Cyrix processors was also included as a BIOS setting, with no jumper settings necessary.
Since the board supports bus speeds above 66MHz, the asynchronous PCI bus feature was also included, with a little twist. At 75MHz bus speed, the user has the option of running the PCI/AGP bus either synchronously or asynchronously. At 83MHz the board will automatically use the asynchronous option. What this means is that those who like to ‘push the envelope’ can run their PCI cards at 37.5MHz and their AGP card at 75MHz, while those more conservative souls can run them at the ‘official’ 33MHz/66MHz speeds even with a Cyrix processor installed.
There are only 2 SIMM slots on the board, which will limit the amount of EDO memory that can be used. This is a particular problem for those who are migrating from an older Pentium class machine and have smaller modules (less than 16MB). On the other hand, 3 DIMM slots allows up to 384MB of SDRAM (though the maximum cacheable limit is ‘only’ 256MB). For a reason unknown to us at this time, M Tech elected to not implement the parity/ECC functionality of the chipset. While the memory slots limit the ‘downward compatibility’ of the memory, M Tech included 3 ISA slots to complement the 4 PCI slots (one is shared, however). Of course, there is one AGP slot as well.
The sample board we were shipped did not include the ATX (20-pin) power connector, however the production models appear to include this feature. In addition, there is a 3-pin connector for a hardware monitoring fan, and the BIOS includes such information as CPU Temperature, CPU/System Fan Control and CPU Voltage. The board also sports a switching voltage regulator with voltage settings from 2.1v through 3.5v, in 1v increments. This should provide compatibility with any processors in the foreseeable future.
Cool and/or Unique Features
The board sports 1MB of L2 cache. This provides the 256MB cacheable range (512k cache would only allow 128MB to be cached on the SiS chipset), and some additional performance as well. When compared with the Gigabyte GA586SG (same chipset, but 512k cache) the performance improvement is about 5% – just about what one would expect.
Layout
The motherboard is laid out very well, considering the limitations of the AT form factor. M Tech learned from their R548 mistake with the CPU socket, and placed it strategically behind the last two PCI slots. This prevents it from getting in the way of the drive bays on smaller tower cases, but also allows for 3 full length ISA cards!
Another item that was well thought out is the SIMM/DIMM placement. Since the DIMM modules can be inserted straight into the socket, while SIMM modules need to be ’tilted’ in, M Tech put the DIMM slots next to the edge of the board (near the PS) and the SIMM slots just to the inside of them. One of the most frustrating experiences is to have a small desktop or mini-tower case where one of the SIMM modules can’t be installed without removing the motherboard due to the proximity of the power supply. This design will eliminate that problem.
The only component that might have been situated a bit differently are the IDE/floppy connectors. These are right in front of the keyboard connector, so you will still have to experience the cable clutter that so many AT boards are prone to. Since the CPU socket was moved slightly towards the drive bays, there just isn’t enough board real estate to have put the IDE connectors there also. They might have placed them behind the PCI slots – next to the memory slots, but that may have caused some problems with laying the trace lines, so we can’t be too critical of this.
Compatibility & Stability
As stated in the preamble, we threw everything we could at this board. AMD K6-200/233, Intel 166MMX, Cyrix 6x86MX PR200/PR233 and an IDT WinChip 200MHz. Every test ran completely and without any errors, and it recognized the IDT chip natively. The board also ran at 83MHz, but only the AMC SDRAM would complete the Winstone 97 test suite at X-2-2-2 timings. While the board would get through POST at the 90+MHZ setting, it would not finish loading Windows no matter how we tweaked the memory timings. It is very likely that current SDRAM simply cannot function at that speed without additional wait states, which are not available in the BIOS at this time.
We also tested with EDO memory (64MB total) and SDRAM from AMC, Crucial Technology and Macrotron (32MB modules). While the default SDRAM timings are X-2-2-2, we took them down to X-1-1-1 to see what would happen. Lo and behold, even at 75MHz the board performed flawlessly. Only at 83MHz did we need to set the timings back to X-2-2-2, and at that speed only the AMC modules would run properly. We filled all three DIMM slots with 3 different manufacturer’s models at 66MHz, and the board didn’t even flinch.
We used two hard drives (neither one was UDMA, unfortunately) – one a Seagate 1.7GB (PIO 4), the other a W.D. 850MB (PIO 3). Even at 75MHz, we had no problems with loading any data or programs. Video cards used were a Diamond 3D 3000 (S3 chipset) and a Jaton Video 57P (Trident chipset). We also located a Viper 330 AGP card, which also worked without any problems, though there was no noticable difference in the light graphics tests we sued (as expected).
Documentation
These days most of the major manufacturers produce good user manuals, and this one is no exception. A ‘quick-reference’ card is included with jumper settings for all current processors. The manual provides settings for all jumpers (except the 90 & 100MHz bus speeds, which you have to figure out yourself). All of the BIOS settings are discussed, though some background knowledge is necessary if you plan to play with most of them. The language is clear and displays no glaring grammar or syntax errors. All-in-all a nice reference manual.
Conclusion
Our conclusion with this motherboard is that it is a good candidate to be called the best M Tech board to date. M Tech has always focused upon compatibility rather than pure speed, and with this product they have achieved that goal with flying colors. The 1MB cache does provide some performance benefits, though it will likely not be among the top products on the more popular performance oriented web sites.
The observation we would have to make is that this product is a very worthy successor to the highly popular R534 series of motherboards. For those who wish to run a Cyrix, this looks like the motherboard to have at this time. Those who are AMD or Intel aficionados still have some good alternatives, but you certainly couldn’t go wrong choosing this product with it’s rich set of features, quality design and great compatibility
This board may not be the fastest Socket 7 board in existence, but it certainly does set the standard to meet or beat for the new generation of ‘Super Socket 7’ boards that will soon hit the market en masse.
Be the first to discuss this article!